The serious case against historical materialism is not the one most readers meet. The version in circulation is that Marx predicted revolution, revolution did not come where he said it would, and the theory is therefore dead. That argument is answerable in a paragraph, and answering it has allowed the tradition to avoid the objections that are much harder to deal with: that its central form of explanation is one philosophers of science abandoned for good reasons, that its own best historians have shown its core causal claim cannot explain the case it was built to explain, that it has no account of law that is not either circular or artificial, and that it cannot generate history anyone would recognise as history.

Six objections belong in a serious case, and each has a named author who advanced it in an identifiable work. Karl Popper on the prediction record and how it was handled. Jon Elster on functional explanation without a mechanism. Robert Brenner, from inside Marxism, on the failure of productive force primacy to account for the agrarian origins of capitalism. The Weberian and institutionalist tradition on multiple causal orders and on law as constitutive rather than derivative. E. P. Thompson, also from inside, on a structural model with no room for experience and agency. And the empirical record on class identity, which lost to national and religious identity at the moments the theory said it should not.

This article states each objection as its author stated it, says what it gets right without defensiveness, gives the best available reply, and reaches a verdict of lands, partially lands, or misses. A Marxist reader should finish better armed than they started; a critical reader should finish better informed. Both outcomes depend on one rule, which is the article’s namable claim and which almost nothing written on this subject observes.

The rung rule

Historical materialism is not one claim. It is a ladder of claims of decreasing strength, running from the assertion that the productive forces determine the relations of production which determine everything else in a fixed universal sequence, down through productive force primacy, down through last-instance determination, to the bare statement that material conditions constrain what is historically possible. The complete guide to historical materialism sets the ladder out rung by rung with the textual source for each, and this article assumes it rather than restating it.

The rung rule is this: an objection only counts against the rung it actually targets. The two most common failures in this debate are symmetrical and both are everywhere. A critic refutes the top rung, the universal stage sequence that Marx explicitly repudiated in a letter drafted in 1877, and announces that historical materialism is finished. A defender retreats to the bottom rung, the claim that material conditions matter, and announces that the theory is obviously true. Neither has engaged the other, because they are standing on different rungs and neither has said which.

Applying the rule is a matter of asking, of every objection, which claim would have to be abandoned if the objection succeeded. Some of the six objections below take out the top two rungs and leave the rest standing. One of them, the functional explanation objection, threatens the middle of the ladder where most serious defences of the theory live. One, Thompson’s, is not really an objection to any rung but to a way of using the theory. Sorting them is most of the work.

Which objection to historical materialism is hardest to answer?

The functional explanation objection, because it targets the rung where the theory’s best defenders stand rather than the crude version. If relations of production obtain because they develop the productive forces, the explanation runs from effect to cause, and no mechanism has been specified that would select for relations with that property.

Objection one: the prediction record, and what Popper actually argued

Popper’s argument is misstated more often than any other in this literature, usually by people invoking it in his support.

He did not claim that Marxism was unfalsifiable from the beginning. He claimed the opposite, and the claim is the premise of his whole case. In The Poverty of Historicism, in The Open Society and Its Enemies, and most compactly in the essay on the problem of demarcation later collected in Conjectures and Refutations, his position is that Marx’s theory was scientific in the strict sense because it made risky predictions that could have failed. It predicted that socialist revolution would come first in the most industrially advanced countries. It predicted that the condition of the working class under capitalism would worsen. It predicted that intermediate strata would be squeezed into the two great contending classes. These are testable and they were tested.

What Popper objected to was what happened next. Rather than treating the failures as refutations, followers of Marx reinterpreted the theory and the evidence so that they agreed, introducing auxiliary hypotheses that rescued the framework at the cost of its testability. He called this a conventionalist twist, and his charge is that the procedure destroyed a scientific status the theory had genuinely possessed. The accusation is therefore about the conduct of a research tradition rather than about a doctrine, which makes it far harder to answer than the crude version, since a defender cannot reply by producing a testable claim from Marx. That there were testable claims in Marx is Popper’s own starting point.

The record itself, stated plainly

The predictions and what happened should be set out without euphemism, because a critique article that softens them has failed its purpose.

Revolutions carried out in Marx’s name occurred in agrarian economies at the periphery of the world market, in Russia and later in China, rather than in Britain, Germany, or the United States. This is not a marginal miss. The expectation that the most developed country shows the less developed the image of its own future is stated in Marx’s own preface to the first volume of Capital, and it was the basis on which the movement organised for two generations.

The condition of the working class in the industrial core improved substantially over the long run rather than declining towards subsistence. The measurement disputes here are real and should be stated as such: the timing of the improvement, whether real wages fell during the early phase of industrialisation before rising, how to price goods whose quality changed, how to treat urban disamenities, and how far gains were distributed across the whole class are all contested in the economic history literature, and the range of estimates for the early industrial period is wide enough that both an optimistic and a pessimistic account have been sustained by serious scholars. What is not contested is the direction over a century and a half in the industrial core, and no reading of the theory that requires absolute immiseration there can survive it.

Intermediate strata did not disappear. Professional, technical, managerial, and service occupations expanded rather than being ground between the two great classes, and the class structure of developed economies became more differentiated rather than less.

Class identity lost decisively to national identity in the summer of 1914, when the socialist parties of Europe voted for war credits against their own international commitments, and it has lost to national and religious identity repeatedly since.

What survives the record

Three replies are available and they differ sharply in quality.

The weakest is that Marx never made predictions, only conditional statements. This is not tenable. He made predictions, and defending the theory by denying that he did is the manoeuvre Popper described.

Stronger is the reply about uneven and combined development, which explains revolution at the periphery by the position of a society within a world economy rather than by its internal level of development, so that the chain breaks at its weakest link. This has a real virtue that critics rarely concede: it was formulated before the event rather than as a rescue afterwards. Whether it is a legitimate development or a conventionalist twist depends on whether it generated new expectations that could be checked, and the honest answer is that it generated some, including the expectation that revolutionary crisis concentrates where advanced and backward forms of production are most violently combined, and that this expectation has a mixed record.

Strongest is the distinction between an explanatory framework and a forecasting instrument. Historical materialism, on this reply, is a set of claims about what explains institutional change, not a machine for predicting dates. Frameworks are assessed by whether they identify the right things to look at and generate good explanations of cases, and by that standard the record is much better than the forecasting record. This reply is legitimate and it has a cost that defenders should acknowledge: it moves the theory down the ladder, and once forecasting is abandoned, the question of what the framework forbids becomes urgent. That question is worked out in the debate over economic determinism, which owns it.

Verdict: partially lands. Popper’s charge succeeds against the top rung and against a good deal of twentieth-century practice. It does not establish that the framework is unfalsifiable in principle, because the qualified version can be made to forbid things, and a research programme that protects its core while generating new work is doing what every research programme does. The full treatment of Popper’s attack on Marxist method works through the historicism and essentialism arguments this summary compresses.

Objection two: functional explanation without a mechanism

This is the objection that should worry a defender most, and it is the one least often stated in general accounts, because it requires setting out the strongest version of the theory before it can be made at all.

Cohen’s defence, which is also the biggest target

G. A. Cohen’s Karl Marx’s Theory of History: A Defence, published in 1978, is the most rigorous statement historical materialism has ever received, and it was written by a defender. Its achievement was to state the theory as a set of precise propositions that could be assessed, which nobody in the tradition had previously done with comparable care.

Two theses carry the argument. The development thesis holds that the productive forces tend to develop throughout history, grounded in a claim about human beings as rational agents in a situation of scarcity, with the intelligence to improve their situation and the motivation to do so. The primacy thesis holds that the relations of production are as they are because they are the relations that best develop the productive forces at that level, so that the character of the relations is explained by their effect on the forces. Cohen was explicit that this is a functional explanation: the relations obtain because of what they do.

He was also explicit that functional explanation is legitimate in principle, drawing the parallel with biology, where the claim that a species has a trait because the trait confers a survival advantage is respectable precisely because natural selection supplies the mechanism connecting the effect to the persistence of the cause. He argued that a functional explanation can be well supported by evidence of a consequence law even where the mechanism is unknown, as Darwin’s own predecessors had functional explanations without knowing about genes.

Elster’s objection

Jon Elster’s Making Sense of Marx, published in 1985, attacked precisely this point, and the attack is the origin of the microfoundations requirement in social theory.

His argument runs in three steps. Functional explanation in biology is licensed by natural selection, a well-understood mechanism that operates through differential reproduction over many generations with random variation supplying the raw material. Nothing in the social world corresponds to it. Societies do not reproduce differentially in the required sense, variation in social arrangements is not random, and the number of cases is far too small for a selection story to do the work. Therefore a claim that relations of production obtain because they develop the forces is not a functional explanation at all; it is an observation of a correlation with a causal arrow attached to it by assertion.

