Analytical Marxism attracts criticism from two directions that rarely meet, and a reader who hears only one of them receives half an education. From outside, classical Marxists object that the school’s modelling habits smuggled foreign assumptions into Marxist questions, so that its negative conclusions were purchased in advance by its premises. From inside, a harder charge circulates among the school’s own members and sympathisers: the current demanded clarity, argument and testability from everyone else, yet the defence of functional explanation in its founding text was fragile by those very tests, its central methodological slogan was repeated more often than it was established, and its scientific self image coexisted with a strikingly thin record of empirical inquiry. Both directions deserve a hearing before any verdict is reached, because each exposes weaknesses the other cannot see. The external critics notice what the models leave out. The internal critics notice what the school failed to demand of itself. An assessment that stages only the first becomes polemic, and one that stages only the second becomes an in house quarrel detached from the tradition the school claimed to serve. This article keeps both in view, states what the school got right at full length, and refuses to settle the one question its partisans on either side most want settled.

A seminar table holding the criticisms ledger beside the founding texts of analytical Marxism, Insight Crunch

The school itself, its membership and its constructive achievements are treated in the companion guide to analytical Marxism complete guide, and readers who need the narrative of the September Group, the founding texts and the sequence of abandonments will find it there. This article presupposes that material without rehearsing it, because its business is different: not what the school built, but where the building failed inspection, and by whose measuring instruments. The comparison with the classical tradition it broke from is developed in analytical vs classical Marxism, and the present assessment borrows that comparison only where a criticism cannot be stated without it. Keeping the territories separate matters for fairness. A criticism lands differently when its target is a method, a model, a conclusion or a career, and much of the published quarrelling around this school consists of parties answering charges that were never quite the charges made. The ledger format adopted below exists to prevent that slippage, by fixing each objection to its source, its best reply and a verdict that carries its own condition of defeat.

A word about grading standards is needed at the outset, because the whole article turns on it. Every intellectual current announces, implicitly or explicitly, the terms on which it wishes to be judged. A prophetic tradition asks to be judged by fidelity and fruitfulness across generations. A scientific programme asks to be judged by the accuracy and reach of its explanations and predictions. Analytical Marxism asked, loudly and repeatedly, to be judged by clarity of statement, validity of argument, specification of mechanisms and openness to evidence. Those requests were honourable, and they transformed the discussion of Marx for the better. They also created a liability that the school’s founders perhaps underestimated: a current that insists on explicit standards hands its critics a ready made instrument, because its own work can be measured against the published criteria without any appeal to alien values. The internal inconsistency named in this article is exactly that measurement performed. The school’s opening theoretical move, G. A. Cohen’s defence of functional explanation in historical materialism, was subjected within four years to a dismantling by Jon Elster, a member of the same circle, using nothing more exotic than the standards the circle had jointly proclaimed. That episode is not an embarrassment to be hurried past. It is the most instructive event in the school’s history, and it organises everything that follows here.

Fairness also requires registering the asymmetry of the surviving record. The school’s members wrote their criticisms of one another in the open, in books and journals that any reader can consult, whereas the classical tradition’s internal doubts were often expressed in correspondence, in factional documents or not at all. A current that publishes its self criticism will always look more troubled than a tradition that keeps its troubles private, and an honest ledger corrects for that optical effect rather than exploiting it. Several of the sharpest points in the pages below were first made by the school’s own people, a fact that hostile summaries tend to omit because it complicates the portrait of a complacent clique. The school was many things, but intellectually complacent it was not. Its failure, where failure is the right word, was of a subtler kind: it applied its instruments unevenly, rigorously to inherited doctrine and intermittently to its own founding commitments, and it mistook the stating of a methodological creed for the arguing of it.

The plan of the article follows the two directions in turn. The classical objections come first, because they are older, more numerous and more often quoted: the charge against methodological individualism, the charge that exchange modelling imports voluntariness into the wage contract, the complaint that the reformulated concept of exploitation cannot separate domination from inequality, the complaint that class became a possession rather than a relation, the silence surrounding the labour process, and the claim that the normative turn left Marxism behind while presenting itself as its completion. Then the internal objections are stated in their strongest form, since they are the article’s centre of gravity. A middle section records what the school got right, without grudging brevity, because an assessment that cannot name the achievements cannot be trusted on the failures. The closing sections take up the trajectory question, the pattern of abandoned theses that both camps cite as proof, and explain why that pattern proves nothing for either side. Throughout, every criticism is attached to a named source, every verdict states what would overturn it, and comparisons between the school and the wider tradition are drawn across the same decades and on declared dimensions, never between a small group’s output and a century and a half of accumulated writing.

Two directions of criticism, and why the second cuts deeper

Criticism arriving from an opposing camp can always be discounted by the target, and usually is. When classical Marxists attacked the school’s models, the school’s members could reply, with some justice, that their attackers disliked the conclusions and therefore quarrelled with the tools. That defensive move is unavailable against criticism arising from shared premises. When a member of the September Group argued that the group’s founding text failed tests the group itself had set, no question of camp loyalty could soften the blow, and none was attempted: Cohen revised, qualified and in part conceded, in print, over the following decade. The episode illustrates a general truth about intellectual movements that this article uses as its organising device. External criticism tests whether a school’s conclusions offend its rivals. Internal criticism tests whether a school’s practice honours its own proclaimed method. The second test is the more revealing one, and it is the test on which analytical Marxism presents the most paradoxical profile in modern Marxist history: exemplary in its willingness to publish internal dissent, uneven in the standards it applied to its own foundations while that dissent was being voiced.

What did analytical Marxism get wrong?

classical critics argue its models assumed voluntary exchange between agents with given preferences, which builds in non-Marxist conclusions. By its own standards, its founding defence of functional explanation was weak, its central methodological claim was asserted rather than argued, and it produced little empirical research.

That compressed answer conceals a distinction the rest of the article exists to draw. The classical half of the charge is about substance smuggled in through form: preferences treated as given, exchange treated as the elementary social act, assets treated as the substance of class. The internal half is about procedure: arguments not supplied where the school’s creed required them, explanations defended by a logic the school elsewhere rejected, tests demanded of others and seldom performed at home. A reader who merges the two halves will misjudge both. The substantive charges can be contested, and the school’s defenders contested them vigorously, sometimes persuasively. The procedural charges are harder to escape, because escaping them requires showing that the founding texts met standards whose content the school itself had fixed. Where the school’s people attempted that defence, as Cohen did in his later reflections on functional explanation, the result was a partial retreat conducted with exemplary candour, which is itself evidence about the school’s character: its standards were real enough to wound their own authors.

Why does grading the school by its own standards matter?

Grading by shared standards removes the easiest escape available to any criticised school, the plea that its judges measure by alien values. When the criteria are the school’s own published demands for clarity, mechanisms and evidence, a failure cannot be dismissed as a misunderstanding between traditions. The verdict then concerns consistency, and consistency is the one virtue no research programme may decline.

The device also disciplines the critic. Anyone may dislike rational choice modelling on instinct, or regret the passing of dialectical prose on aesthetic grounds, and such reactions, however sincere, settle nothing. Requiring the critic to work with the school’s own criteria forces engagement with the strongest version of its project, in the same way that the school’s founders required themselves to engage with the strongest version of Marx. There is a pleasing symmetry in turning the founders’ method upon the founders, and the school’s best members recognised the symmetry when it was applied to them. Elster’s internal critique of Cohen was received not as treachery but as the programme working as advertised, a reception that speaks well of the circle’s norms even as the critique itself speaks badly of its founding text. Both facts belong in the ledger, and an assessment that records only the second has misunderstood the first.

A further reason concerns the audience. Students meeting this material in examinations are routinely asked to evaluate the school, and evaluation questions reward exactly the discrimination this article practises: which criticisms are external and contestable, which are internal and therefore binding, and which achievements survive both audits. A student who can say that the exchange paradigm objection depends on premises a classical Marxist holds and an analytical Marxist need not, while the functional explanation objection depends on premises every analytical Marxist professed, has learned something transferable about how arguments bind. The distinction generalises far beyond Marxism. Any tradition that imports a formal apparatus, whether game theory into political science or optimisation into ecology, faces the paired questions staged here: does the apparatus import conclusions, and does the importer obey its own rules? The analytical Marxist episode remains the clearest case study of both questions arising together, partly because the participants were unusually explicit about their rules and unusually willing to record their own breaches.

What separates a criticism from outside from one raised within?

An external criticism disputes a school’s premises, conclusions or choice of problems, and can be answered by defending those choices. An internal criticism accepts the premises and shows the practice departing from them. The first kind invites debate between traditions. The second kind, when it succeeds, leaves the school no reply except repair, concession or silence about its own standards.

With the distinction fixed, the classical charges can be given their full force without pretending they are knock down blows. Each of them identifies a real feature of the school’s work. The models did begin from individuals. The exchange framework did place endowed agents at the centre. The reformulated exploitation concept did travel by way of asset distributions. The labour process did recede from view. The normative writings did converge with liberal egalitarianism. None of these observations is fabricated, and the school’s defenders rarely denied the observations themselves. What they denied was the inference: that beginning from individuals begs the question against relational explanation, that modelling exchange presumes voluntariness in the morally loaded sense, that an asset based concept of exploitation loses what matters, that class as endowment cannot capture class as relation, that the labour process omission reflects method rather than division of labour, and that convergence with liberal conclusions reveals a departure rather than a discovery. The sections that follow take the charges one at a time, attribute them, state the best replies, and grade the exchange. Readers should notice how often the honest grade is partial: a charge that identifies a genuine cost, met by a reply that identifies a genuine gain, with the balance depending on judgements the ledger can clarify but not make on the reader’s behalf.

One preliminary caution applies to the whole classical dossier. The tradition lodging these charges was itself divided about the standards they presuppose. Classical Marxism contained its own individualists and its own holists, its own celebrants of the labour process and its own theorists for whom production relations mattered more than the experience of work, its own moralists and its own stern opponents of moralising. When a critic speaks in the name of the classical view, the phrase usually selects one strand of a quarrelsome inheritance and presents it as the whole. That selectivity does not invalidate the charges, since a strand can be right against the whole, but it does mean the charges arrive as arguments to be weighed rather than as a tradition’s unanimous testimony to be recorded. The school’s members made this point in reply more than once, and it stands. What does not stand, and what the replies could never dissolve, is the narrower observation that certain modelling choices have consequences whether or not their makers intend them, and that a school proud of making assumptions explicit was oddly reluctant to trace the consequences of its own.

The classical charge against methodological individualism

The oldest and most fundamental classical objection concerns the starting point. Analytical Marxist models characteristically begin with individuals: agents endowed with preferences, beliefs, assets and capacities, whose interactions aggregate into social outcomes. To classical critics, that starting point is not a neutral modelling convenience. It is the very proposition Marx’s mature work contests. On the classical reading, persons are formed within relations they did not choose; the worker and the capitalist confront one another as bearers of positions defined by a structure that precedes both; preferences themselves, including the preference for security over risk or leisure over income, bear the marks of the social order in which they were formed. A method that takes the individual and the preference schedule as given therefore assumes, at the first step, a picture of society as an aggregate of pre social atoms, and every conclusion drawn from the model inherits that assumption. Critics writing in this vein, among them Alex Callinicos in his extended engagement with the school’s historians and philosophers, argued that the school’s celebrated rigour was rigour within a frame, and that the frame decided in advance which Marxist claims could even be formulated, let alone confirmed.

The charge gains precision when stated against particular texts. Elster’s reconstruction of Marx proceeded by sorting Marx’s claims into those translatable into individual level mechanisms and those dismissed as holistic residue, and the sorting criterion was never itself neutral between the traditions. Claims about class formation, ideology and historical direction that resist translation were not refuted by the procedure; they were disqualified by it, which classical critics regarded as a quieter and less honest operation than refutation. A refutation meets a claim on its own ground and shows it false. A disqualification changes the ground so that the claim cannot be stated, then reports that nothing of value was lost in translation. The school’s defenders answered that untranslatable claims were usually untranslatable because they were unclear, and that the burden lay on the classical side to produce one clear, true, irreducibly holistic explanation. That challenge was fair, and it was rarely met with the rigour it demanded. Yet the challenge does not wholly dissolve the objection, because translatability is itself a matter of degree and patience: a mechanism sketch that a sympathetic modeller could supply for a holistic claim counts as translation available, and the school’s judgements about which claims lacked mechanisms were sometimes delivered with a speed that suggested the verdict preceded the search.

What does methodological individualism assume about persons?

It assumes that explanation should terminate in individuals, their preferences, beliefs and endowments, treated as given at the point where analysis begins. Classical critics reply that persons and preferences are themselves products of social relations, so the assumption installs as a premise the atomistic picture of society that Marx’s relational account exists to deny.

The school’s own ranks were divided on how strong a doctrine methodological individualism should be, and the division matters for grading the charge. Andrew Levine, Erik Olin Wright and Elliott Sober, writing jointly, defended a version modest enough to permit institutions, relations and structures as genuine explanatory furniture, provided their operation could in principle be traced through the actions of situated individuals. Elster’s version was more austere, and his austerity set the tone that classical critics attacked. A charge aimed at the austere version lands; aimed at the modest version, it mostly misses, because the modest version asserts a constraint on explanation rather than a metaphysics of the person. Much of the published quarrel proceeded at cross purposes for precisely this reason, with critics refuting Elster and defenders sheltering behind Levine, Sober and Wright, sometimes within the same exchange. The ledger therefore records the objection as partly answered: answered against the modest doctrine the school could retreat to, standing against the austere doctrine several of its prominent members actually practised in their reconstructive work.

There remains a subtler layer that neither version fully addresses. Even a modest individualism must decide where to stop the regress of formation. If preferences are admitted to be socially formed, the modeller who takes them as given for one inquiry owes an account, somewhere in the research programme, of their formation, or else the givenness hardens from a local convenience into a global habit. The school’s record on that debt is thin. Its models of class, exploitation and collective action take wants and beliefs as inputs with remarkable consistency, while its promised accounts of preference formation remained programmatic. Ideology, which in the classical tradition named exactly the social formation of belief, was the school’s least developed topic relative to its importance, a gap Elster himself acknowledged in his later reflections on his Marx book. A criticism that began as metaphysics thus ends as bookkeeping: the school borrowed against future work on formation that largely never arrived, and the borrowing was invisible because each individual model was permitted to treat its inputs as somebody else’s problem.

Set against that, the defence deserves its due. Treating agents as given is what every tractable model in every science does somewhere, and the classical alternative had its own unexamined givens, above all the class subject whose interests and consciousness were read off from structural position with a confidence no survey ever quite vindicated. The school’s insistence that interests do not announce themselves, that workers may rationally decline the collective action their position supposedly dictates, and that any theory of class formation must pass through the reasoning of actual persons, permanently raised the standard for Marxist explanation. The classical charge identifies the cost of the school’s starting point correctly. It errs when it implies the rival starting point was costless, or when it treats a modelling decision as a confession of political faith. Graded as an objection to specific reconstructive practices and to an unpaid debt on preference formation, the charge stands in part. Graded as a refutation of the school’s entire method, it overreaches, and the school’s replies, at their careful best, showed exactly where the overreach begins.

Exchange, endowments and the voluntariness built into the frame

The second classical charge follows from the first but targets a specific modelling idiom rather than a general doctrine. The school’s characteristic constructions, above all John Roemer’s, represent an economy as a population of agents carrying endowments of assets and labour power into exchanges whose outcomes can then be interrogated for exploitation, class and inequality. The full reconstruction is treated in Roemer theory of exploitation, and only the critical point belongs here. Classical critics objected that the idiom itself performs ideological work. Exchange between endowed agents is the grammar of the market viewed from inside, and a theory that adopts the grammar inherits its most contestable suggestion: that the transactions modelled are voluntary meetings of independent parties, differing only in what they happen to own. On that reading, the wage contract enters the model as one exchange among others, and the classical claim that it is a compelled transaction, entered under the duress of having nothing else to sell, must be re introduced as an afterthought about background conditions rather than appearing as the phenomenon to be explained. The model does not deny compulsion. It relocates compulsion to the scenery, where analytical attention seldom lingers.

Roemer’s defenders replied, accurately, that his framework was designed precisely to interrogate background conditions: the distribution of assets is the variable his definitions manipulate, and his celebrated counterfactuals ask what would happen if agents withdrew with per capita shares of alienable wealth. Compulsion, on this reply, is not scenery but the subject, since the whole apparatus exists to show which inequalities of endowment generate which patterns of transfer. The reply is strong against the crude version of the charge, in which the school is accused of celebrating market exchange. Roemer was a socialist critic of property arrangements, not an apologist for them, and his models were built to convict distributions, not to bless transactions. The subtler version of the charge survives the reply, however. It observes that in the exchange frame, the social relation between capitalist and worker appears only as a consequence of differing portfolios, never as a constitutive relation that shapes what the agents are. Two agents with identical preferences and different assets exchange; reverse the assets and the positions reverse with them, cleanly, without residue. Classical Marxism’s claim is that the relation is stickier than that: that occupying the position of buyer of labour power forms interests, information, habits and power that no redistribution of portfolios within the model quite captures. Whether that claim is true is contestable. That the model cannot state it is not.

Why do classical critics call the exchange model question begging?

Because the model represents the wage relation as an exchange between independent owners before any argument has been given, so the feature classical theory most wants explained, the compelled character of selling labour power, is settled by the representation rather than investigated by it. A frame that presumes the answer is accused of begging the question.

The verdict recorded in the ledger for this charge is therefore divided, as the division in the secondary literature suggests it should be. As a claim about the school’s intentions or political sympathies, the voluntariness charge fails; the documentary record of socialist commitment is unambiguous. As a claim about expressive limits, it succeeds in part: the exchange idiom purchases its celebrated clarity about distribution at the price of muteness about constitution, and the school’s writing seldom acknowledged the price as a price. There is a recognisable pattern here that recurs across the classical dossier. Again and again, the school’s reply demonstrates that a crude charge is false, and the demonstration is allowed to stand in for a response to the refined charge, which is true in a narrower and more damaging way. A school whose founding virtue was distinguishing versions of claims, strong from weak, charitable from uncharitable, might have been expected to insist on the distinction when the claims were aimed at itself. That it so often answered the weak version is a small but genuine entry on the internal side of the ledger, where the charge is not bad modelling but uneven self application of its own hermeneutic discipline.

A related complaint, pressed by critics in the tradition of Ellen Meiksins Wood, concerned what the exchange frame does to politics. When class outcomes are functions of asset distributions, political action appears in the model as a device for redistributing portfolios, and the classical emphasis on transforming the relations within which production occurs fades into a preference for one distribution over another. Wood’s broader argument against the retreat from class, directed at several currents of the period, held that such frameworks translate a theory of social transformation into a theory of distributive adjustment, and that the translation loses the specifically Marxist claim that the wage relation itself, however cushioned, constitutes a form of domination. The school’s egalitarians could and did embrace distributive aims openly in their later work, which rather confirms than refutes the observation that the frame and the politics fitted one another. What the critics could not show, and what honest grading must record, is that the fit was a defect rather than a discovery: perhaps distribution really is where the moral action lies, and the classical insistence on relations was the residue of a metaphysics the evidence never supported. The ledger marks the charge as identifying a real transformation, with its evaluation left dependent on the domination question taken up next.

Exploitation without domination, and class without relation

The third and fourth classical charges are companions, since both complain that a relational phenomenon was converted into a distributional one. Roemer’s reformulation defined exploitation through the distribution of alienable assets: a group is exploited when it would fare better withdrawing from the economy with its per capita share of those assets. The construction achieved real analytical gains, separating the moral objection to capitalism from the labour theory of value and showing that exploitation like outcomes can arise under many institutional arrangements. Its critics, including sympathetic readers inside the school’s orbit, pressed a difficulty that Roemer himself came to acknowledge in part. The definition cannot distinguish domination from inequality. A group may be asset poor, may satisfy the withdrawal test, and may therefore count as exploited, in a situation where nobody commands its labour, directs its working day or stands over it in a relation of personal or structural subordination. Conversely, relations of command that classical theory placed at the centre of the factory regime register in the definition only insofar as they correlate with asset holdings. What dropped out of the concept was the experience and the structure of being subject to another’s direction within production, which for the classical tradition was not an ornament on the concept of exploitation but its core.

What is lost when exploitation cannot mark domination?

A definition keyed to asset distributions records who would gain from equalisation, but it cannot record who gives orders, who may be dismissed at will, whose time is commanded and whose is not. Critics argue that this subordination inside production is the phenomenon the concept of exploitation was built to name, and that a concept silent about it has changed the subject.

The school’s reply had two layers. The first held that domination and exploitation name distinct wrongs, each deserving a concept of its own, and that a theory which separates them explains more than a theory which fuses them into a single indignant blur. There is force in that: the analytical habit of distinguishing is not automatically an evasion, and Roemer’s framework generated a family of precise notions, including socialist and status exploitation, that let arguments proceed by cases. The second layer conceded ground gracefully. Roemer’s later writing treated the property relations definition as capturing an important species of injustice rather than the whole genus, and the egalitarian philosophy he and Cohen developed supplied separate accounts of domination, luck and responsibility. Graded against the original ambition, however, the charge stands in its central contention. The reformulation was announced as a reconstruction of Marx’s concept, preserving its explanatory and critical content in defensible form. What it preserved was the distributive skeleton. The relational content, the claim that exploitation names a way persons are governed within production and not merely a way goods are divided after it, was not reconstructed but discarded, and the discarding was seldom flagged with the candour the school demanded when classical texts quietly changed the subject.

The class charge mirrors this one. Erik Olin Wright’s class analysis, the school’s most empirically serious construction and the subject of Erik Olin Wright class analysis, located class positions in relations to productive assets: means of production, organisational resources and credentials or skills. The schema proved extraordinarily fruitful for survey research, generating hypotheses about income, consciousness and political alignment that could be tested across countries. Classical critics nonetheless objected that a schema of asset locations converts class from a relation between persons into a possession held by persons, a slot in a taxonomy rather than a bond of dependence and antagonism. On the classical account, a capitalist is not someone who happens to hold certain assets, as a tall person holds a height; the capitalist exists only in relation to wage labour, and the relation is one of mutual constitution under conditions of conflict. A taxonomy can sort the population with great precision while the relational claim, that the positions create and recreate one another through an ongoing process of extraction, slips through its categories unrecorded.

Can an asset based account capture class as a relation?

It captures the distribution of advantages that relations produce, and it predicts attitudes and incomes with testable precision, which the classical vocabulary rarely attempted. What it does not capture is the process by which positions constitute one another through extraction and command. Critics conclude that the schema maps the sediment of class while the river runs elsewhere.

Wright, to his credit, engaged this objection more directly than most, revising his schema across decades, adding exploitation centred mechanisms to the asset framework and conceding that location within relations, not possession of resources, was the concept’s proper home. His intellectual itinerary is the school’s best exhibit for the defence that its frameworks were corrigible under criticism. Yet the itinerary also documents the charge’s substance: each revision moved the schema closer to the relational understanding the classical critics had defended from the start, which suggests the original conversion of relation into possession was a genuine loss and not merely a difference of taste. The ledger grades the class charge as substantially answered by Wright’s own development of the framework, with the answering done by movement toward the critics’ position, a pattern the trajectory section will meet again in a more general form.

Behind both charges sits a question of explanatory taste that the article should state plainly rather than pretend to resolve. The school preferred concepts that sort, count and compare. The classical tradition preferred concepts that narrate a process and locate a contradiction. Each preference illuminates what the other obscures, and the history of Marxist theory can be read as an oscillation between them long before the September Group existed. The school’s distinct failing was not its preference, which it was entitled to, but its rhetoric: it presented the sorting concepts as replacements that had exposed the process concepts as confused, when the defensible claim was the weaker one that the two vocabularies answer different questions. Cohen’s treatment of class in his historical materialism writings, Wright’s taxonomies and Roemer’s definitions all exhibit moments of that overclaim, and the overclaim is what transformed a fruitful methodological pluralism into a quarrel about which side had dissolved Marxism. On that quarrel the ledger will rule later; here it is enough to record that the conceptual charges, exploitation and class alike, identify real losses, partially repaired from within, and consistently overstated by critics who treated every loss as a refutation and every repair as a surrender.

The silence about the labour process

Of all the classical charges, the one concerning the labour process is the simplest to state and the hardest for the school to answer, because it rests on an absence rather than an argument. Marx’s analysis of capitalism reserved its most detailed attention for the interior of production: the working day, the cooperation of labourers under one roof, the division of labour in manufacture, the machinery system of modern industry, the real subsumption of labour under capital. In those chapters the distinctive claims of the theory are cashed out in concrete descriptions of how human activity is organised, paced, fragmented and surveilled, and how the organisation generates both surplus and resistance. The labour process tradition descending from Harry Braverman’s study of deskilling made that interior the centre of Marxist inquiry for a generation of sociologists and historians. Against that background, the analytical school’s corpus is striking for what it does not contain. Its characteristic scenes are markets, distributions, games and constitutions. The factory, the office floor, the call centre, the warehouse: these settings, where most wage labourers spend most of their waking lives under relations the theory exists to explain, appear in the school’s work mainly as black boxes within which labour power, once purchased, is somehow converted into labour performed.