The positive requirement Elster derived is the microfoundations principle: any claim about a social structure must in principle be cashable into claims about the beliefs, desires, and actions of individuals and the interactions between them. Without that, the explanation names a relation without saying how it is produced.

The force of the objection is that it does not depend on rejecting Marxism. Elster considered himself to be improving it, and the analytical Marxist current he belonged to was an attempt to rebuild the theory on foundations that would survive the objection.

What replies are available

Three, and they are of unequal strength.

The first is to supply the missing mechanism. Cohen himself gestured at candidates, including the tendency of societies to adopt arrangements that make them militarily and economically competitive with neighbours, and the pressure exerted by individuals and groups with an interest in removing obstacles to production. The competitive mechanism is the most promising because it does supply something analogous to selection: arrangements that leave a society systematically less productive than rivals capable of invading, underselling, or absorbing it do not persist indefinitely. The difficulty is that this mechanism operates only where competitive contact exists and is strong enough, which restricts the theory’s scope considerably and makes it a claim about societies in contact rather than about history as such.

The second is to abandon the functional form and rebuild the theory around class conflict, so that relations change because classes with an interest in changing them acquire the capacity to do so. This is a real repair and it is what the political Marxists did, but it is a substantial retreat: it gives up the primacy of the forces, and with it the claim that the direction of history is explained by the development of productive capacity. What it gains is a mechanism made of nothing but the actions of identifiable people.

The third is to deny that microfoundations are required, arguing that structural explanations in the social sciences are legitimate without reduction to individual action, as they are in other fields dealing with aggregate phenomena. This reply has serious defenders and it is not obviously wrong. Its weakness in this particular case is that Cohen’s explanation is not merely structural but functional, and the objection to functional explanation without selection is not answered by defending structural explanation in general.

Verdict: lands. The objection succeeds against the functional form of the theory, and the most rigorous defence of historical materialism ever mounted has not been shown to survive it. Cohen continued to work on the problem and conceded in later writing that the primacy thesis faced difficulties he had not resolved, and his subsequent turn towards political philosophy was in part a consequence. The profile of Cohen and the functional explanation debate sets out the exchange in full and traces what he conceded and when.

Objection three: the transition to capitalism, from inside Marxism

The third objection is the most damaging, and it comes from a Marxist historian using Marxist categories on the case the theory was built to explain.

Brenner’s argument

Robert Brenner’s essay on agrarian class structure and economic development in pre-industrial Europe, published in the journal Past and Present in 1976, set off what became known as the Brenner debate. Its target was not Marxism but the two dominant explanations of European economic divergence: the demographic model, which explained changes in the position of peasants and lords by long swings in population, and the commercialisation model, which explained the dissolution of feudal relations by the growth of trade and markets.

His objection to both is that they treat the same cause as producing the same effect everywhere, and the record shows it did not. The same demographic collapse in the fourteenth century was followed in western Europe by the erosion of serfdom and in eastern Europe by its intensification. The same expansion of long-distance trade coincided with the development of capitalist agriculture in England and with the consolidation of a peasantry secure in its holdings in France, alongside a nobility that turned to state office rather than to agricultural improvement. If the same cause produces opposite outcomes, the cause is not doing the explanatory work.

What does the work, on Brenner’s account, is the outcome of class conflict over rent, tenure, and the terms on which surplus is extracted. Where lords succeeded in dispossessing the peasantry but could not restore direct coercive extraction, as in England, the result was a triangular structure of landlords, capitalist tenant farmers, and wage labourers, in which all parties were subject to competitive pressure and therefore to the imperative to improve productivity. Where peasants secured effective possession, as in much of France, the result was an agriculture with no such imperative and a state that extracted through taxation rather than through market compulsion.

Why this is an objection rather than an application

Because the explanation reverses the order the theory specifies. Productive force primacy says that relations obtain because they develop the forces. Brenner’s account says that a particular configuration of relations, arrived at through a conflict whose outcome was not determined in advance, is what generated the imperative to develop the forces. Relations explain forces, not the other way round, and the specific relations in question were produced by struggles that could have gone otherwise and did go otherwise a few hundred miles away.

The current that developed from this argument, associated also with Ellen Meiksins Wood and often called political Marxism, drew the general conclusion explicitly: there is no transhistorical tendency for productive forces to develop, because the compulsion to improve productivity is itself a historically specific feature of capitalist social property relations rather than a universal feature of human societies.

That conclusion, if accepted, removes the development thesis and the primacy thesis together, which is to say the whole of Cohen’s ladder above the level of class conflict.

The replies

The strongest reply is that Brenner’s account is itself a materialist explanation and remains inside the framework, since it explains institutional outcomes by the structure of property relations and the balance of class forces rather than by ideas or accident. This is correct and it is important, but it does not save the rungs under attack; it relocates the theory to a lower one.

A second reply questions whether the divergence was as sharp as the account requires and whether the English case is being made to carry too much weight, and there is a substantial critical literature on exactly this. The debate is genuinely unresolved, which is a fact worth stating rather than a reason to dismiss the objection.

A third reply, more concessive, accepts the argument for the transition specifically while holding that productive force primacy may still hold within capitalism, where competitive compulsion is general. This is coherent and it is a further narrowing of scope.

Verdict: lands, and lands from inside. The internal provenance is what makes it decisive: this is not an outsider misunderstanding the theory but a practitioner testing its central causal claim against the case it was designed for and finding that the arrow points the other way.

Objection four: multiple causal orders and the problem of law

The fourth objection comes from the Weberian and institutionalist traditions and has two parts, one general and one specific. The specific part is the harder.

The general form

Max Weber’s position, developed across his sociology of religion, of law, and of domination, is that social life contains several orders that are analytically distinct and none of which is reducible to the others: the economic order, the political order concerned with the distribution of power, and the status order concerned with honour and social estimation. Each has its own logic, each can be a source of stratification, and the relations between them are contingent and vary by case. His account of class, status, and party is the compressed statement of this, and the comparison of Marx and Weber on class owns that particular argument.

The objection to historical materialism follows directly. If several orders are causally effective and their relations vary, then a theory assigning explanatory priority to one of them across all cases has assumed what needs to be shown, and the correct procedure is comparative investigation of how the orders interacted in each case.

Weber’s own methodological warning matters here and is routinely omitted by those who invoke him. At the close of his study of the Protestant ethic he cautioned against replacing a one-sided materialist interpretation of culture and history with an equally one-sided spiritualist one, and said that each is a preparation for investigation rather than a conclusion. He is objecting to monocausal explanation, not proposing an idealist alternative, and a writer who uses him as an anti-Marxist weapon is using a reading he rejected in the text itself.

The specific form: law

The specific objection is sharper and it is a genuine structural difficulty rather than a matter of emphasis.

Relations of production are relations of effective control over productive resources and over labour. In any developed society those relations are constituted by legal categories: ownership, tenure, contract, employment, corporate personality, security interests, and the rest. If the base is constituted by law, and law belongs to the superstructure that the base explains, the explanation is circular. This is not a Weberian invention; it is a problem the tradition has recognised and worried at for a century, and the institutionalist literature on the legal foundations of markets has sharpened it considerably.

Two solutions exist and each costs something.

Cohen’s solution is to describe relations of production in terms of powers rather than rights, so that the base can be specified without legal vocabulary. A person has effective power over a resource if they are in fact able to use and dispose of it, whatever the legal position, and law is then explained as the form in which those powers are stabilised, made predictable, and enforced. This is logically clean and it is the most rigorous answer available. Its cost is artificiality: nobody experiences effective control over a factory in abstraction from the legal title, the two are not separable in practice, and the description of a power structure that is deliberately silent about the rules constituting it is a description of something no participant would recognise.

The second solution denies the sharp separation, treating law as internal to the relations of production rather than derived from them. This is closer to how a working historian describes an eighteenth-century property regime or a colonial land settlement, where the law is the arrangement rather than a reflection of it. Its cost is that it weakens the model substantially, since if law is part of the base, the claim that the base explains the superstructure has lost much of its content.

Verdict: partially lands. The general form is a caution rather than a refutation: it establishes that priority must be argued case by case, which good Marxist historical work already does. The specific form on law is a real and unresolved difficulty for any version of the theory that keeps a sharp base and superstructure distinction, and it is the point at which the model is weakest. The guide to base and superstructure works through the boundary problem in full.

Objection five: agency, experience, and the writing of history

The fifth objection comes from a historian and is directed less at a proposition than at a practice.

Thompson’s argument

E. P. Thompson’s The Poverty of Theory, published in 1978, is a polemic against Althusserian structuralism, and behind the polemic is a claim about what a theory of history has to be able to do. The positive statement of it is in the preface to The Making of the English Working Class, published in 1963, where he insists that “class is a relationship, and not a thing”, that it happens when people as a result of common experiences articulate the identity of their interests against others, and that it cannot be found by taking a cross-section of a society at a single moment and looking for a structure.