What is missing from the school’s picture of production?

The interior of production itself is missing. Its models treat production as a function converting inputs into outputs, and its class schemas locate positions by assets rather than by place in the organisation of work. Defenders call this a division of labour among researchers. Critics answer that the omission follows from the exchange framework, which has no interior to examine.

The defence from division of labour deserves a fair statement. No school can work on everything, and the analytical Marxists concentrated on questions, the logic of historical explanation, the coherence of exploitation, the structure of class locations, where they judged the tradition’s confusions to be most damaging. Braverman’s successors, meanwhile, were producing labour process research in abundance, so the topic was hardly orphaned within Marxism as a whole. A charge of omission, the defence continues, proves too much: every research programme omits nearly everything, and omission becomes culpable only when the omitted material would have corrected the programme’s positive claims. That is a strong argument in general, and the ledger accepts its logic. The difficulty is that the omitted material plausibly would have corrected, or at least complicated, several of the school’s positive constructions, and the school’s own standards required it to check. The asset based account of exploitation abstracts from the extraction of effort, which is where the indeterminacy of the labour contract, the gap between labour power bought and labour actually performed, does its theoretical work. The class schema’s organisational assets gesture toward authority at work without examining how that authority is exercised hour by hour. The collective action models treat the workplace as one arena among others, although the classical tradition’s claim was always that shared subjection within production is what makes workers’ collective action different in kind from consumers’ or voters’ coordination problems.

Elster came closest to the missing interior, in his discussions of technology, his treatment of the forces of production and his sceptical examinations of deskilling claims, and even there the engagement is with theses about the labour process rather than with the process itself. The school’s characteristic evidence was textual and formal; the labour process tradition’s characteristic evidence was ethnographic and archival. A reader formed by the second tradition who opens the first tradition’s books experiences a genuine estrangement, as if a theory of seafaring had been written by scholars who modelled fleets and freight rates with great sophistication while displaying only a theoretical interest in water. The estrangement is data. It records that the school’s Marxism was a Marxism of the study and the seminar, continuous with the biography of its members as academics, and that its picture of capitalism was drawn from the economy’s surface of circulation more faithfully than from its depths of production.

Grading this charge requires separating its two forms. The weak form, that the school wrote little about the labour process, is simply true and was never seriously denied. The strong form, that the silence was dictated by the method, is partly true: the exchange and game frameworks have no native vocabulary for the interior of production, and importing one would have required the school to develop tools it did not possess. But the strong form overreaches when it implies that analytical tools as such cannot enter the workshop, since principal agent models of effort extraction, incomplete contract theory and the economics of authority all represent analytical attempts on exactly that territory, some of them by economists adjacent to the school. The school could have gone there and largely chose not to, which returns the verdict to a matter of priorities rather than necessities. Priorities, however, are exactly what a school’s critics are entitled to judge, and a Marxism that deprioritises the site Marx himself treated as decisive has some explaining to do. The ledger records the labour process charge as standing, with the explanation lying in the school’s formation and instruments rather than in any argued conclusion that the topic did not matter. Silence of that kind, unargued and unexplained, is precisely the sort of move the school’s creed existed to prevent, and its presence here is an early sign of the pattern the internal criticisms will state in full.

The normative turn: completion or departure

The final classical charge concerns the school’s destination. Beginning in the nineteen eighties, several leading members migrated from the theory of history and exploitation toward normative political philosophy: Cohen toward the defence of equality against incentives based justifications of inequality, Roemer toward a designed egalitarianism of vouchers and coupon stock markets, others toward the theory of justice, responsibility and luck. The resulting work was formidable, and much of it entered the mainstream of liberal egalitarian thought, where Cohen’s critique of Rawlsian incentives and his later defence of a socialist ethos were debated by philosophers who had no particular interest in Marx. Classical critics drew a pointed conclusion. The later work, they argued, is continuous with liberal political philosophy in its questions, its methods and its conclusions, and its connection with Marxism runs mainly through the biographies of its authors. A trajectory that begins with the defence of historical materialism and ends with a camping trip thought experiment about socialist principles has not completed Marxism’s project. It has exchanged that project for another, more tractable one, while retaining the ancestral name as a mark of origin rather than a description of content.

What changed when the school turned to normative theory?

The subject itself changed. Where the earlier work asked how capitalism works and whither history tends, the later work asked what justice requires and which distributions can be defended, in continuity with liberal political philosophy. Critics err only in calling the change a betrayal, since it was announced, argued and defended in the open as the inquiries led there.

The school’s reply begins by denying that the normative inquiry was optional. The classical tradition, its members argued, had always traded on moral condemnation while officially disowning moral theory, treating exploitation as a scientific finding and its wrongness as too obvious to argue. That posture could not survive the school’s own reconstructive work, because Roemer’s analysis had shown that the wrongness of exploitation, once the concept was clarified, depended on prior judgements about the justice of asset distributions, judgements the tradition had never supplied. The normative turn was therefore forced by the analytical project itself: follow the arguments about exploitation honestly and one arrives, whether one wishes to or not, at the theory of distributive justice. The question of whether exploitation is a moral or a technical notion, pursued in is exploitation moral or technical, is exactly the hinge on which the turn swung. On this account the later philosophy is Marxism’s unfinished business conducted with better tools, and its convergence with liberal egalitarianism records a discovery, that the defensible core of the socialist complaint is an egalitarian complaint, rather than a defection.

Critics answered that the discovery was purchased by the earlier clarifications, each of which had quietly discarded a distinctively Marxist element: history as a directional process, exploitation as domination, class as relation, the labour process as the theory’s home ground. Strip those away and of course what remains resembles liberal egalitarianism, since the resemblance was manufactured by the stripping. The reply has rhetorical force, and the ledger has already recorded partial verdicts supporting its premises. Yet it proves less than its users hope. A tradition is not obliged to retain elements merely because they are distinctive; astrology is distinctive too. The question is whether the discarded elements were true or fruitful, and on that question the critics’ case varies element by element, strong on the labour process, contested on functional explanation, weak where it defends obscurity as depth. The normative turn charge, graded whole, identifies a real discontinuity of subject matter and a real continuity of personnel and motivation. Calling it a departure is accurate. Calling it a completion is the school’s own preferred description and is defensible in the limited sense that the turn answered questions the tradition had posed and evaded. Calling it a betrayal imports a standard of loyalty that the school’s creed explicitly rejected, since the creed subordinated loyalty to argument from the first day. The ledger’s verdict is departure, openly conducted, with the evaluative quarrel about whether departure was loss or liberation referred to the reader’s judgement on the earlier charges.

One more consideration belongs here, because it reverses the usual polarity of the debate. The normative work is, by common consent including that of its philosophical opponents, the part of the school’s output with the strongest claim to permanence. Cohen’s writings on equality and incentives altered how a generation of political philosophers framed the difference principle. Roemer’s proposals for market socialism forced defenders and critics of markets alike to specify which features of capitalism do the distributive damage. If the school’s Marxism dissolved into liberal egalitarian philosophy, it dissolved upward, into the most rigorous conversation available on its questions, and the host discipline was changed by the arrival. Classical critics who cite the turn as evidence of failure must explain why failure so often looks like influence, and their explanations, usually turning on the alleged sterility of academic philosophy as such, tend to indict the university rather than the school. The charge as a description stands. The charge as a condemnation depends on premises about where serious thought happens that the critics seldom defend outright, and the ledger declines to smuggle those premises into a verdict.

The criticisms ledger

The criticisms ledger gathers the charges examined in this article into a single findable form. Each row names a criticism, its source, whether it arises from classical Marxism or from the school’s own standards, the best reply available to the school, and a verdict. Every verdict in the ledger carries the statement of what would reverse it, since a verdict without a defeater is a slogan, and the discipline of stating defeaters is borrowed from the school’s own practice at its best. Readers should treat the verdicts as graded judgements on the published record, open to revision by the evidence each defeater specifies.

Criticism Source Direction Best reply Verdict and defeater
Methodological individualism assumes pre social agents with given preferences Callinicos and classical critics Classical The modest version defended by Levine, Sober and Wright constrains explanation without denying social formation Partly standing; defeated if the school’s practice is shown to have modelled preference formation as a central, sustained inquiry
Exchange models build voluntariness into the wage relation Classical critics of Roemer Classical The framework interrogates asset distributions, and its counterfactuals target background compulsion directly Standing in its refined form about expressive limits; defeated by a demonstration that the frame states constitutive relations without residue
The reformulated exploitation cannot distinguish domination from inequality Critics within and beyond the school’s orbit Classical Domination and exploitation are distinct wrongs and separating them is clarification, not loss Standing against the original reconstructive ambition; defeated if the property relations definition is shown to entail subordination within production
Class locations convert a relation into a possession Classical critics of Wright’s schema Classical The schema generates testable claims the relational vocabulary never produced, and Wright revised it toward relations Substantially answered by Wright’s revisions; the verdict would be reversed if the original schema were shown to have encoded relational process from the start
The labour process is absent from the school’s work The Braverman tradition and labour process sociologists Classical Division of labour among researchers, and the topic flourished elsewhere in Marxism Standing; defeated by identification of a substantial labour process literature produced within the school’s programme
The normative turn departed from Marxism while claiming completion Classical critics of Cohen and Roemer Classical The turn was forced by the school’s own findings about exploitation and justice Departure established, betrayal not established; defeated if the later philosophy is shown to depend essentially on distinctively Marxist premises
The founding defence of functional explanation failed the school’s own tests Elster, from within the group Internal Cohen’s revisions supplied qualifications and partial mechanisms in later work Standing; defeated if the original defence is shown to have met the mechanism standard the school applied to others
The school’s claim of no distinctive Marxist method rested on assertion rather than argument Analytical philosophers and internal critics Internal The claim was a negative existential supported by case by case failures of proposed distinctive methods Standing in part; defeated by production of a sustained argument for the claim meeting the school’s standards of explicitness
The school produced little empirical research despite its scientific self image Internal critics and sociologists of the school Internal The international class survey project was a major empirical achievement, and theory was the school’s chosen contribution Standing, with the survey as the substantial exception; the judgement that the school produced little empirical work would be defeated by identification of a substantial empirical research programme beyond the class survey project that the school generated and sustained
The trajectory of abandoned theses proves the method was biased Classical critics and some defenders, in mirror image Both The trajectory is equally consistent with honest inquiry following arguments Not evidence for either side; defeated only by evidence distinguishing the two explanations of the same sequence

How should the criticisms ledger be read?

The ledger orders charges by the strength of their binding force on the school, not by their volume in the polemical literature. Internal charges come last in the table but weigh most, because they engage premises the school itself professed. Classical charges weigh according to how far their refined forms survive the school’s best replies.

Reading the ledger vertically teaches a lesson that the polemical literature, organised horizontally by camp, tends to hide. The classical charges cluster around a single theme, the conversion of relations into distributions, while the internal charges cluster around a different one, the uneven application of proclaimed standards. The two themes are independent: a school could be guilty on both counts, innocent on both, or guilty on either one alone, and the evidence for each must be gathered separately. Much confused writing on this school results from treating a verdict on one theme as a verdict on the other, so that a reader who finds the exchange models illuminating concludes the internal charges must be spiteful, or a reader convinced by Elster’s internal critique concludes the classical charges must all be sound. The ledger exists to block that inference. Its rows are severally graded, and the article’s overall assessment is nothing more than the sum of the rows, with no gestalt verdict imposed on top. That restraint is itself a judgement about how schools should be assessed: in parts, against named standards, with defeaters attached, rather than in the round, against a mood.

The internal inconsistency: functional explanation judged by its own school

The internal criticisms begin where the school began. Cohen’s 1978 defence of historical materialism rested on functional explanation: the claim that production relations are as they are because they develop the productive forces, and superstructures are as they are because they stabilise those relations, in something like the way a biological trait is explained by its contribution to fitness. The defence was the school’s founding document, the text around which the September Group coalesced, and its argumentative strategy was to legitimate a pattern of explanation by analogy and by the alleged primitiveness of the demand for mechanisms. Within four years Elster, the group’s own most formidable methodologist, published a systematic dismantling of that strategy, arguing that functional explanation in social science is legitimate only where a feedback mechanism can be specified, that Cohen had specified none for his central historical claims, and that the biological analogy fails because societies lack the selection processes that make trait explanation respectable in biology. The episode is treated in full in G. A. Cohen functional explanation; its significance here is jurisdictional. Elster did not import foreign criteria. He applied the criteria of clarity, mechanism and testability that the group jointly professed, the same criteria Cohen’s book had wielded against vaguer Marxisms, and the founding text failed them.

Why did the defence of functional explanation fail internally?

Because its central explanations invoked consequences as causes without supplying the feedback loop that would connect them, and the school’s own stated standards required such a loop wherever a functional claim was made. Elster’s critique succeeded from within because it needed no premise that Cohen and the group had not already endorsed in public.

The internal inconsistency, the namable claim of this article, is the structural fact that episode reveals. A current whose opening theoretical move would not pass the entrance examination it set for others carries a difficulty that no external critic needs to manufacture. The difficulty is not that Cohen’s book was worthless; by wide agreement it remains the most rigorous statement historical materialism has received, and its failure under audit is partly a measure of how exacting the audit was. The difficulty is one of sequence and candour. The school announced its standards as the price of admission to serious discussion, charged the classical tradition with never having paid, and then discovered, through its own internal processes, that its founding text had not paid either. To the school’s credit, the discovery was published rather than buried, and Cohen’s subsequent work, qualifying the functional claims, supplying candidate mechanisms and eventually restating the theory’s core in terms less dependent on the contested pattern, is a model of how a thinker should behave when his own side’s referee blows the whistle. But credit for the correction does not erase the entry in the ledger. The founding defence was weak by the school’s criteria, the weakness was demonstrated by a member within four years, and every later statement of the school’s methodological superiority over the tradition must be discounted by that founding fact.

It is worth pausing on why this charge binds more tightly than any in the classical dossier. A classical critic who dislikes exchange models can be answered by a defence of modelling. A classical critic who mourns the labour process can be answered by a plea of priorities. There is no corresponding answer to Elster’s critique available within the school’s own commitments, because the critique’s premises are those commitments. The only exits are to lower the standards, which dissolves the school’s raison d’être, to show the mechanisms were present all along, which decades of subsequent commentary have not managed, or to concede, which is what in substance occurred. Internal criticism of this purity is rare in the history of intellectual movements, rarer still when the target is a founding text and the critic a founding member, and its existence is the strongest reason for taking the school’s standards seriously even while recording their uneven application. A movement that generates its own most damaging critic has something real at its centre. A movement whose founding text cannot survive that critic has something real to answer for. The ledger holds both sentences together, because both are true, and any assessment that keeps only one has chosen a camp rather than reached a judgement.

The pattern repeated in miniature across the school’s later history. Its members’ criticisms of one another, Roemer on the labour theory of value’s remaining defenders, Wright on the cruder versions of class reductionism, Elster on nearly everyone, were consistently more exacting than anything the classical journals produced against the school, and consistently more effective, because they proceeded from shared premises to unwelcome conclusions without the static of camp warfare. If the school is to be faulted for the internal inconsistency, it must in fairness be credited with the institutional virtue that exposed it: a seminar culture in which the referee’s whistle blew at home. The classical tradition, for all its other strengths, seldom institutionalised an equivalent. Its internal critics were more often expelled than answered, a contrast the school’s defenders are entitled to press, and one that complicates any simple morality tale in which rigour lives on one side of the quarrel.

An asserted method, and the missing empirical programme

The second internal charge concerns the school’s most quoted slogan: the claim that there is no distinctive Marxist method, that the tools of good inquiry are the common property of the sciences and humanities, and that Marxism stands or falls as a set of substantive claims rather than as a logic of its own. The slogan did valuable polemical work against dialectical exceptionalism, the doctrine that Marxist thought obeys special laws of reasoning exempt from ordinary standards. Yet judged by the school’s own criteria, the slogan’s epistemic status in the school’s writings is embarrassing. It was asserted in prefaces, repeated in interviews, presupposed in reconstructions, and wielded against opponents, far more often than it was argued. The argument it needed is not trivial: a negative existential over all possible methods, supported by examining the serious candidates for a distinctive Marxist logic and showing, case by case, that each is either ordinary reasoning in costume or no reasoning at all. Fragments of that argument exist, in Elster’s sortings of Marx’s claims and in Cohen’s discussions of dialectics, but the sustained demonstration the school’s standards would demand of anyone else’s central claim was never assembled. An assertion repeated by rigorous people remains an assertion, and the school’s creed had a word for building on assertions, a word it applied unsparingly to the tradition’s habits.

What does it mean that the method claim was asserted rather than argued?

It means the school’s most consequential methodological commitment rested on accumulated confidence rather than on a demonstration meeting its own published standards for central claims. Supporters can point to case by case sortings of Marx’s arguments as partial evidence. The ledger records the gap between that evidence and the slogan’s sweeping scope.

Defenders of the school have a reply that deserves recording at strength. Some claims function as working hypotheses whose credentials are their fruits: the presumption that no special logic exists guided reconstructions that repeatedly found ordinary arguments, good and bad, where dialectical ones had been advertised, and each successful sorting added inductive support. On this view the slogan was argued the way research programmes argue their hard cores, by accumulation rather than by a single set piece demonstration. The reply is respectable, and it moves the charge from standing to partly standing. What it cannot do is restore parity, because the school did not extend the same inductive charity to the tradition’s hard core claims, historical materialism’s among them, which were required to meet explicit standards in explicit texts or be marked down as articles of faith. Charity for one’s own core, audit for everyone else’s: the asymmetry is the charge, and the reply, by describing the asymmetry accurately, confirms more than it refutes.

The third internal charge is the plainest of all. The school presented itself as bringing scientific standards to a tradition careless of evidence, yet its own empirical output was remarkably slight. Bibliographies of the school are dominated by philosophy, formal theory and textual reconstruction. Its members produced models of economies that never existed, analyses of classes in societies they did not survey, and explanations of historical transitions they did not test against archives. The great exception, and it is a genuinely great one, is the international class structure project associated above all with Wright: a coordinated survey programme across multiple countries, built on the school’s own class schema, generating data and findings that entered mainstream stratification research and are discussed in the article on Wright’s class analysis. One swallow, however sustained and admirable, does not make a research programme’s summer. Beyond the survey project, the school generated no comparable empirical enterprise: no historical testing programme for the theory of history its founder had defended, no measurement tradition for exploitation to set beside the national accounting of income shares, no archival school, no data infrastructure that outlived its founders’ individual projects.

How much empirical research did the school generate?

Far less than its scientific self presentation promised. The international class survey stands as the substantial exception, a real achievement of design and coordination. Around it lie models untested against data, historical explanations unexamined in archives, and programmatic statements about testability that the school itself seldom converted into tests.

Why the shortfall occurred is a question the article should answer carefully, because the easy explanations are unflattering and incomplete. Part of the cause lies in personnel: the school recruited philosophers and theorists, whose skills and incentives pointed toward the article and the monograph rather than the dataset. Part lies in the division of intellectual labour the school inherited from its disciplines, in which theory and evidence are separate crafts with separate rewards. Part, less comfortably, lies in the logic of the school’s own trajectory: as its members’ conclusions drifted away from Marxist theses, the motive to test those theses empirically weakened, since a claim one no longer believes needs no research programme to sustain it. Whatever the weighting among causes, the consequence for assessment is fixed by the defeater discipline this article practises. The verdict on this charge is issued with its defeater attached, in the form the ledger records: name a further programme of empirical research, substantial in scale, that the school brought into being and kept running alongside the survey project, and the verdict falls. No such programme has been named in the literature surveyed for this article, and the invitation to name one stands open in the ledger, which is where such invitations belong. Until it is answered, the charge stands: a current that demanded testability of a tradition produced few tests of its own, and the tradition, for all its theoretical untidiness, could point to generations of historians, sociologists and economists whose empirical work, however contested, filled libraries the school’s founders themselves consulted.

The three internal charges compose a single portrait when laid side by side. A founding explanation that failed the mechanism test. A methodological slogan that outran its argument. A scientific posture unmatched by empirical production. Each charge is documented from the school’s own record and judged by the school’s own criteria, and each survives the best replies in at least a substantial part. That is the internal inconsistency in its full extent: not hypocrisy, which would imply the standards were never meant, but unevenness, the systematic tendency of the school’s rigour to relax at exactly the points where its own commitments were at stake. Unevenness of that kind is the most common failing of intellectual movements and the least often documented against any of them, because documenting it requires a critic who shares the movement’s premises. Analytical Marxism had such critics on its own membership list, which is why its ledger, alone among modern Marxisms, can be compiled from internal sources. The compilation is damaging. It is also, in a sense the school’s founders might have appreciated, a backhanded vindication of their method: the standards worked, even on the school, once somebody bothered to apply them there.

What the school got right

An assessment that stopped at the charges would commit the offence it has been documenting, uneven application, by auditing the school’s failures with a care it denied its successes. The school’s achievements were considerable, and several of them permanently improved the intellectual environment in which Marx is read, including for readers who reject everything else the school stood for. The first achievement was the demand for clarity itself. Before the school’s intervention, a great deal of Marxist writing sheltered behind a specialised idiom in which the failure of an argument to conclude could be disguised as dialectical movement, and in which requests for definitions were treated as positivist pedantry. The school’s insistence that claims be statable in sentences whose truth conditions a reader could inspect did not merely embarrass particular authors. It changed what counted as a move in the game, so that subsequent Marxist theory, even in traditions hostile to the school, became measurably more careful about distinguishing empirical claims, conceptual claims and normative claims. Obscurity had protected weak arguments for decades, and the school’s founders were right, simply and importantly right, that a tradition confident of its insights should welcome the light.

What did the demand for clarity achieve?

It deprived weak arguments of their traditional shelter, in which specialised vocabulary disguised the absence of inference as dialectical subtlety. After the school’s intervention, Marxist claims across rival traditions were more often sorted into empirical, conceptual and normative kinds, each answerable to its own tests. Even the school’s opponents absorbed the discipline.

The second achievement was the critique of functional explanation itself, which belongs on the credit side of the ledger precisely because it was an internal production. Elster’s demand for mechanisms, generalised by later writers into a broad movement across the social sciences, permanently raised the standard for explanations that cite the beneficial effects of a practice as the reason for its existence. Historians, sociologists and political scientists who have never read a page of Marx absorbed the lesson that a consequence is not a cause until a feedback story connects them, and the lesson’s route into the mainstream ran substantially through the school’s quarrel with its own founder. A movement should be judged partly by the quality of the tools it leaves on the bench for others, and the mechanism demand is a tool of the first quality. The third achievement was the school’s treatment of collective action. By taking seriously the free rider problem that Mancur Olson had posed, the school forced Marxist theory to explain, rather than presume, the formation of class wide organisations, and its models of thresholds, selective incentives and assurance games gave the tradition a realistic psychology of mobilisation it had previously replaced with the incantation of objective interests. The problem remains unsolved in its general form, as the school’s writers conceded, but a problem honestly posed and partially modelled is worth more than a solution falsely announced, and the school’s posing of it was exemplary.

Why is the collective action problem a lasting contribution?

Because it converted a slogan into a research question. Where the tradition had inferred organisation from shared position, the school asked what could move situated individuals to bear costs for collective goods, and modelled thresholds, incentives and expectations accordingly. The problem survives unsolved, which the school admitted, and the admission itself raised the tradition’s standard of honesty.

The fourth achievement was the demand for a normative theory, which the classical dossier counted as a departure and which belongs equally in the credit column. The tradition had condemned capitalism in a moral vocabulary it officially disavowed, deriving indignation from a theory that claimed to have dispensed with morality. The school exposed that incoherence and then did the harder thing: it built the missing theory, in Cohen’s writings on equality and community, in Roemer’s designs for egalitarian institutions, and in a generation of work on responsibility, luck and need. Whether the resulting philosophy is labelled Marxist matters less than the fact that the gap it filled was real, was located by the school’s diagnostic work, and would still be yawning had the school waited for permission from the tradition’s custodians. The fifth achievement returns to the survey project, whose credit entry balances its use in the internal charge. Wright’s class structure research designed instruments that measured a Marxist schema against the populations of several countries, published its data, invited replication and criticism, and saw its categories absorbed into comparative research programmes with no Marxist allegiance. That is what an empirical achievement looks like, and its singularity in the school’s record reflects on the record, not on the achievement.