The objection that follows has three parts. A model in which classes are positions in a structure and individuals are bearers of those positions cannot explain why people in the same position behaved differently in different places and periods, which is the thing a historian actually has to explain. A model with no place for experience cannot account for the process by which a common situation becomes a common identity, which is neither automatic nor guaranteed and which is the whole subject of labour history. And a theory that treats the categories as prior to the evidence produces history in which the archive illustrates the theory instead of testing it, which Thompson regarded as a corruption of the discipline.

His charge against the structuralist version was blunt: it converted a historical method into a closed system in which the terms defined each other and nothing external could disturb them.

The reply

The reply concedes more than defenders usually notice, because Thompson’s objection is largely to a version of Marxism rather than to Marx. Nothing in the 1859 formulation requires that classes be treated as static positions or that experience be excluded, and Marx’s own historical writing, particularly the analysis of French politics after 1848, is full of the specific political conduct of specific groups behaving in ways their economic position did not dictate. Thompson himself insisted throughout that he was working within the tradition rather than against it.

What the objection does establish is a constraint on how the theory may be used. A structural account that cannot say how a position becomes a disposition, and a disposition an action, has not explained anything historical; it has classified. The requirement is close to Elster’s microfoundations principle arrived at from the opposite direction, by a historian who distrusted analytical philosophy entirely, and the convergence is worth noticing.

Verdict: partially lands, against a use rather than a claim. The objection is fatal to a mechanical structuralism and leaves the theory’s substantive claims untouched. Its practical value is high: it is the best available diagnostic for distinguishing Marxist history that explains from Marxist history that labels.

Objection six: class identity and its rivals

The sixth objection is empirical and is the one that has done most damage in practice, because it concerns the mechanism by which the theory expected its predictions to be realised.

The argument

Historical materialism holds that the relations of production distribute people into groups with opposed interests, and that these groups become the agents of historical change. The transition from a shared situation to a shared political identity is essential to the account; without it, classes are categories rather than actors.

The record shows that this transition is unreliable and that it competes with other bases of identity which have repeatedly proved stronger. In 1914 the socialist parties of Europe, which had committed themselves to opposing a war between their governments, voted for war credits in their national parliaments. Religious identity has organised political life in large parts of the world where the theory expected class to do so, and continues to. Ethnic and communal identity has repeatedly cut across class alignment, including within workforces sharing an identical position in production. National liberation movements mobilised across class lines with far more success than class movements mobilised across national lines.

The objection is not that people sometimes act on other identities. It is that the theory specifies a mechanism, the formation of class consciousness from a common material situation, and the mechanism has an inconsistent record at exactly the moments the theory identifies as decisive.

The replies

The most substantial reply is that the theory never held that class identity forms automatically, and that the whole tradition from Lenin’s insistence on the necessity of political organisation to Gramsci’s account of the work required to build a collective will is an admission that it does not. On this reading the objection targets a spontaneist version of the theory that its most serious exponents rejected. The complete guide to class consciousness sets out how the tradition handled the gap between a class in itself and a class for itself, and this article does not restate it.

A second reply is that national and religious identities are themselves historically produced and have material conditions of existence, so explaining their strength is a task for the theory rather than a refutation of it. There is real work behind this, including analyses of nationalism that connect it to the requirements of industrial economies and to the formation of states, and it is not an evasion. It does however concede something important: if the theory has to explain why class loses, it can no longer treat class as the default outcome that other factors merely delay.

The weakest reply, and one still encountered, is that identities other than class are forms of false consciousness. This explains nothing, is not supported by Marx, and converts a failed prediction into an accusation against the people whose behaviour falsified it.

Verdict: lands against the strong version, and reshapes the weaker one. The theory can accommodate the record only by treating class formation as a contingent political achievement rather than a structural consequence, which is a considerable change to what it originally claimed.

The objections ledger

The table is the instrument this article exists to produce. Each row names the objection, its author and the work in which it was made, the rung of the ladder it targets, the best reply available, and an honest verdict. The verdict column is the one to read first, and its most important feature is that it is not uniform.

Objection	Who made it, and where	Rung targeted	Best reply	Verdict
The predictions were testable, failed, and were rescued by auxiliary hypotheses	Karl Popper, The Poverty of Historicism and The Open Society and Its Enemies, developed in the demarcation essay in Conjectures and Refutations	Top rung: the universal sequence culminating in revolution in the most advanced economies	Uneven and combined development was formulated in advance rather than after the fact; and an explanatory framework is not a forecasting instrument, though conceding this moves the theory down the ladder	Partially lands: fatal to the top rung and to much twentieth-century practice, not to the framework as such
Functional explanation without a selection mechanism is not explanation	Jon Elster, Making Sense of Marx, against G. A. Cohen’s Karl Marx’s Theory of History: A Defence	Productive force primacy, the rung where the theory’s best defenders stand	Competitive pressure between societies in contact supplies something analogous to selection, at the cost of restricting the theory’s scope to societies in such contact	Lands: the most rigorous defence has not been shown to survive it
Productive force primacy cannot explain the agrarian origins of capitalism, since identical causes produced opposite outcomes east and west	Robert Brenner, in Past and Present, and the debate that followed; developed by Ellen Meiksins Wood	The development thesis and the primacy thesis together	The account is itself materialist and remains inside the framework, explaining outcomes by property relations and class forces	Lands, and lands from inside: it relocates the theory to a lower rung rather than refuting materialism
Several causal orders are effective and their relations vary by case, so priority cannot be assumed	Max Weber, on class, status, and party, and the sociology of law and religion; extended by the institutionalist literature	Any rung that assigns priority across all cases in advance	Good Marxist historical work already argues priority case by case rather than assuming it	Partially lands: a methodological caution rather than a refutation
Law constitutes the relations of production, so it cannot be a superstructure the base explains without circularity	The same tradition, sharpened by work on the legal foundations of markets	Last-instance determination and every version keeping a sharp base and superstructure boundary	Describe the base in terms of powers rather than rights, so it can be specified without legal vocabulary	Partially lands: the reply is logically clean and descriptively artificial, and the problem is unresolved
A structural model with no room for experience and agency cannot produce history	E. P. Thompson, The Poverty of Theory, with the positive statement in the preface to The Making of the English Working Class	Not a rung: a way of using the theory	Nothing in Marx requires static classes, and his own historical writing is full of contingent political conduct	Partially lands, against a use rather than a claim: fatal to mechanical structuralism
Class identity lost to national and religious identity at decisive moments	The historical record, from the war credits votes of 1914 onwards	The mechanism by which classes become historical agents	Class formation was always held to require political work rather than to follow automatically, from Lenin through Gramsci	Lands against the strong version and reshapes the weaker one

Reading down the verdict column produces the finding this article was written to establish. Two objections land outright, one of them from inside the tradition. Four partially land. None of the seven refutes materialism as such, and none leaves the strong version standing.

The strongest Marxist reply, stated at full strength

A critique article that states objections at full strength and replies in fragments has not been fair, so the best case on the other side is set out here as its defenders would state it.

The reply begins by noticing what the objections have in common. Every one of them either targets the strongest version of the theory, which its most serious modern exponents had already abandoned, or converts into a constraint on how the theory should be used. What none of them does is offer a better account of the thing the theory exists to explain.

That thing is a specific and difficult question: why the institutional order governing production and appropriation changes at all, and why it changes in the ways it does. Not why a particular war broke out, or why a doctrine spread, but why arrangements as durable and self-reinforcing as serfdom, chattel slavery, guild regulation, and the various forms of tenancy were eventually replaced by others, and why the replacements have the character they do. This is the question historical materialism was constructed to answer, and the alternatives on offer answer it less well or do not attempt it.

The reply then makes three specific claims. Frameworks are displaced by better frameworks, not by lists of difficulties, and no rival account of long-run institutional change in the organisation of production has established itself. The objections that land establish that the strong version is false, which the tradition’s own best work established independently and earlier, and the collapse of a strong version is the ordinary fate of an ambitious theory rather than evidence of emptiness. And the framework continues to generate research that other frameworks do not, in economic history, in the study of labour, in the analysis of colonial land settlement, in agrarian studies, and in the study of the state, which is the practical test of a research programme.

The reply concludes with a point about provenance that is not merely rhetorical. The sharpest objections in the ledger came from inside: Brenner and Wood on the transition, Thompson on structure and experience, Cohen and then Elster on the logic of the explanation, all of them working within a tradition that was arguing with itself. A framework that produces its own most damaging critics is behaving like a research programme rather than a doctrine, and the pattern recurs across the whole of this series: the strongest criticisms of Marxist positions are overwhelmingly internal ones.