To these five, a sixth and quieter credit should be added: the school’s members modelled, in their public revisions, a relationship to their own past claims that academic life rarely rewards. Cohen annotated his earlier positions with their defects. Roemer restated his definitions as counterexamples accumulated. Wright rebuilt his schema in the open across three decades, each edition a record of criticisms absorbed. Intellectual honesty of that visibility is a contribution to the culture of inquiry independent of any doctrine, and it is the deepest reason the internal criticisms in this article can be documented so fully: the school’s people wrote the evidence down. A tradition whose critics must rely on the school’s own candour for their strongest charges owes that candour an acknowledgement, which this section pays without discounting the charges the candour supports.

The trajectory question, and why it settles nothing

No feature of the school’s history is cited more confidently, by more parties, to more incompatible conclusions, than its trajectory. The bare sequence is not in dispute. A circle forms in 1979 to defend Marxism by rigorous means. Its founder’s theory of history is dismantled from within the group in four years. The labour theory of value is abandoned by its ablest reconstructors. Exploitation is redefined in terms that sever it from labour and from domination. Class theory migrates toward a sociology its classical custodians barely recognise. The leading figures end their careers as egalitarian philosophers, institutional designers and stratification sociologists, several of them stating plainly that they no longer hold the doctrines they gathered to defend. Critics of the school read the sequence as a confession: apply the method and Marxism evaporates, therefore the method was a solvent all along, chosen, consciously or not, for its dissolving power. Defenders read the identical sequence as a vindication: honest inquirers followed arguments to unwelcome destinations, therefore the method was genuine inquiry, and its results are the tradition’s overdue correction. Both readings cannot be right, and this article’s position, argued in the companion guide and maintained here, is that neither is entitled to the evidence, because the trajectory underdetermines its explanation.

Why can the trajectory settle nothing?

Because both candidate explanations predict the same observed sequence. A method biased against its object would produce a steady abandonment of the object’s claims. So would honest inquiry into claims that happen to be false. An outcome predicted equally by rival hypotheses is evidence for neither, however often partisans on both sides present it as decisive.

The logic deserves spelling out, because its neglect powers a small industry of polemic. Evidence discriminates between hypotheses only when the hypotheses assign it different probabilities. A trajectory of abandonment is highly probable if the school’s method was rigged, since rigged tests fail their targets. The same trajectory is also highly probable if the method was sound and the targeted claims were, in the main, indefensible, since sound tests fail false targets. Observing the trajectory therefore leaves a rational reader’s relative confidence in the two explanations almost exactly where it began, which is the precise meaning of saying the trajectory is evidence for neither. What would discriminate is different in kind: evidence about the method’s operation in cases where the truth is independently known, demonstrations that the same tools, applied to non Marxist doctrines, spare them for no good reason, or internal documents showing conclusions fixed in advance. Critics have gestured at the second kind, pointing to the school’s gentler handling of the mainstream economics it borrowed from, and that gesture identifies the right sort of test even where its execution remains anecdotal. Until such discriminating evidence is produced and weighed, the honest sentence about the trajectory is the one both camps resist: it is consistent with everything and therefore proves nothing.

Did members of the school raise these criticisms themselves?

Several of the most damaging ones, yes. The functional explanation critique came from a founding member. The empirical thinness was acknowledged in the school’s own retrospective assessments. Members annotated their earlier claims with their defects in later editions. The school was not a complacent target surprised by hostile scrutiny; it was its own severest examining board.

There is a corollary that partisans resist even more strongly. If the trajectory cannot convict the method, it also cannot acquit the tradition. Defenders of classical Marxism sometimes argue as though the school’s destination, liberal egalitarianism, itself refutes the school, as if arriving somewhere unfashionable were a logical error. It is not. Arguments are not graded by the company their conclusions keep, and the classical tradition’s own history, a century of schisms in which every faction’s trajectory was cited by every other as proof of apostasy, should have inoculated its members against trajectory arguments in all forms. The school’s members, for their part, sometimes wrote as though the mere fact of having changed their minds under argument certified the change as progress, a vanity the record does not support either: people also drift, tire, follow incentives and mistake the fashion of adjacent disciplines for the verdict of reason. The trajectory is a fact to be explained by the patient, row by row methods of the ledger, not a verdict to be quoted. Each abandonment must be assessed on the arguments that produced it, and those arguments are exactly what the preceding sections have graded. Where the arguments were strong, as on functional explanation, the abandonment reflects credit on the inquirers and discredit on the abandoned claim. Where the arguments were partial, as on class and exploitation, the abandonment overshot, and the ledger has said so. Aggregation into a single story of dissolution or of enlightenment is the one move the evidence forbids.

What general lesson does the case offer other traditions?

That importing a formal apparatus is never a neutral act of translation, and that a movement’s first obligation is to apply its proclaimed standards to its own founding texts before applying them to rivals. The analytical Marxist episode is the clearest documented case of both lessons, which is why it rewards study far beyond Marxist circles.

The adjacent disciplines confirm the generality. Economics itself, when it formalised the theory of the firm or of growth, discovered that its apparatus had quietly decided questions its practitioners believed they were investigating. Political philosophy, when it adopted the contract device, found conclusions about obligation prefigured in the device’s setup. Formalisation clarifies by excluding, and the excluded material does not announce its departure. A tradition importing such tools therefore needs two disciplines the school practised unevenly: an inventory of what the apparatus cannot say, kept as carefully as the results it produces, and a standing audit of founding commitments by the imported standards, conducted before opponents think to demand it. The school’s history shows the cost of the first omission in its classical charges and the cost of the second in its internal ones, and it shows something rarer, that both omissions are corrigible in public by people willing to publish their corrections. That willingness, finally, is the school’s best defence against the harshest available verdict, that it dissolved Marxism. Dissolution implies an agent and an intention. The record shows instead a group of scholars who asked questions in a sequence, published the answers they reached, and revised when the answers were damaged, including by one another. If that process, applied to Marxism, leaves less of Marxism standing than its custodians wish, the finding belongs in the first instance to Marxism’s account, and the school’s genuine faults, uneven standards, an unpaid argumentative debt, a thin empirical record, are faults within the process rather than evidence that the process was a fraud.

The strongest reply, and what survives it

Fairness to the school requires stating its collective reply at the strength its members would have wished, before the article’s verdicts are allowed to settle. That reply, assembled from the school’s defences across many texts, runs as follows. The criticisms amount, in the end, to the complaint that when Marxist claims were stated clearly enough to be assessed, several did not survive assessment. A tradition that treats the demand for clarity as an imposition is not defending its claims against a foreign method; it is protecting them from scrutiny as such. Every specific charge dissolves under that light. Methodological individualism is the refusal to let unexplained wholes do explanatory work. The exchange framework is the price of saying exactly what exploitation means. The labour process was studied by others while the school repaired the foundations those others built on. The normative turn supplied the moral theory the tradition had always needed and never admitted needing. And the internal criticisms, far from embarrassing the school, are its proudest product: name another current in the tradition’s history that generated, published and absorbed criticism of that quality from its own ranks. On this telling, the ledger’s rows record not a school’s failure but a tradition’s discomfort at being examined by examiners it could not dismiss as enemies.

What is the strongest reply the school can make?

That the criticisms reduce to resentment at examination itself: stated clearly, several Marxist claims failed, and the school’s internal critics prove its standards were real. The reply is formidable against charges that target clarity as such. It does not reach charges about substantive assumptions installed by the models, which are a different species of objection.

Which criticisms concern clarity, and which concern assumptions?

The clarity charges concern obscurity, unargued slogans and missing mechanisms, and the school’s reply defeats them, since a tradition may not plead its own vagueness as a defence. The assumption charges concern what the models build in: given preferences, exchange as the elementary act, assets as class. Those stand or fall on the models’ expressive limits, untouched by the reply.

The division drawn in that answer is the article’s final grading instrument, and applying it sorts the ledger cleanly. The school’s reply succeeds completely against the complaint, heard often in the classical journals, that formalisation as such betrayed the tradition’s spirit. It succeeds against the suggestion that internal criticism discredits the criticised, since the opposite inference is the sound one. It succeeds against the betrayal version of the normative charge, for the reasons already given. It does not touch the refined classical charges, because those charges are not complaints about being examined; they are claims about what the examining instruments were built to see and built to miss. A microscope is not refuted by the observation that it cannot weigh, and it is not vindicated either: the question is whether the weighing mattered. On preferences given rather than formed, on relations constituted rather than distributions recorded, on the interior of production where the tradition’s central process runs, the weighing mattered, and the school’s instruments, for all their precision, were not built for it. The concession the school’s reply demands in return must also be paid in full, and this article pays it: the first group of criticisms, those targeting clarity, has no answer, and the tradition’s long resistance to formal statement did protect weak arguments that deserved no protection. A reader who finishes this article believing the school was simply wrong has misread it as thoroughly as a reader who finishes believing the school was simply betrayed.

How does this assessment compare the school fairly?

By comparing like with like over identical periods and on declared dimensions: the school’s empirical output against the tradition’s empirical output of the same period, its internal criticism against the tradition’s, its theoretical yield against the yield of comparable groups. It never weighs a small circle’s production against the accumulated writing of a hundred and fifty years.

What, then, is the sum? Analytical Marxism got wrong, first, the even application of its own standards, and that failure is documented from its own record in its founding text, its central slogan and its empirical shortfall. It got wrong, second, the expressive bookkeeping of its models, repeatedly presenting partial vocabularies as complete replacements and answering refined charges with refutations of crude ones. It got wrong, third, the estimate of its own empirical vocation, announcing a scientific reformation and delivering, the survey project apart, a library of theory. Against that, it got right the demand for clarity, the mechanism standard, the collective action problem, the necessity of normative theory and the practice of public self correction, achievements that outlast the school and now belong to everyone, including its critics. The school did not dissolve Marxism, a task beyond any seminar’s power, nor did it save it. It examined Marxism with instruments that revealed genuine weaknesses in the patient and genuine limitations in the instruments, and the double revelation is its legacy. Readers who want the constructive story should return to the companion guide to the school, keep a study record of the ledger’s rows and their defeaters in a tool such as the VaultBook study notes, and treat every verdict in this article as an invitation to find the evidence that would overturn it.

The seminar as institution: criticisms of the school’s sociology

Some criticisms of the school concern neither its models nor its method but its social form. The September Group was small, self selecting, overwhelmingly located in elite universities in Britain and the United States, and bound by personal acquaintance as much as by doctrine. Critics in the classical tradition, accustomed to movements with journals, factions, publishing houses and roots in labour organisations, regarded that sociology as explanatory. A Marxism produced by a dining and discussion circle of professors, they suggested, would predictably prize the virtues professors prize, argumentative elegance, publishable precision, distance from the untidy evidence of workplaces and movements, and would predictably drift toward the concerns of the neighbouring disciplines its members wished to impress. The school’s trajectory, on this sociological reading, requires no hypothesis about method at all: it is the record of able academics following the incentive gradients of their profession, which reward novelty within philosophy and economics departments and offer no prize for fidelity to a nineteenth century doctrine.

The school’s defenders had answers, and some of them bite. Small groups with high internal standards have repeatedly changed intellectual fields, and the relevant comparison is not with mass movements but with other seminar born currents, several of which the classical critics admire when their conclusions please them. The charge of professional incentive, moreover, proves elastic enough to explain any outcome whatever: had the members defended every Marxist thesis, the same sociology would be invoked to explain their conformity to the expectations of Marxist journals. An explanation compatible with all outcomes explains none, a standard the school’s own teaching supplies and its sociological critics seldom meet. Yet the sociological charge is not wholly empty, because it identifies a real selection effect at the point of recruitment. The group gathered people already disposed to believe that Marxism’s troubles were troubles of statement and argument. Scholars who believed the troubles were troubles of political organisation, of historical evidence or of the labour process itself did not join, were not invited, or left no mark on the programme. A research programme’s founding selection shapes its later blind spots more reliably than any single argument does, and the school’s blind spots, the workplace, the archive, the movement, map closely onto the competences its founding circle did not contain.

There is also the matter of tone, which hostile accounts exaggerate and friendly accounts omit. The school’s writing could be brisk to the edge of contempt toward the tradition it claimed to serve, and its members’ habit of sorting predecessors into the clear headed and the confused tried the patience of readers who had sustained Marxist scholarship through decades when it carried professional costs the seminar room never imposed. Elster’s judgements on revered figures were famously unsparing. Cohen, by contrast, wrote of Marx with an affection his arguments did not always spare, and the difference in register between the two men maps a real difference in the school’s internal culture between those who experienced the project as rescue and those who experienced it as audit. Critics who met only the audit register concluded the whole enterprise was hostile, and the school’s failure to manage its own presentation, to distinguish in public between the claims it was testing and the tradition it was testing them for, cost it allies it might have kept. Presentation is not everything, but a school that understood mechanisms might have foreseen that contempt is a mechanism too, generating resistance that argument alone then cannot dissolve.

None of this overturns the article’s main gradings, and the ledger does not give the sociological charges a row of their own, because they explain the school’s character better than they evaluate its claims. They belong in the assessment nonetheless, as context for the two patterns the ledger does record. The uneven application of standards is easier to understand in a group bound by friendship, where the founding text was a friend’s book and the methodological slogan a shared badge. The thin empirical record becomes intelligible in a group whose members’ careers were made in theoretical disciplines. Explanation is not exculpation, and the school’s own standards, which it applied to the beliefs of historical agents without much interest in their excuses, would not have accepted biography as a verdict on argument. The same courtesy is extended here in reverse: the sociology explains why the failures occurred where they did, while the failures remain failures, graded on the arguments alone.

Cohen under audit: the founding text and its long revision

Because the functional explanation episode carries so much of this article’s argument, its details deserve a slower treatment than a summary allows. Cohen’s 1978 book proposed that historical materialism’s central claims be read as functional explanations of a distinctive and legitimate kind. The development thesis held that productive forces tend to grow across history. The primacy thesis held that the character of a society’s production relations is explained by the level of development of its forces: relations are of the kind they are because relations of that kind, at that level, promote further development. Superstructures, in turn, were explained by their tendency to stabilise the relations they crown. Cohen knew the pattern resembled the consequence explanations of biology, and he defended the resemblance directly, arguing that a consequence explanation can be confirmed by elaboration of the connection between the item and its effect even before the detailed mechanism, the story of how the effect maintains the item, is available. His analogy was with Darwinian explanation before genetics: Darwin could explain the giraffe’s neck by its function without knowing the machinery of inheritance, and the explanation was good science in the interim.

Elster’s rejoinder attacked every joint of that defence. In biology, he argued, the interim is underwritten by a known general process, natural selection, whose existence does not await the genetics of any particular trait; the promissory note is drawn on a bank known to be solvent. In social life no equivalent general process had been shown to exist. Candidate mechanisms, deliberate design by ruling classes, competition among societies, imitation of successful forms, each explained some cases and failed others, and none operated with the reliability the functional pattern required. Worse, the ruling class design story smuggled back the intentional explanations the functional form was meant to transcend, while the competition story required conditions, tight coupling between a society’s survival and its relations’ efficiency, that history routinely violates, since inefficient arrangements persist for centuries behind walls of coercion and luck. Elster’s conclusion was not that Cohen’s historical claims were false. It was that their explanatory form was unearned: the book had identified correlations between levels of development and types of relations, dressed them in the grammar of explanation, and mistaken the grammar for the achievement.

The sequel is as instructive as the critique. Cohen did not denounce the critic or leave the group. He replied, refined, distinguished consequence explanations from their stronger cousins, conceded the need for what he called elaborations, and over subsequent editions and essays moved the theory’s weight toward claims that could survive without the full functional apparatus. Philosophers of social science continued the argument for years, with defenders of Cohen urging that Elster’s mechanism demand, applied consistently, would invalidate much respectable explanation in economics and evolutionary theory itself, and with Elster’s allies answering that the demand was always for a mechanism story of some credible type, not for a completed microphysics of society. The debate’s residue is a standard that neither camp’s partisans can now escape: functional claims in social explanation carry a burden of mechanism specification proportioned to their scope. A school whose founding text provoked that standard into existence, by failing it conspicuously and then arguing about the failure in the open for a decade, contributed something to method that a more cautious founding text never would have. The ledger’s row stands, the founding defence failed the criteria its own school proclaimed, and the row beside it in any honest accounting records that the weakness was the making of the criteria themselves, since the school’s standards were sharpened on the whetstone of its own first book.

One further criticism of Cohen belongs to the classical rather than the internal dossier, and it concerns the theory of history rather than its explanatory form. Historians objected that the development thesis, read as a claim about what actually happened, sits poorly with long periods of stagnation, regression and divergence that the record supplies in abundance, and that the primacy thesis achieves its fit by redescribing recalcitrant cases until the forces can be said to have selected the relations after all. A thesis that accommodates every outcome by adjusting its descriptions courts the same charge the school levelled at dialectics, immunity to evidence. Cohen’s defenders replied that the theses were tendencies, not timetables, and that tendency claims are not refuted by counter instances where countervailing causes are independently identifiable. The exchange ended, like many in this literature, with each side’s standards vindicated to its own satisfaction, and the ledger records the historical objection as contested rather than standing. Its presence matters for a different reason: it shows the classical dossier contains serious empirical arguments and not only methodological laments, a fact the school’s brisk public manner sometimes obscured.

Elster as critic and as target

No assessment of the criticisms is complete without turning the instruments on the school’s chief internal critic, because Elster’s reconstructions attracted objections as serious as the ones he lodged. His great sorting of Marx’s writings into the living and the dead proceeded claim by claim, and critics charged that the sorting criterion did quiet violence to its material. Texts were read as contributions to a social science Marx never quite announced, their rhetorical, polemical and philosophical registers flattened into propositions, and propositions that resisted the flattening were marked as confusion rather than as evidence that the sorting frame was too narrow. Readers formed by the philosophical Marx, the Marx of alienation and praxis, found in Elster’s reconstruction a thinker they scarcely recognised, and they said so with some bitterness. The school’s reply, that recognition is not a criterion of truth, is correct and insufficient, since the question is not whether the reconstruction feels familiar but whether it discards content along with obscurity, and on that question the alienation material remains the contested exhibit: what Elster treated as youthful residue, other readers, including some within analytical philosophy, treat as a theory of human flourishing whose loss impoverished the school’s later normative work by depriving it of a conception of the good life richer than distribution.

Elster’s positive programme attracted a parallel set of charges. His methodological individualism, in its austere form, required explanations to terminate in individual actions and interactions, and critics within philosophy of social science asked whether the requirement could survive its own application to the school’s favourite objects: a class, a market, a state are not obviously decomposable without remainder into the doings of persons, since the descriptions under which persons act already contain the institutions. The more modest doctrine that Levine, Sober and Wright defended absorbed that point by allowing structures as explanatory furniture under a tracing constraint, but Elster’s rhetoric seldom advertised the modest version, and his practice in polemic was to wield the austere one. There is an irony the ledger is obliged to note. The school’s most formidable auditor stands himself accused, by critics using recognisably analytical standards, of asserting a stronger doctrine than his arguments supplied and of applying his sorting standards unevenly to texts he admired and texts he did not. The internal inconsistency, it turns out, was not one man’s founding lapse but a recurring temptation of the school’s style: confidence outrunning demonstration at the points of deepest commitment. That recurrence strengthens rather than weakens this article’s central claim, since a failing distributed across the school’s ablest members is a property of the school and not of a biography.

Elster also illustrates the limits of the trajectory argument from the other side. He travelled further from Marx than any other member, ending as a theorist of constitutions, bargaining and the passions whose Marx book he came to regard with mixed feelings. Critics cite the distance travelled as the method’s reductio. Yet Elster’s later work is among the most respected in its fields, and his studies of precommitment and of local justice are used by scholars with no stake in the Marxist quarrel. If trajectory is evidence, it must be allowed to testify in both directions, and a witness that testifies to everything testifies to nothing, which is the article’s standing point. The fairer judgement on Elster is the ledger’s usual kind: his critique of functional explanation stands among the most successful internal criticisms in modern social theory; his reconstruction of Marx was partial in ways its confidence concealed; his methodological austerity was asserted beyond its argument. Those three sentences can be held together, and holding them together is the discipline this article exists to model.

Roemer under audit: elegance and its price

Roemer’s constructions drew the most technically specific criticisms in the classical dossier, and they repay attention because they show the exchange frame’s costs at the level of theorems rather than slogans. His general theory of exploitation defined exploitative relations for feudalism, capitalism and socialism by a single withdrawal test, an achievement of unification that critics granted even while disputing its foundations. The disputes concentrated on three points. First, the test’s dependence on the distribution of alienable assets made exploitation a function of a counterfactual redistribution, and critics asked whether a relation between actual persons should be defined by the fortunes of hypothetical coalitions in imaginary economies. Second, the framework treated skills and credentials as assets alongside land and capital, generating the category of socialist exploitation by the skilled, a result some readers took as a discovery and others as a sign that the concept had drifted free of the phenomenon, since the credential holder’s advantage involves no extraction from any identifiable other. Third, and most damaging in the school’s own subsequent judgement, the framework’s cleanliness about exploitation coexisted with silence about the production relation in which, on the classical account, the transfer actually occurs, so that the theory explained the pattern of gains from an economy while declining to describe the economy’s interior transactions at all.

Roemer’s responses across his career constitute one of the school’s better exhibits. He conceded the force of the domination objection by rebuilding his egalitarianism on broader foundations, treated his definitions as instruments whose value lay in the discriminations they enabled rather than as captures of an essence, and in his work on equality of opportunity moved the entire discussion toward responsibility sensitive metrics that owed little to the withdrawal apparatus. A critic grading the exchange must record that the target kept moving, and moving largely in the directions the criticisms indicated. What the critics can fairly claim is that the movement was always away from the original Marxist content of the concepts and never back toward it, a one way ratchet that the trajectory section has already declined to treat as proof of anything beyond itself. What the school can fairly claim is that each move was argued in the open with the counterexamples displayed, which is more than the custodians of the unreconstructed concepts ever offered in their defence. The ledger’s row on exploitation records the standing part: against the announced ambition of reconstructing Marx’s concept with its critical content preserved, the reformulation preserved the distributive skeleton and discarded the relational core, and the discarding was a loss whether or not it was also, as Roemer came to suggest, a price worth paying for a workable egalitarian theory.

A word should also be said about the criticism Roemer attracted from the opposite flank, since an assessment confined to classical and internal charges might imply the school was criticised only from the left. Economists in the mainstream found his market socialist designs vulnerable on incentive and information grounds continuous with the calculation debate treated elsewhere in this series, and his coupon socialism proposal was examined, and largely rejected, by specialists in mechanism design for reasons having nothing to do with Marx. The school’s members generally welcomed such engagement as the compliment of being treated as contributors rather than as curators. Its presence in the record matters here for one reason: it confirms that the school’s work was genuinely continuous with the disciplines it entered, exposed to their referees and their standards, and not a protected enclave in which Marxist conclusions were guaranteed by Marxist premises. Whatever else the school got wrong, it did not cheat in that particular way, and several of its critics, whose own traditions maintained enclaves of their own, were poorly placed to make the corresponding claim.

Wright under audit: the survey and its discontents

Wright’s programme attracted criticisms of a more empirical flavour than any other part of the school’s work, precisely because it ventured onto ground where evidence could answer back. The class schema’s middle locations, managers, supervisors, credential holders, the famous contradictory locations between the great classes, were charged by classical critics with multiplying categories until the theory of polarisation that gave the schema its Marxist point had been qualified away. A map on which most of the population occupies intermediate terrain, the charge ran, describes a stratification order rather than a class structure in Marx’s sense, and calling the terrain contradictory locations names the embarrassment without removing it. Wright’s reply, that the map’s complexity was the world’s complexity and that a two class model of advanced capitalism was a catechism rather than a finding, had the school’s characteristic strength of appealing to evidence over loyalty. It also had the characteristic cost recorded throughout this article: each defence moved the framework closer to mainstream stratification sociology and further from the classical claims it was built to test, so that admirers and critics could read the same revisions as maturation or as dissolution according to taste.

The survey findings themselves supplied critics with their most concrete ammunition. The project’s measures of class consciousness and class formation returned associations that were real but modest, varying by country in ways the schema alone did not predict, and critics argued that decades of coordinated research had established, at considerable expense, that class location explains less of political attitude than the tradition had assumed and less than rival variables such as education, sector and nation explained with simpler tools. Wright’s own reporting of these results was scrupulous, and his later theoretical work absorbed the lesson by rebuilding class analysis around mechanisms of exploitation and domination rather than around location alone, a rebuilding that conceded in practice much of what the critics had urged in principle. The episode cuts in two directions for the ledger. Against the school, it shows its one great empirical programme returning verdicts lukewarm toward the tradition’s expectations, which the school’s defenders must count as the method working and its critics as the method’s predictable yield. For the school, it shows the programme doing exactly what empirical programmes are for, and doing it in public, with data others could reanalyse. The internal charge about empirical thinness is a charge about quantity and institutionalisation, not about the survey’s quality, and conflating the two is one of the standard errors this article exists to prevent.