That reply deserves its own scrutiny, and two things can be said against it. The argument that no better framework exists is only as good as the survey behind it, and the institutionalist literature on long-run economic change is a serious competitor that Marxist writers have often engaged with less carefully than they engage with each other. And the observation that a tradition produces its own critics is compatible with the tradition being wrong, since a degenerating research programme also produces internal critics, often in larger numbers.

What each objection gets right

Stating this separately, without a reply attached, is a discipline worth observing, because the reply-shaped paragraph tends to swallow the concession.

Popper is right that the theory made predictions, that they failed, and that the failures were frequently handled by reinterpretation rather than by revision. Anyone who has read the polemical literature of the interwar period will recognise the pattern he described.

Elster is right that a claim about a social structure needs a story about how it is produced by the actions of people, and that the analogy with natural selection does not transfer. The requirement he formulated has improved social theory generally and it improved Marxism specifically.

Brenner is right that identical causes produced opposite outcomes across Europe, and that any theory in which the growth of trade or the movement of population explains the dissolution of feudal relations is refuted by the eastern European case. The demonstration is straightforwardly empirical and it has not been overturned.

The Weberian tradition is right that causal priority is a question for investigation rather than a premise, and the institutionalist tradition is right that the constitutive role of law in economic relations is a genuine problem for a model that puts law in a derived level.

Thompson is right that a class is not a thing found by taking a cross-section, and that the formation of a shared identity from a shared situation is the historical process requiring explanation rather than a step that can be assumed.

The record is right that class has repeatedly lost to national and religious identity, including at moments the theory identified as decisive, and no amount of subsequent explanation converts that into a success.

The objections that do not work

An honest case against a theory has to say which of the standard attacks fail, and several of the most popular ones do. Including them here is not a defensive manoeuvre; a critic who leads with these will be answered easily and will have wasted the opportunity to make the objections that land.

Marxism is unfalsifiable because Marxists explain everything after the fact

This is the crude version of Popper’s argument and it is not Popper’s argument. It also proves too much: any framework can be applied badly after the event, and the ability to accommodate a case is not by itself evidence of vacuity, since a good theory should be able to explain the cases in its domain. The objection only bites when it is made specific, by identifying a claim, showing it failed, and showing that the response was an ad hoc modification rather than a development with new implications. That is work, and the crude version is a way of avoiding it.

The Soviet Union collapsed, therefore the theory is refuted

This confuses a theory with a regime, and the confusion runs in both directions. Nothing in historical materialism predicts that a state calling itself socialist will endure, and the collapse of such a state is evidence about that state. Conversely, nothing in the theory is vindicated by any regime’s achievements. Keeping the analytical question separate from the moral and political one is a requirement for anyone writing seriously about this, and the tradition’s own most careful work on the socialist states has always done so. The historical record of those states, including famine, repression, and mass death, is a subject in its own right and is dealt with accurately in the cluster devoted to it, and it is not an argument about the theory of history.

Marx got human nature wrong

This objection is popular and is usually made against a position Marx did not hold. He did not claim that human beings have no fixed characteristics; he claimed that many of the characteristics presented as fixed by the political economy of his day were the characteristics of people living under a specific set of relations. The claim is historical rather than metaphysical, and the sixth of the Theses on Feuerbach defines the human essence as the ensemble of social relations rather than denying that there is one. There is a serious version of this objection concerning motivation and incentives, and it belongs to the argument about socialist economic organisation rather than to the theory of history.

The theory is deterministic and denies free will

The historical claim and the metaphysical claim are independent, as the debate over economic determinism sets out. Aggregate regularities are compatible with individual indeterminacy, and Marx did not address the metaphysical question. The serious version of the objection is about the practical incoherence of fatalism for a political movement, and it was made most effectively from inside the socialist movement rather than from outside it.

Marx predicted absolute immiseration and wages rose

This one requires care with the text, and a critic who states it carelessly will be corrected. The general law of capitalist accumulation in the twenty-fifth chapter of the first volume of Capital states that accumulation at one pole is at the same time “accumulation of misery” at the opposite pole, and the passage has been read as absolute immiseration and as relative immiseration for over a century. The relative reading has substantial textual support, since Marx elsewhere explicitly allows that real wages can rise while the share going to labour falls and while the intensity of work increases, and the mechanism he specifies, the reserve army of labour, produces pressure on wages rather than a guaranteed decline. The reading also has to accommodate the strongly worded passages about degradation, which are not obviously about relative shares.

The objection that does work here is narrower and is worth stating precisely. Whatever the correct reading of the text, the movement built on it expected the material position of workers in the industrial countries to deteriorate, organised on that expectation, and was wrong. That is a claim about the tradition’s expectations rather than about the exegesis of one chapter, and it does not depend on winning the textual argument.

The misreadings that distort this debate

Five recur often enough to be worth naming.

Popper’s objection is not the simple unfalsifiability charge. He argued that the theory was scientific and was made unscientific by the way its followers handled refutation. Getting this wrong forfeits the strongest version of the argument.

The failure of a prediction does not automatically refute an explanatory framework. It refutes the prediction, and whether the framework survives depends on whether the modification that saves it generates new work or merely absorbs the anomaly. That is a judgment about a research programme over time and it cannot be made in a sentence.

Internal criticism is the tradition working, not defecting. Treating Brenner, Thompson, Cohen, or Elster as apostates misdescribes what they were doing and, more practically, deprives a critic of the best material available.

Immiseration was not straightforwardly a claim about absolute wages everywhere, and a critic who asserts that it was will be corrected on the text and will lose the exchange.

Criticising historical materialism does not commit anyone to an alternative. A reader can hold that the primacy thesis fails without adopting institutionalism, Weberian pluralism, or any other framework, and the absence of a preferred rival is not a reason to accept the theory. The converse also holds, which is why the strongest Marxist reply above, that no better framework has established itself, is a consideration rather than a proof.

The honest verdict

Stating a verdict requires saying what would change it, and both parts belong here.

The strong version of historical materialism is false. The universal sequence of stages is not in Marx and was repudiated by him in a letter drafted in 1877, where he refused to have his sketch of western European development converted into a “historico-philosophical theory of the general path every people is fated to tread”. The development thesis and the primacy thesis, the most rigorous form the strong version has received, face an unanswered objection about mechanism and a demonstrated counter-case in the transition to capitalism. The prediction record is bad. The expectation that class would reliably become the operative political identity has been disappointed repeatedly.

What survives is a research programme rather than a science of history, and its content is real. It holds that the organisation of production and appropriation is a primary object of historical explanation rather than a background condition. It holds that durable institutional arrangements require material reproduction and that an account which cannot say who bears the costs and who captures the benefits is incomplete. It holds that conflict over the terms on which surplus is produced and appropriated is a recurrent and often decisive source of institutional change. It holds that the categories through which an economy is understood are historically specific rather than natural. Each of these is a substantive commitment, each can be wrong, and each has generated historical work that other frameworks did not.

What would change this verdict in the theory’s favour is a specified mechanism connecting economic structure to institutional outcomes that is precise enough to generate expectations and general enough to apply across cases. That is the missing piece, it has been missing since Elster named it, and supplying it is the most valuable thing anyone working in this tradition could do.

What would change the verdict against it is a demonstration that the surviving commitments are routinely violated in the record: that durable institutional orders regularly persist with no material basis, that arrangements systematically blocking productive development survive competitive contact indefinitely, and that large-scale transformations in how societies are organised occur with no relation to conflicts over surplus. Nothing in the current historical literature suggests any of these, which is why a theory whose strong version has been refuted several times over is still in use by working historians who have no political attachment to it at all.

How the objections interact

Treating the six as a list understates the case, because several reinforce one another and one of them cuts against the rest.

Elster’s objection and Brenner’s converge from opposite directions on the same point. Elster argues that the functional form lacks a mechanism; Brenner shows, in the case the theory was built to explain, that the mechanism people had assumed does not operate, since the same commercial and demographic pressures produced opposite institutional outcomes. One is a logical objection and the other is an empirical one, and together they leave productive force primacy without a defence at either level.

Thompson’s objection and Elster’s also converge, which is remarkable given their mutual hostility. Both demand that a structural claim be cashable into something people actually did, one in the language of microfoundations and the other in the language of experience and agency. A defender who satisfies one has largely satisfied the other.

The objection about class identity connects to Thompson’s directly. If the formation of a class as a political agent is a contingent achievement rather than a structural consequence, then the record of class losing to national and religious identity is not an anomaly to be explained away but the expected variance in a process that was never guaranteed. Conceding Thompson’s point substantially defuses the sixth objection, at the price of abandoning the claim that made the theory politically confident.