A further criticism, pressed by sociologists outside both camps, questioned whether survey instruments can capture class in the relational sense at all. A questionnaire records attributes of respondents, their occupations, credentials, supervisory duties and attitudes, and assembles locations from the pattern of answers. The relational claim, that classes exist in and through their antagonistic interdependence, is not an attribute of any respondent and cannot be read off aggregated attributes without a theoretical bridge the surveys presupposed rather than built. Wright was aware of the difficulty and wrote about the levels of class analysis, structural, organisational and ideological, with a care that acknowledged how much of the classical concept escaped his instruments. The acknowledgement is the school’s pattern at its best: a limitation identified, named and priced by the practitioner himself. It is also, once more, the pattern’s shadow, since the pricing occurred in methodological asides while the programme’s public presentation continued to speak of measuring class tout court. Readers of the ledger should grade the survey as the school’s outstanding empirical credit and its rows on class as substantially answered by revision, while noting that the revisions and the caveats together amount to a long, honourable retreat from the schema’s original claims, conducted, as the school conducted everything, in the open and with the arguments shown.

Rational choice, games and the psychology the school presupposed

A distinct strand of classical criticism targeted the school’s use of game theory and rational choice modelling in the theory of class action and historical change. The models represented workers, peasants and capitalists as strategic agents weighing costs and benefits, and derived the conditions under which collective action, revolution or compromise would emerge. Critics objected on two levels. At the level of psychology, the calculating agent of the models bore little resemblance to the beings who actually make revolutions, whose conduct is saturated with loyalty, rage, habit, honour and misperception, and a theory that must assume those away to get started has assumed away its subject. At the level of politics, the models’ most publicised results, the fragility of collective action, the rationality of compromise, the instability of revolutionary coalitions, were charged with a systematic tilt: formalising a situation tends to privilege the equilibria visible from the status quo, and a generation of students taught that rebellion is a collective action problem may absorb a quietism no theorem strictly implies.

The school’s replies were partly concessions dressed as clarifications. Elster himself became the leading theorist of the limits of rational choice within the school’s orbit, writing extensively on emotion, irrationality and the ways agents bind themselves against their own future calculations, so that the psychology charge was answered from inside by complicating the agent until the models’ elegance thinned considerably. On the political tilt, defenders replied that a theorem has no politics and that the same apparatus models successful revolutions as readily as failed ones, the outcomes turning on parameters the historian must supply. Both replies have force, and both illustrate the article’s recurring pattern in a new register: the school’s framework survived its critics by absorbing them, at the price of the framework’s original crispness. A rational choice theory enriched with emotions, norms, endogenous preferences and bounded cognition is a better psychology and a worse slogan, and the school’s public identity never quite caught up with the enrichment, continuing to trade on a crispness its own best members had spent decades qualifying.

There is a methodological moral here that generalises beyond this school, and the article states it because the case teaches it so cleanly. A formal apparatus earns its keep by the surprises it generates, the results its users did not install and do not welcome. By that test the school’s game theoretic work scores genuinely well in places: the assurance and threshold structures it developed explained patterns of mobilisation and demobilisation that purely structural accounts had waved at, and its models of class compromise gave the comparative study of welfare states a grammar it still uses. The same test convicts the weaker stretches of the programme, where the model’s conclusion could have been written down before the model was built and the formalism served as ceremony rather than inquiry. Critics who condemned the apparatus wholesale missed the first set of cases. Admirers who celebrated it wholesale missed the second. The ledger’s discipline, row by row and case by case, exists because wholesale verdicts on methods are the one commodity this literature already produces in surplus.

The polemical record: what the hostile accounts get wrong

Because this article has prosecuted the case against the school at length, intellectual honesty requires a section for the errors of the prosecution’s popularisers. The hostile literature, in journals, pamphlets and the rougher corners of academic debate, repeats a set of claims about the school that the record does not support, and a reader who absorbs them will misunderstand the criticisms that do stand by lumping them with ones that do not. The first error is that the criticisms come only from classical Marxists defending dogma. In fact the most damaging criticisms, as this article has shown, came from members of the group and from analytical philosophers with no dogma to defend, and several classical Marxists, far from attacking the school wholesale, borrowed its tools while disputing its conclusions. The second error is that the school produced no empirical work at all. It produced little, which is the graded charge, and the difference between little and none is the difference between the survey project’s existence and its erasure; polemic prefers the erasure because it scans better. The third error is that the normative turn was a betrayal concealed from the tradition. It was a departure announced in prefaces and defended in print against named critics, and whatever else betrayal means it cannot mean that.

The fourth error is the mirror of the third: that the school’s members drifted away from Marxism because they were never serious about it, careerists wearing a fashionable label. The biographical record refutes the insinuation in detail. Cohen’s socialism was lifelong and cost him standing in rooms where it was unfashionable. Roemer designed institutions for an egalitarian order long after abandoning the label would have eased his passage. Wright spent decades building empirical instruments for a theory his profession rewarded him for neglecting. Whatever the school’s faults, want of seriousness is not among them, and critics who reach for motive when argument runs short repeat, in a coarser key, the school’s own worst habit of answering the weak version of a charge. The fifth error concerns the school’s alleged ignorance of the criticisms themselves. The record shows a group that institutionalised its own criticism, published its defeats and revised in public, a record several of its attacking traditions, with their unexamined founding texts and their expelled internal critics, would struggle to match. A hostile account that omits that record is not a stronger version of this article’s case; it is a weaker one, vulnerable to the first competent reply, and it discredits the standing charges by association.

Correcting the polemicists is not a courtesy to the school. It is a condition of the genuine criticisms being heard. Every exaggerated charge hands the school’s defenders an easy victory that they can mistake, and present, as a victory over the charges entire. The pattern is familiar from the school’s own conduct, which answered crude versions and let the answers stand for refined ones, and it would be a poor outcome if the school’s critics, armed with better arguments than its defenders ever faced from outside, threw them away by refusing the discipline of accuracy. The ledger’s rows are deliberately unexciting in their phrasing, sourced, graded, defeater attached, because excitement is what this literature has in surplus and accuracy is what it lacks. Readers who find the rows less satisfying than the pamphlets should ask themselves which document they would rather have read about a school they belonged to, and grade accordingly.

The dialectics question behind the method slogan

The asserted slogan denying any distinctive Marxist method deserves a fuller hearing than the internal charge alone gives it, because the claim’s target, dialectical method, had serious defenders whose arguments the school engaged only intermittently. The dialectical tradition, in its sophisticated forms, did not claim a licence to contradict oneself or to substitute metaphor for inference. It claimed that certain objects, social relations above all, are internally related, changing through their own tensions, and totalities whose parts cannot be understood in isolation without distortion, and that a method adequate to such objects must therefore privilege relation, process and contradiction over the static categories of an atomistic logic. Thinkers in that tradition argued that the school’s analytical sorting, which decomposed doctrines into discrete propositions and tested each in isolation, systematically destroyed exactly the content the dialectical method existed to preserve: a proposition about capital and a proposition about labour, assessed separately, can each look confused, while the claim that matters is about the relation that constitutes both, and relations do not survive decomposition into their relata.

The school’s counter case, when its members troubled to make it, was that internal relatedness is a metaphysical picture in search of an argument, that every allegedly holistic claim produced in debate had so far yielded, under patient pressure, either an ordinary causal story about situated individuals or a confession of vagueness, and that the burden of producing one clear, true, irreducibly relational explanation had been outstanding for a century. That challenge, issued in seminar rooms across two decades, met responses of varying quality. Some dialectical writers supplied candidate cases, the value form, the capital relation itself, the tendential laws, and argued them with a rigour the school’s caricatures had not led its members to expect. Others retreated into the claim that the demand for a case already begged the question, since cases are what an atomistic method sees. The ledger grades the exchange as unfinished rather than decided, and that grading is itself a criticism of the school, because a slogan announcing that no distinctive method exists claims a finished result. Had the school’s public position been the modest one, that no distinctive method had yet been demonstrated to its satisfaction, the internal charge on this row would fall away. The distance between those two formulations, the demonstrated and the satisfactory to us, is the distance the school’s standards existed to police, and on its most famous methodological commitment the school stood on the wrong side of its own line.

It is worth adding that the school’s practice was often better than its slogan. Cohen’s reconstructions of historical materialism are saturated with relational and process claims, stated in his own idiom. Wright’s class analysis, especially in its later forms, treats exploitation and domination as relational mechanisms before treating them as distributions. Roemer’s models, whatever their limits, model interactions and not merely aggregates. A reader who knew the school only by its work might conclude that it practised a supple methodological pluralism, while a reader who knew it only by its manifestos would expect a sterner uniformity than any member achieved. The gap between manifesto and practice is common in intellectual movements and is not, in itself, discreditable. It becomes an entry in this ledger because the school made the policing of such gaps its signature activity when the gaps were other people’s, annotating the tradition’s distance between its professed dialectics and its actual arguments with relish. Symmetry, the virtue this article has demanded throughout, requires the annotation to be returned.

Historians against the theory of history

A further body of criticism came from practising historians, whose complaint differed from the philosophers’ and deserves separate registration. The school’s theory of history, in Cohen’s reconstruction, offered a powerful schema: forces develop, relations correspond or fetter, transitions occur when fettering becomes acute. Historians of actual transitions, the long dissolution of feudal relations in Europe, the varied paths into capitalism across regions, objected that the schema’s fit to the record was achieved at a level of abstraction where almost any sequence could be redescribed to fit, and that the school’s members showed limited interest in the archival literatures where the redescription might be tested. The Brenner debate, although its central figure stood adjacent to the school rather than inside it, illustrated the friction: an account of agrarian class relations driving divergent economic outcomes, built from demographic and tenurial evidence, sat awkwardly with a forces first schema, and the ensuing exchanges showed the school’s framework absorbing the challenge by reinterpretation rather than by prediction. A theory of history that no archive can embarrass has traded explanation for insurance, the historians charged, in a phrase that echoes the school’s own vocabulary of testability with pointed intent.

The school’s reply leaned on the distinction between a theory’s framework and its applications. Historical materialism, Cohen argued, is a research programme’s hard core, a claim about what ultimately explains the broad shape of epochs, and hard cores are properly judged by the fertility of the work they organise rather than by single recalcitrant episodes. The reply invokes a respectable philosophy of science, and the ledger accepts that framework claims are not falsified like weather forecasts. But the reply sits uneasily beside the school’s treatment of rival frameworks, whose hard cores were seldom granted the same protective patience, and it leaves the empirical charge standing in a specialised form: the school’s flagship theory was a theory of history produced almost entirely without doing history, by a group that included no practising archival historian of transitions, and defended, when historians objected, by methodological distinctions its members wielded more aggressively against others. The pattern is by this point in the article a familiar one, and its familiarity is the finding. Again and again the school’s difficulties reduce to a single asymmetry between the standards announced for the tradition and the standards accepted at home, and the historians’ row differs from Elster’s only in arriving from outside the group while depending, for its force, on criteria the group had published.

Against this, the credit side should record that the school’s schema did organise genuine historical inquiry in others. Social historians took the primacy and development theses as hypotheses to be localised, qualified and sometimes discarded in the face of evidence, which is a more honourable fate for a theory than reverent citation, and the comparative literature on transitions is unthinkable without the questions Cohen’s formulation made precise. A framework can be simultaneously under tested by its makers and fertile in the hands of its users, and the ledger, true to its method, records both facts in different rows rather than averaging them into a blur.

What the school owed the archive, the field and the movement

Running beneath several rows of the ledger is a single complaint about evidence, which this section gathers into one statement. The school’s characteristic sources were texts and models. Its members read Marx closely, read one another closely, and read the mainstream literatures of their disciplines with professional competence. What they rarely did was generate primary evidence about the social world their theory described: no archives opened, no workplaces studied, no movements accompanied, no datasets built, the survey project always excepted. Classical Marxism, whatever its theoretical disorders, had been built by people who also compiled factory inspectors’ reports, organised unions, edited newspapers and wrote histories from documents, and the tradition’s empirical sediment, the Blue Books behind Marx’s chapters, the party inquiries, the sociological classics, formed a deposit the school drew on without replenishing. A reader can test the asymmetry concretely. The school’s bibliographies cite the tradition’s empirical monuments constantly. The tradition’s successors, compiling their own bibliographies a generation later, cite the school’s theoretical monuments and look elsewhere for evidence about capitalism as it is lived and recorded.