The one that cuts against the others is Popper’s, and this is worth noticing. His complaint is precisely that the theory absorbs difficulties by modification. A defender who responds to Brenner by relocating the theory to a lower rung, to Elster by narrowing its scope to societies in competitive contact, and to the record by treating class formation as contingent, has performed a sequence of modifications that fits Popper’s description. Whether the sequence is progressive or degenerating depends on whether the modified theory generates new work, which is a question about a research programme over decades rather than about a proposition.

Do the objections add up to a refutation?

They refute the strong version and leave a narrower research programme standing. Two land outright, four partially. None supplies a better account of why the institutional order of production changes, which is the question the theory exists to answer, and none establishes that materialist explanation as such fails.

What a rival framework would have to do

Criticism is cheap unless it is measured against alternatives, and the alternatives deserve statement.

The institutionalist tradition in economic history is the most serious competitor. It explains long-run divergence in economic performance by the character of institutions, particularly those securing property rights and constraining rulers, and by the political settlements that produced them. Its strengths are real: it has quantitative purchase, it engages the comparative record systematically, and it supplies mechanisms connecting institutional form to economic outcome. Its recognised difficulty is explaining where institutions come from, and the more sophisticated work in this tradition has moved towards accounts of distributional conflict that are closer to Brenner than its practitioners usually acknowledge.

The Weberian tradition offers a comparative sociology of several causal orders with no assigned priority. Its strength is fidelity to the variety of the record; its weakness is that a framework which refuses priority in advance provides less guidance about where to look and produces case studies more readily than general accounts.

Demographic and ecological accounts explain change by population pressure, disease, climate, and resource constraint, and they have had considerable success in specific periods. Their difficulty is the one Brenner identified against the demographic model: the same pressure produced different institutional outcomes in different places, so the pressure cannot be the whole explanation.

Evolutionary and selection-based accounts of institutional change attempt to supply exactly the mechanism historical materialism lacks, by specifying how variation arises and how some arrangements are retained. They are the most interesting current competitor precisely because they address the gap Elster identified, and their difficulty is the smallness of the sample and the non-randomness of the variation, which are the same problems Elster raised against Cohen.

Is there a better framework for explaining institutional change?

None has established itself. The institutionalist tradition is the strongest competitor and shares the difficulty of explaining institutional origins. What can be said is that the field now contains several frameworks with partial success rather than one that has displaced the others, which is a different situation from the one either side usually describes.

The honest summary is that historical materialism is one of several partially successful frameworks, that it retains a distinctive question and a distinctive answer, and that its principal weakness, the missing mechanism, is a weakness shared in some degree by every rival except the evolutionary accounts, which pay for the mechanism with problems of their own.

Using this material under deadline

A writer who needs to make a confident claim about Marxism being discredited should check four things first, because this is the area where quick writing goes wrong most reliably.

Has Marxism been debunked?

The strong version of its theory of history has been refuted, on the record and by argument, much of it internal. Materialist explanation of institutional change has not been, and remains in routine use by historians with no political attachment to it. Claims that the whole framework has been disposed of overstate what the objections establish.

Check whether the claim being attributed is one anybody holds. The universal five-stage sequence is the version most often attacked and it is not Marx’s; it was formalised much later, most influentially in a Soviet statement of the 1930s, and Marx explicitly refused to universalise his account. Attacking it is easy and establishes nothing about the positions actually defended.

Check the Popper attribution before using it. The claim that Popper showed Marxism to be unfalsifiable is the most common misattribution in this field. He argued that it was falsifiable, was falsified, and was then rendered unfalsifiable by its defenders, which is a different and more interesting claim, and quoting him for the crude version invites correction from anyone who has read him.

Check whether a historical record is being used to settle a theoretical question. The performance of states that described themselves as Marxist is a subject with its own evidence and its own scholarship, and it is not evidence about a theory of historical change in either direction. Writers under time pressure conflate the two constantly, in both directions.

Check the immiseration claim if you intend to use it. The textual position is genuinely disputed between absolute and relative readings, and the defensible form of the objection concerns what the movement expected and organised for rather than what one chapter of Capital means.

The reliable material sits in the primary works named in the ledger above, each of which is a book or article by an identifiable author making an identifiable argument, and in the historiographical literature on the Brenner debate, which is unusually well documented because it was conducted in a single journal over several years. Keeping a card for each objection with its author, work, target, and the standard reply attached is the practical defence against the errors above, and you can save your notes and build a citation-linked reading list free on VaultBook so the attribution travels with the argument.

Writing about this in an examination

Evaluation questions on historical materialism are where marks are won, and this material is what wins them, provided it is used with the rung rule rather than as a list.

The single distinction that earns marks is between an objection to the strong version and an objection to materialist explanation as such. A candidate who writes that the theory has been criticised for determinism, for being unfalsifiable, and for ignoring culture has produced a list. A candidate who writes that the functional form of the primacy thesis faces an unanswered objection about mechanism, that this objection came from within the analytical Marxist current, and that the theory survives at the level of class conflict rather than productive force primacy, has produced an argument.

The standard trap is the all-critique answer, which lists objections, concludes that the theory is refuted, and never reconstructs what survives. Markers penalise this because it demonstrates reading without judgment. The mirror trap is the answer that treats every objection as answered, which demonstrates the opposite.

Two evaluative moves are reliably rewarded. Naming an internal critic, since it shows the candidate knows the tradition argues with itself, and Brenner is the most useful single name available for this. And naming what would change the verdict, since it converts an opinion into a testable position. The answer structures into which these moves fit are set out in the exam answer guide for historical materialism, which owns the marks logic, and the wider survey of criticisms across the whole of Marxism sits in the complete guide to critiques of Marxism.

Teaching this as a disagreement rather than a rout

The difficulty in teaching this topic is that it can go two ways, both of them useless: a demolition in which the theory is presented as an exploded curiosity, or a defence in which every objection is disposed of. Both leave students unable to argue.

The structure that works is to assign objections to groups with the requirement that each group state the objection in its author’s terms and then produce the best available reply themselves, rather than having a rival group supply it. This prevents the exercise from becoming a debate in which winning matters, and it forces the intellectual work that the format is for.

The misconception that dominates such a class is that a criticism of a theory is a criticism of the politics associated with it, and students on both sides arrive with it. The question that surfaces it is to ask whether a person could accept every objection in the ledger and remain a socialist, and whether a person could reject them all and be a conservative. Both answers are yes, and establishing that early removes most of the heat.

The extract that resolves the substantive issue best is a paired reading of a short passage from Cohen stating the primacy thesis precisely, alongside Brenner’s summary of the eastern and western European divergence. The class can be asked whether the second is compatible with the first, and the discovery that a careful statement of a theory makes it easier rather than harder to test is worth more than any conclusion about Marx.

The objection from outside Europe

A seventh line of attack deserves separate treatment because it is now among the most frequently made and because it is often stated badly, which allows it to be dismissed badly.

The charge is that historical materialism universalises a European trajectory. Its sequence of forms was derived from western European evidence, its account of the transition to capitalism describes an English case, its treatment of societies outside Europe relies on a category, the Asiatic mode of production, that was assembled from limited and often unreliable sources and that describes stagnation where it should describe a different dynamic, and its expectation that capitalist development would sweep away pre-existing forms underestimated their persistence and their reworking under colonial conditions.

Three parts of this are strong. The Asiatic mode is a genuinely weak construction, contested inside Marxism from early on and abandoned officially in the Soviet Union in the 1930s for reasons that were political rather than scholarly, and it should not be used as though it were a settled part of the theory. The historical writing on India and China that Marx produced for the press in the 1850s contains judgments about stagnation and about the effects of colonial rule that later scholarship has not sustained, and a serious treatment says so plainly rather than explaining them away. And the expectation that capitalist relations would displace earlier forms rather than combining with them has been contradicted repeatedly, from plantation slavery within a world market to the persistence of unfree and semi-free labour inside modern supply chains.

One part of the charge is weaker than it appears. Marx himself explicitly refused to convert his account into a universal path, in the letter of 1877 and again in the drafts on the Russian commune, which means the Eurocentric universalism being attacked is more accurately attributed to the tradition after him than to the texts. This is worth stating precisely rather than as a defence: the objection lands squarely against the stage theory that dominated official Marxism for most of the twentieth century, and lands less squarely against Marx.

The most productive response within the tradition has been to treat combination rather than succession as the normal case, so that a society contains several forms of production articulated together, with the dominant form shaping but not dissolving the others. That move preserves the materialist question while abandoning the sequence, and a great deal of the best work on colonial and post-colonial economies has been done on that basis. It also concedes what the objection alleges about the original formulation.

Verdict: lands against the stage theory, partially against the texts, and has produced the most substantial repair on this list.

The pattern underneath these disputes

Something recurs across the objections that is worth naming, because it recurs across the whole field and not only here.

In case after case, what looks like a disagreement about theory turns out on inspection to be a disagreement about measurement or about the individuation of cases. Whether working-class conditions improved depends on which basket of goods is priced, over what period, for which segment of the class, and whether urban disamenities are counted. Whether the middle strata disappeared depends on how occupational categories are drawn. Whether the transition to capitalism happened once in England or several times elsewhere depends on what counts as capitalist relations, and the political Marxist definition, which requires market dependence for the reproduction of the producers, produces a different count from a definition based on wage labour and commodity production.

This has two consequences for anyone using this material. Arguments that proceed by asserting a fact are weaker than arguments that name the measurement choice driving the result, and a critic who says that wages rose will be answered, while a critic who says that on any reasonable price index the long-run direction in the industrial core was upwards, whatever happened in the first decades, will not. And the appearance of irresolvable theoretical conflict is often an artefact: the parties are counting differently and have not said so.

Stating the measurement choice is therefore the single most useful discipline in this literature, on both sides.

What remains open

Four questions in this debate are genuinely unsettled and a reader should treat any confident answer to them with suspicion.

Whether a mechanism can be supplied for the primacy thesis that is neither the discredited functional form nor a retreat to class conflict. Competitive selection between societies in contact is the best candidate and it restricts the theory’s scope severely.

Whether the Brenner account of the transition is correct. The debate it started has continued for decades, the critical literature is substantial, and the question of how much weight the English case can carry is live.

Whether law can be handled without either circularity or artificiality. Both available solutions have costs that their own defenders acknowledge.

Whether the surviving research programme is progressive or degenerating, in the sense of generating new findings or absorbing anomalies. This cannot be settled by argument at a point in time; it is a judgment about a body of work over decades, and reasonable people assessing the same literature reach different conclusions.

How to use a critique

A closing note on method, because the commonest misuse of an article like this one is to treat it as ammunition.

The purpose of stating objections at full strength is to find out what is true, which requires being willing to discover that an objection you were relying on does not work. Three of the seven treated here are weaker than their popularity suggests, and the ones that do the real damage, Elster’s and Brenner’s, are technical, come from inside the tradition, and require some effort to state. That distribution is normal. The objections that circulate most widely against any well-developed position are usually the ones that circulate because they are easy, not because they are strong.

The reciprocal point applies to defenders. A reply that disposes of the weak objections and does not engage the strong ones has not defended anything, and the tradition’s own best figures did not proceed that way. Cohen stated the theory precisely enough to be refuted and was; Brenner tested the central claim against the archive and found against it; Thompson attacked a version of Marxism that had become influential because he thought it was ruining historical writing. That is what a research programme looks like when it is working, and it is the standard against which any use of this material should be measured.

The objections have a history inside the movement

Presenting the case against as an external assault misdescribes it. Every major objection in the ledger was anticipated by someone inside the socialist movement, usually decades earlier, and knowing that history prevents a reader from treating the tradition as though it had been ambushed.

Eduard Bernstein made the first systematic version at the end of the 1890s. His argument, developed in articles and then in the book known in English as Evolutionary Socialism, was empirical before it was philosophical: the concentration of capital was not proceeding as predicted in agriculture, the number of property owners was increasing rather than shrinking, crises were becoming less rather than more severe, and the condition of workers in Germany and Britain was improving. From this he concluded that the theory’s predictive apparatus was failing, that socialism would have to be argued for on ethical grounds rather than presented as a historical outcome, and that the movement should organise for reform. His conclusion that the movement was everything and the final goal nothing was a repudiation of the framework rather than an adjustment to it.

The controversy that followed is the origin of most of the positions the tradition still occupies. Karl Kautsky’s reply defended the orthodox reading and disputed Bernstein’s statistics, and the statistical argument was more evenly balanced than either side admitted, since the categories in dispute were exactly the ones the measurement discussion above concerns. Rosa Luxemburg’s reply attacked Bernstein while refusing the fatalism he was attacking, insisting that the objective development produces the conditions for revolution without producing the revolution, and that a movement which treats the outcome as guaranteed will not build the capacity to seize it. Her formulation of the choice facing society, between a transition to socialism and a descent into barbarism, is the sharpest statement in the tradition that nothing is settled.

Werner Sombart’s question about the absence of socialism in the United States, posed in the first decade of the twentieth century, put a further empirical difficulty on the table that has never been fully answered: the most advanced capitalist economy produced the weakest socialist movement, which is precisely the inverse of the expected relation. Explanations have been offered, including the effects of immigration on the composition of the workforce, the availability of land, the structure of the political system, and the intensity of employer resistance, and each is plausible without any of them being decisive.

The failure of the German revolution after the First World War, and the survival of capitalism in western Europe, produced the next wave of internal revision, and the whole of what became known as Western Marxism can be read as an extended response to a question the theory had not expected to face: why the revolution did not happen where the conditions were supposedly ripest. Gramsci’s account of hegemony, the Frankfurt School’s turn to culture and to the analysis of authority, and Lukacs’s work on reification are all attempts to explain a failure of the mechanism the theory had relied on.

The pattern is worth stating explicitly because it recurs throughout the series. The tradition’s most substantial theoretical developments were produced by defeats, and its most damaging critics have been its own participants. A reader who wants the strongest case against historical materialism will find more of it in the socialist literature than in the anti-socialist literature, and that fact is itself evidence about what kind of intellectual enterprise this is.

Why did revolutions not happen in Britain or Germany?

The theory expected them where industry was most developed, and they came instead to agrarian societies at the periphery. The standard replies invoke uneven and combined development, the integration of labour leaderships into national institutions, the strength of parliamentary and welfare concessions, and the capacity of dominant classes to secure consent rather than rely on force.

The theory in historical practice

A final consideration belongs in any assessment and is usually left out: what the framework has produced when people have used it to write history.

The record here is stronger than the record of prediction and it is the evidence a fair verdict has to weigh. The study of enclosure and of agrarian change, the history of work and of the labour process, the analysis of colonial land settlement and revenue systems, the history of slavery within an Atlantic economy, the study of the making of working classes in a dozen countries, and much of what is now standard in the economic history of early modern Europe were produced by historians working with materialist categories. A good deal of it was produced by people who did not accept the strong version of the theory and were explicit that they did not.

The point is not that good work under a banner vindicates the banner. It is that a framework’s value is partly a matter of what it makes visible, and this one has reliably directed attention to questions that were otherwise not asked: who worked, on what terms, under what compulsion, and who took what they produced. Those questions were not being asked systematically before, and much of what is now common ground in social and economic history entered the field through people asking them.

The corresponding weakness is equally visible in the record. Where the framework has been applied as a template rather than as a set of questions, it has produced history in which the archive illustrates a conclusion reached in advance, which is exactly what Thompson objected to. The distinction between the two uses is not subtle and any reader can make it: ask whether the account could have come out differently.

Testing the theory on one case

A critique is more useful when it shows the theory being tested rather than discussed, and the abolition of Atlantic slavery is the best available case because the materialist and the idealist explanations were both stated precisely by serious historians and the evidence has been fought over for decades.

Eric Williams argued in Capitalism and Slavery, published in 1944, that the profits of the slave trade and of the plantation colonies contributed to British industrialisation, and that abolition came not from a change of heart but from a change of interest: once the West Indian sugar economy had declined in relative importance and industrial capital had different requirements, the political coalition sustaining the system dissolved. The argument is materialist in the strict sense. It explains a moral transformation by a shift in the structure of economic interests, and it makes the humanitarian campaign the form in which the shift was expressed rather than its cause.

Seymour Drescher’s response, in the book published under the title Econocide in 1977, attacked the decline thesis directly on the evidence, arguing that the British slave system was economically buoyant rather than failing at the moment abolition was carried, and that Britain destroyed a profitable branch of its economy. If that is right, the materialist explanation of the timing fails, and the political mobilisation, the petition campaigns, the religious networks, and the parliamentary struggle have to carry more explanatory weight than the Williams account allows.

The debate has not closed, and the current position is instructive for this article’s purposes. The contribution of slave-trade profits to industrial finance is now generally regarded as real but smaller than Williams proposed, and the estimates depend heavily on how contribution is defined and measured, which is the measurement pattern described above appearing again. The decline thesis has fared worse than the finance thesis. Meanwhile the strongest recent work has tended to combine the accounts rather than choose between them, explaining the timing by the interaction of a mass political movement, the strategic position of the colonies during a period of war, resistance and revolt in the colonies themselves, and the changing structure of British commercial interest.

Two lessons follow for the assessment of historical materialism. The theory was testable on this case, in the strong sense that a specific claim about timing and interest was made, checked against evidence, and substantially defeated. That is a point for the theory’s testability and against the unfalsifiability charge in its crude form. And the surviving materialist contribution is the weaker one: not that interests explain the outcome, but that no account of the outcome is complete without them, since the campaign succeeded where and when it did partly because of what it cost whom. That is the lower rung of the ladder doing real work, which is what this article’s verdict predicts.

The case is also a useful corrective to the assumption that materialist and idealist explanations are exhaustive. The role of resistance and revolt by enslaved people themselves, which both the original accounts underweighted, is neither a matter of British economic interest nor of British moral opinion, and its addition to the explanation came from historians working in neither of the original camps.

Where this leaves a reader

The position a reader should be able to occupy after working through this material is narrower than either side’s public statements and more defensible than both.

Historical materialism made large claims, several of which are false and have been shown to be false, in some cases by its own adherents. The strong version is not available to anyone who has read the objections. Defending it requires either ignoring the record or performing the sequence of modifications that Popper identified and that a defender should at least acknowledge as a pattern.

What is available is a set of questions and a set of commitments about what an explanation must include, and these have a genuine record in historical practice. They are not the property of Marxism in the sense that using them requires accepting anything else in the tradition, and a great deal of the historical work that uses them is done by people with no political attachment to it.

The unresolved item is the one to watch. Until someone specifies a mechanism connecting the organisation of production to institutional outcomes that is precise enough to generate expectations and general enough to travel across cases, the theory will continue to be strongest as a guide to what to investigate and weakest as an account of why things had to happen as they did. That is a smaller claim than the tradition has usually made for itself, and it is the claim the evidence supports.

Is Engels responsible for the version being attacked?

A recurring move in these disputes is to concede the objections and assign the position they refute to Engels rather than to Marx, and the move deserves examination because it is both partly right and frequently abused.

The case for it is real. Engels wrote the popular expositions through which the theory reached the movement, and those expositions are more systematic, more confident, and more schematic than anything Marx published. He extended the dialectic to nature in work that Marx did not see in finished form, he supervised the editing of the second and third volumes of Capital from manuscripts their author had left incomplete, and he presided over the codification of the theory in the decade after Marx’s death. A tradition that received historical materialism through his summaries received something tidier than the texts warrant.

The case against it is equally real. Engels was the one who explicitly rejected the crude version, in the correspondence of the 1890s, and diagnosed how it had arisen, admitting that polemical emphasis on the economic side had encouraged it. If he is to be blamed for the schematic version, he must also be credited with the most authoritative statement of the qualified one. The two men collaborated for four decades and Marx read and approved a great deal of what Engels published in his lifetime, including the popular exposition of scientific socialism, which makes a sharp separation between them difficult to sustain on the evidence.

Did Engels distort Marx’s theory of history?

He systematised it, and the systematisation was tidier than the texts support. But he also rejected the vulgar determinist reading in terms, and Marx read and approved much of what he published. Assigning the discredited version to Engels alone is a convenience rather than a finding.

The practical rule is to treat the Engels question as a matter for evidence rather than as a device. Where a specific formulation is Engels’s and not Marx’s, say so and cite it. Where the argument requires that Marx would have rejected it, produce the passage in which he does. What is not acceptable, and what the move usually amounts to, is protecting a favoured author by attributing every difficulty to a collaborator who cannot answer. The profile of Engels as an independent thinker works through what he contributed and what he altered.

Separating the theory of history from the rest of Marxism

One more distinction prevents a great deal of wasted argument, and it belongs at the end because readers who have followed the objections will be tempted to draw conclusions wider than the evidence supports.

The theory of history is one component among several. The analysis of value and exploitation is a separate body of argument with its own evidence and its own critics, and a reader can reject productive force primacy while holding that the account of surplus extraction is sound, or the reverse. The analysis of capitalist dynamics, including accumulation, concentration, and crisis, stands or falls on economic evidence rather than on the philosophy of history. The theory of ideology, the account of the state, and the political strategy each have their own literatures. So does the ethical critique, which several critics of the theory of history have found compelling precisely because it does not depend on any claim about historical inevitability. G. A. Cohen is the clearest illustration available: having written the most rigorous defence of the theory of history ever produced, he spent his later career on egalitarian political philosophy, on the ground that the case for equality had to be made on its own terms once the historical guarantee had gone.

The objections in this article target one component. They do not transfer automatically, and a writer who moves from the failure of the primacy thesis to a general verdict on Marxism has performed an inference that nothing here supports. The reverse error is equally common: defending the analysis of exploitation and treating that as a defence of the theory of history.

The reason this matters practically is that the components have different records. The prediction record of the theory of history is poor. The record of the analysis of concentration and of the tendency of capital to expand its sphere of operation is considerably better. Treating Marxism as a single object that succeeds or fails as a whole guarantees a verdict that is wrong about most of its parts, and it is a habit shared by hostile and sympathetic writers in almost equal measure, which is one reason the public argument about it has been so unproductive for so long, and the survey of the objections to each component is set out in the complete guide to critiques of Marxism.

Frequently Asked Questions

Q: What is the strongest criticism of historical materialism?

The objection that functional explanation without a selection mechanism is not explanation, made by Jon Elster in Making Sense of Marx against G. A. Cohen’s defence of the theory. It is the strongest because it targets the rung where the theory’s best defenders stand rather than the crude version. Cohen’s primacy thesis holds that relations of production obtain because they are the relations that best develop the productive forces, which explains the cause by its effect. In biology such explanations are licensed by natural selection, a well-understood mechanism operating through differential reproduction over many generations with random variation. Nothing in the social world corresponds to it: societies do not reproduce differentially in the required sense, variation in social arrangements is not random, and the number of cases is far too small. The best available answer, competitive pressure between societies in contact, supplies something like selection at the price of restricting the theory’s scope considerably.

Q: What would falsify historical materialism?

The strong version was falsifiable and was falsified: it predicted revolution first in the most industrially advanced countries, deteriorating conditions for workers under capitalism, and the squeezing out of intermediate strata, and none of these occurred as stated. The harder question concerns the surviving version, and it has an answer if anyone is willing to state it. Three findings would count against it. Durable institutional orders persisting with no material basis, so that an arrangement is sustained by belief alone with nobody bearing costs or capturing benefits. Arrangements that systematically block productive development surviving indefinitely in competitive contact with more productive rivals. And large-scale transformations in how societies are organised occurring with no relation to conflicts over the production and appropriation of surplus. Nothing in the current historical literature establishes any of these, which is why a theory whose strong version failed remains in use.

Q: Why did revolutions happen in poor countries instead of rich ones?

The expectation that the most developed country shows the less developed the image of its future is stated in Marx’s own preface to the first volume of Capital, and the revolutions came instead to agrarian economies at the periphery of the world market. The standard reply is uneven and combined development, which holds that a society’s revolutionary potential depends on its position within a world economy rather than on its internal level of industry, so that the chain breaks at its weakest link where advanced and backward forms of production are violently combined. The reply has a genuine virtue critics rarely concede: it was formulated before the events rather than as a rescue afterwards. Its weakness is that it converts a prediction about development into a prediction about combination, and the record of that second prediction is mixed rather than clean.

Q: Did Marx’s prediction about worker poverty come true?

Not in the industrial core, and the textual question of what he predicted is genuinely disputed. The general law of capitalist accumulation in the twenty-fifth chapter of the first volume of Capital has been read as claiming absolute immiseration and as claiming relative immiseration for over a century. The relative reading has substantial support, since Marx allows elsewhere that real wages can rise while labour’s share falls and the intensity of work increases, and the mechanism he specifies, the reserve army of labour, produces downward pressure rather than guaranteed decline. Measurement disputes about the early industrial period are real: the timing, the pricing of goods whose quality changed, the treatment of urban disamenities, and the distribution across the class are all contested. What is not contested is the long-run direction in the industrial core, and the defensible objection concerns what the movement expected and organised for rather than the exegesis of one chapter.

Q: Does nationalism disprove historical materialism?

It refutes the strong version of the mechanism by which classes were supposed to become historical agents. The theory holds that relations of production distribute people into groups with opposed interests and that these groups act; the record shows national and religious identity repeatedly overriding class alignment, most starkly when the socialist parties of Europe voted for war credits in 1914 against their own international commitments. Two replies have force. The tradition never held that class identity forms automatically, and the whole line from Lenin’s insistence on organisation to Gramsci’s account of building a collective will concedes that it does not. And nationalism is itself historically produced with material conditions of existence, so explaining its strength is a task for the theory. The second reply concedes something important: if the theory must explain why class loses, it can no longer treat class as the default that other factors merely delay.

Q: What is the microfoundations objection?

It is Jon Elster’s requirement that any claim about a social structure must in principle be cashable into claims about the beliefs, desires, and actions of individuals and the interactions between them. Without such a story, an explanation names a relation without saying how it is produced, and what looks like a causal claim is a correlation with an arrow attached by assertion. Elster developed it against functional explanation in Marxism, arguing that the biological analogy fails because nothing in the social world plays the role natural selection plays. The requirement is not anti-Marxist in intention; the analytical Marxist current he belonged to was an attempt to rebuild the theory on foundations that would satisfy it. Its most interesting feature is that E. P. Thompson arrived at a similar demand from the opposite direction, insisting that a structural model must say how a position becomes a disposition and a disposition an action.

Q: Is functional explanation a valid Marxist argument?

It is valid in principle and unsupported in this case, which is the distinction that matters. Functional explanations are respectable wherever a mechanism connects the beneficial effect to the persistence of the thing that produces it, as natural selection does in biology, and G. A. Cohen argued explicitly that a well-supported consequence law can license such an explanation even before the mechanism is identified. The difficulty is that no social analogue of selection has been established. Societies do not reproduce differentially in the required way, variation in institutional arrangements is not random, and the sample is small. Competitive pressure between societies in contact is the best candidate and it works only where such contact exists and bites, which restricts the claim to societies in competition rather than to history as such. The honest position is that the functional form of the primacy thesis remains without a defence.

Q: How do Marxists reply to these criticisms?

The strongest reply notices what the objections have in common. Each either targets the strong version that the tradition’s own best work had already abandoned, or converts into a constraint on how the theory should be used, and none offers a better account of the question the theory exists to answer, which is why the institutional order governing production and appropriation changes at all. Frameworks are displaced by better frameworks rather than by lists of difficulties, and no rival account of long-run institutional change has established itself. The framework continues to generate research in economic history, labour history, agrarian studies, and the analysis of colonial settlement that other frameworks did not produce. Two things can be said against this reply: the claim that no better framework exists is only as good as the survey behind it, and a degenerating research programme also produces internal critics, often in greater numbers than a healthy one.

Q: Why do political Marxists reject productive force primacy?

Because the evidence from the transition to capitalism runs the other way. Robert Brenner’s essay in Past and Present showed that identical pressures produced opposite institutional outcomes across Europe: the same fourteenth-century demographic collapse was followed by the erosion of serfdom in the west and its intensification in the east, and the same expansion of trade coincided with capitalist agriculture in England and a secure peasantry with an office-holding nobility in France. If the same cause produces opposite effects, the cause is not doing the explanatory work. What does the work, on this account, is the outcome of class conflict over rent and tenure, which was not settled in advance. The general conclusion drawn by Brenner and by Ellen Meiksins Wood is that there is no transhistorical tendency for productive forces to develop, because the compulsion to raise productivity is itself a historically specific feature of capitalist property relations.

Q: What parts of historical materialism survive the criticism?

A research programme rather than a science of history, and its content is substantive. That the organisation of production and appropriation is a primary object of historical explanation rather than a background condition. That durable institutional arrangements require material reproduction, so an account that cannot say who bears the costs and who captures the benefits is incomplete. That conflict over the terms on which surplus is produced and appropriated is a recurrent and frequently decisive source of institutional change. And that the categories through which an economy is understood are historically specific rather than natural. Each is a real commitment that could be wrong, and each has generated historical work other frameworks did not produce. What does not survive is the universal sequence of stages, the functional form of the primacy thesis, the prediction record, and the assumption that class will reliably become the operative political identity.

Q: Is criticising historical materialism the same as rejecting Marxism?

No, and the conflation weakens arguments on both sides. Marxism contains a theory of history, a theory of value and exploitation, an analysis of capitalist dynamics and crisis, a theory of ideology, a body of political strategy, and an ethical critique, and these are logically separable. A reader can hold that the primacy thesis fails while accepting the analysis of exploitation, or accept the theory of history while rejecting the value theory. Several of the critics in this article did exactly that: Brenner, Wood, and Thompson were Marxists attacking specific claims, and the analytical Marxists who pressed the microfoundations objection considered themselves to be repairing the tradition. Rejecting a framework is also not required by criticising it. A reader can hold that no adequate theory of long-run institutional change has yet been established, which is a defensible position and commits them to nothing else.

Q: What is the Brenner debate about?

It is the exchange that followed Robert Brenner’s essay on agrarian class structure and economic development in pre-industrial Europe, published in Past and Present, and continued in that journal and elsewhere for years. Brenner attacked the two dominant explanations of European economic divergence, the demographic model and the commercialisation model, on the ground that identical causes produced opposite institutional outcomes in different regions. He argued that the decisive variable was the outcome of class conflict over rent and tenure, which produced in England a structure of landlords, capitalist tenants, and wage labourers all subject to competitive pressure, and in France a peasantry with secure possession and no such pressure. The debate that followed involved historians of several countries disputing the evidence, the periodisation, and how much weight the English case can bear. It remains unresolved, and it is the single most important internal challenge to productive force primacy.

Q: Does the collapse of the Soviet Union refute historical materialism?

No, and treating it as evidence either way confuses a theory with a regime. Nothing in historical materialism predicts that a state describing itself as socialist will endure, and the failure of such a state is evidence about that state, its economic organisation, and its politics. The converse also holds: nothing in the theory is vindicated by any regime’s achievements in literacy, industrialisation, or life expectancy. The historical record of the socialist states, including famine, repression, and mass death, is a serious subject with its own evidence and its own scholarship, and it should be reported accurately and without euphemism, but it belongs to the assessment of those states rather than to the assessment of a claim about what explains long-run institutional change. Keeping the analytical question separate from the moral and political one is a requirement for writing seriously about either.

Q: Is historical materialism Eurocentric?

The charge lands squarely against the stage theory that dominated official Marxism and less squarely against the texts. Three parts are strong: the Asiatic mode of production is a weak construction assembled from limited sources, contested inside Marxism early and abandoned in the Soviet Union for political rather than scholarly reasons; the journalism Marx produced on India and China in the 1850s contains judgments about stagnation and about colonial rule that later scholarship has not sustained; and the expectation that capitalist relations would displace earlier forms rather than combining with them has been contradicted repeatedly, from plantation slavery inside a world market to unfree labour in modern supply chains. The weaker part is the attribution of universalism to Marx himself, since he explicitly refused to convert his account into a path prescribed for all peoples. The most productive repair treats combination rather than succession as the normal case.

Q: Does the constitutive role of law make historical materialism circular?

It is the sharpest structural difficulty the model faces. Relations of production are relations of effective control over resources and labour, and in any developed society those relations are constituted by legal categories: ownership, tenure, contract, employment, corporate personality. If law constitutes the base and also belongs to the superstructure the base explains, the explanation runs in a circle. Two solutions exist and each costs something. Cohen’s is to describe the base in terms of powers rather than rights, so it can be specified without legal vocabulary, with law then explained as the form in which powers are stabilised and enforced; this is logically clean and descriptively artificial, since nobody experiences control over a factory apart from the title to it. The alternative treats law as internal to the relations of production, which matches how historians describe property regimes and substantially weakens the model. The problem is unresolved.

Q: How do you criticise historical materialism without attacking a straw man?

Name the rung you are attacking. The version most often demolished is the universal five-stage sequence, which is not Marx’s and was formalised much later, most influentially in a Soviet statement of the 1930s. Attacking it establishes nothing about the positions people actually defend. Check the Popper attribution before using it, since the claim that he showed Marxism to be unfalsifiable inverts his argument, which was that the theory was falsifiable, was falsified, and was then rescued by its defenders. Do not use a state’s record to settle a theoretical question. Handle the immiseration claim carefully, because the textual position is disputed and the defensible objection concerns what the movement expected rather than what one chapter means. And use the internal critics, since Brenner, Elster, and Thompson supply objections far stronger than anything in the popular literature.

Q: Did E. P. Thompson reject Marxism?

No. He attacked a version of it, and he did so as a participant. The Poverty of Theory is a polemic against Althusserian structuralism, which he regarded as a closed system in which the terms defined each other and evidence could not disturb them, and behind the polemic is a claim about what a theory of history must be able to do. His positive statement, in the preface to The Making of the English Working Class, is that class is a relationship rather than a thing, that it happens when people articulate the identity of their interests out of common experience, and that it cannot be found by taking a structural cross-section at a moment. That is a demand for an account of how a shared situation becomes a shared identity, and it converges with Elster’s microfoundations requirement despite the mutual hostility between their intellectual styles. Thompson insisted throughout that he was working inside the tradition.

Q: What should a teacher watch out for when covering criticisms of Marxism?

The assumption, held by students on both sides, that a criticism of the theory is a criticism of the politics associated with it. The question that surfaces it quickly is whether a person could accept every objection in the ledger and remain a socialist, and whether a person could reject all of them and be a conservative; both answers are yes, and establishing that early removes most of the heat. The second thing to prevent is the class becoming either a demolition or a defence, both of which leave students unable to argue. Assigning objections to groups with the requirement that each group produce the best reply itself, rather than having a rival group supply it, forces the intellectual work the format exists for. The most productive extract is a precise statement of the primacy thesis paired with the account of the eastern and western European divergence.