Defenders respond that replenishment is a collective task and the school’s shift of the tradition’s centre of gravity toward theory was a legitimate specialisation, even a necessary correction to a culture that had prized reportage over argument. There is truth in the defence, and the ledger does not require every current to do everything. The criticism stands in a narrower and more wounding form. The school did not present its specialisation as a specialisation. It presented its standards as the standards of science itself, and its verdicts on the tradition carried the authority of a tribunal that had weighed the evidence. A tribunal that has not visited the scene should moderate its tone, and the school’s tone, in its founding decade especially, was not moderate. The gap between epistemic posture and evidential practice is the deepest version of the internal inconsistency, deeper than the functional explanation episode, because it concerns not one text but the school’s whole manner of proceeding. Elster’s mechanism demand, the school’s finest methodological product, is itself an evidential demand: show the process, do not merely name the pattern. Applied to the school’s own corpus, the demand asks where the processes are, in the workplaces, the transitions, the class formations the theory describes, and the corpus answers, mostly, with models of processes rather than with studies of them. Models are honourable instruments. They are not evidence, and a school that taught a tradition to distinguish patterns from mechanisms might have been expected to distinguish models from findings with the same care in its own self presentation.

The point should not be overdriven into philistinism. Theoretical work is real work, several of the school’s theoretical results are permanent gains, and the demand that every theorist also be an ethnographer would abolish the division of labour that makes modern scholarship possible. The ledger’s charge is calibrated accordingly: not that the school theorised, but that it theorised about evidence saturated subjects while generating little evidence and claiming scientific authority in full. Calibration matters because the school’s critics habitually overdrive the point into a romance of the archive, as though proximity to documents guaranteed theoretical virtue, a romance the tradition’s own history of learned sterility refutes. Between the romance and the school’s posture lies the defensible position, that a current’s authority should be proportioned to the evidence it commands, and it is a measure of the school’s unevenness that this proportioning, obvious once stated, must be supplied by its critics.

The uses of this case beyond Marxism

The brief for this article notes that the methodological questions it raises arise wherever a tradition imports a formal apparatus, and the point deserves development because it explains why the school’s criticisms matter to readers with no stake in Marx. Economics underwent the same drama when optimisation and equilibrium modelling entered fields built on institutional description: the apparatus clarified, and it also decided, in advance, that whatever could not be represented as constrained choice would count as noise. Political philosophy underwent it when the contract device and later the original position reorganised questions about obligation and justice into questions about agreement under stipulated ignorance, gaining precision and losing, for a generation, the vocabularies of virtue, history and power in which those questions had previously been thought. Ecology, linguistics and parts of history have each hosted a version of the quarrel between the formalisers and the custodians of the untranslatable residue. In every case the same paired questions arise that this article has pressed on the school: what does the apparatus import along with its clarity, and do the importers obey, at home, the standards they announce abroad?

The episode examined in this article is an unusually instructive case for three reasons. The participants stated their standards with unusual explicitness, so the audit has fixed criteria to work with. The participants published their internal disagreements and revisions, so the audit has a record to work from. And the stakes were doctrinally central rather than peripheral, since the claims under test, exploitation, class, historical direction, were the tradition’s core rather than its ornaments. A teacher of methodology in any field can build a term around the episode for exactly those reasons, and several have. The lesson the case teaches is not that formalisation fails, since the school’s formalisations include permanent achievements, nor that traditions should refuse imported tools, since the refusal merely preserves confusion at the price of purity. The lesson is bookkeeping: keep a register of what the apparatus leaves unexpressed, apply the announced standards to the founding texts first, and treat every verdict, including the verdicts in this article, as carrying the evidence that would overturn it. A tradition that keeps those books can import any apparatus safely. A tradition that does not will eventually be audited by its own members, on the record, as this one was.

There is a final, quieter use. The school’s story offers students of intellectual life a rare documented instance of a group changing its collective mind in public, claim by claim, with reasons. Universities are full of schools that changed their minds silently, by retirement and recruitment, leaving the record to suggest a continuity that never existed. The September Group’s members instead left a paper trail of conversions, qualifications and second thoughts that historians of ideas can follow step by step. Critics who cite that trail as evidence of dissolution should consider what the alternative record, the silent kind, would have concealed, and whether a tradition that conceals its changes is really in a stronger position than one that publishes them. The ledger’s verdicts on the school are severe in places, and they are possible only because the school kept the evidence. Severity purchased with the subject’s own candour carries an obligation of gratitude, which this article discharges by recording, one last time before its closing sections, that the audit itself is the school’s invention, applied to the school.

The classical rows revisited in detail

A table compresses, and compression hides the texture of the arguments it summarises. This section and the next unpack the ledger’s rows at the length their difficulty deserves, beginning with the classical rows, so that a reader who uses the table for reference can also see the reasoning that fixed each grade. The individualism row first. Its refined form, stripped of polemic, asserts a conditional: if preferences and identities are substantially formed by the relations a theory seeks to explain, then a method that must take them as given will systematically understate the relations’ causal depth, whatever its local successes. The school’s literature contains the materials for an answer, since its members wrote on ideology, on adaptive preferences and on the formation of wants under deprivation, Elster’s work on sour grapes being the celebrated instance where a mechanism of preference formation is placed at the centre of the analysis. That work complicates the row’s verdict without overturning it, because it remained a specialism of one member rather than a programme of the school, cited by the others more often than it was built upon. The row therefore stands as partly answered, with the answering concentrated where the school’s division of labour happened to place a psychologist of genius, and the unanswered remainder concerning the school’s collective practice rather than any individual’s insight. A defeater for the row would require evidence that preference formation stood near the centre of the school’s collective agenda, in its shared texts and its members’ main lines of work, and that evidence is not what the record shows.

The exchange row repays similar care. Its crude form accused Roemer of endorsing the market’s self image, and that form is false, as the article has recorded. Its refined form asserts a claim about representation: a framework whose primitives are agents, endowments and exchanges can state facts about distribution with unmatched precision, and can state facts about the constitution of agents by relations only by bolting on additional structure that the framework does not supply from its own resources. Every attempt within the school to discuss domination, authority or the experience of subordination confirms the diagnosis by its method, since each proceeds by leaving the exchange frame and importing concepts, command, hierarchy, dependence, whose home is the relational vocabulary the frame was built to replace. The importations are not dishonest; they are admissions. A framework that must import its rivals’ vocabulary at exactly the points where the rivals said the substance lay has conceded the refined charge in practice while denying it in manifesto. The row stands accordingly, and its defeater is correspondingly precise: a demonstration, from within the exchange framework’s own resources, of the constitution of agents by the relations they enter, without imported vocabulary, would overturn it. No such demonstration has been produced in the literature this article surveys, and the invitation to produce one is recorded here in the same spirit as the empirical defeater, as a standing offer rather than a rhetorical flourish.

The exploitation and class rows travel together and can be unpacked together. Both began as charges that a relational phenomenon had been replaced by a distributional shadow of itself. Both met replies of real power: that the shadow could be measured, compared and acted on, while the phenomenon, in its classical telling, resisted every attempt at disciplined statement. And both were then substantially answered from within, not by the replies but by the revisions, as Roemer rebuilt his egalitarianism on foundations where domination regained its own heading and Wright rebuilt his class analysis with exploitation and appropriation as its working mechanisms rather than locations alone. The ledger’s grades, standing against the original ambition for exploitation, substantially answered by revision for class, encode a distinction between the two cases that readers sometimes miss. Roemer’s revisions abandoned the claim that the property relations definition captured Marx’s concept, replacing it with the claim that the definition captured an injustice worth caring about, which is a change of thesis rather than a defence of it. Wright’s revisions, by contrast, retained the thesis that class analysis explains the distribution of material advantage and life chances, and changed the machinery, which is what answering a criticism while keeping one’s project looks like. The distinction matters beyond these rows, because movements under criticism constantly substitute the first manoeuvre for the second and call both of them learning. A reader trained by this article to ask, of any celebrated revision, whether the thesis survived or was exchanged, owns a diagnostic tool of general application, and the school’s record, which contains honourable instances of both manoeuvres, is the ideal training set precisely because its practitioners documented their moves so openly.

The labour process row is the simplest in the table and the least argued about, because the fact it records is visible in any bibliography of the school. Its interest lies in the explanation, and the explanations on offer divide into the sociological, the instrumental and the temperamental. The sociological explanation points to the group’s composition: philosophers and theorists, formed in disciplines where the workplace is a topic for other departments. The instrumental explanation points to the frameworks: exchange and game models have no interior, and a school works where its tools cut. The temperamental explanation points to the school’s founding quarrel, which was with the tradition’s theorists, in texts, and which therefore defined the tradition’s failings as textual failings, to be repaired at the desk. The three explanations are compatible and probably all contribute, and their conjunction yields the row’s real lesson. A school’s sense of what its tradition lacks is formed by the quarrels it happens to have had, and those quarrels are accidents of biography dressed, in retrospect, as diagnoses. Classical Marxism’s actual lacks were many, and the school’s list, obscurity, missing mechanisms, absent normative theory, was a true list as far as it went. It simply was not the whole list, and the items it omitted, the unexamined interior of production, the untested theory of history, the unbuilt empirical programme, were omitted because nobody in the founding circle had quarrelled with them. Criticism, even at its most rigorous, sees the faults it was formed to see. The ledger, compiled from the school’s critics in several traditions, is this article’s attempt to widen the formation, and its rows should be read as a map of several formations’ blind spots correcting one another, the classical critics strong exactly where the school was weak and weak exactly where the school was strong, which is why neither camp’s account can be adopted whole.

The internal rows revisited in detail

The internal rows require a different kind of unpacking, because their evidence is documentary in a stricter sense: the standards are quotable, the founding texts are fixed, and the revisions are dated. The functional explanation row turns on a comparison any reader can perform. Set the criteria for legitimate functional explanation that the school’s methodological writings endorse, a specifiable mechanism, a feedback loop connecting effect to persistence, an alternative to intentional design where design is implausible, beside the founding text’s actual practice, which offers consequence laws, a biological analogy and a plea that elaboration may come later. The mismatch is not subtle and does not depend on contested interpretation, which is why the critique that exposed it could be mounted from inside the group without importing a single foreign premise, and why Cohen’s response took the form of revision rather than rebuttal. What later scholarship added was context rather than rescue: philosophers sympathetic to Cohen showed that consequence explanations have legitimate uses under conditions the founding text had not stated, and historians showed that some of its claims could be restated as mechanism bearing narratives. Both developments improve the theory. Neither alters the row, because the row concerns the founding defence as delivered, judged by the standards as proclaimed, and on that comparison the verdict was reached by the school’s own referee a generation ago and has not been appealed successfully since.

The method slogan row is subtler, because it concerns a claim whose evidence is scattered by design. The school’s defenders can point to hundreds of pages in which allegedly distinctive Marxist reasoning is examined and found to be ordinary reasoning, good or bad, and they can ask, reasonably, how much accumulation a negative claim requires before it counts as argued. The row’s answer, partly standing, reflects a judgement about proportion: the slogan’s scope is total, covering all possible methods and all future proposals, while the accumulation covers the proposals the school happened to examine, and the school’s own standards, applied to a totalising claim by an opponent, would have demanded either a restriction of scope or a different evidential posture. The point is small in isolation and large in pattern, because it is the pattern’s clearest instance at the level of the school’s identity. The slogan was the school’s badge, the sentence by which it announced itself, and badges are precisely the claims movements exempt from audit. That this school, of all schools, exempted its badge is the internal inconsistency in miniature, and Elster’s later career, in which he subjected rational choice itself, his own badge within the badge, to corrosive examination, shows the exemption being corrected by the same member who corrected the founding text, a biographical rhyme the historians of the school have not failed to notice.

The empirical row is the one whose defeater this article has stated most formally, and the unpacking should explain why the formality matters. Charges of empirical thinness are easily made and easily waved away, because thinness is a matter of degree and every party keeps its own measuring rod. The defeater discipline converts the charge into a testable shape: name one further programme of empirical inquiry, comparable in scale to the survey project, that the school created and then maintained over time, and the judgement falls. The terms are chosen with care. Generated excludes work by sympathetic outsiders that the school merely cited. Sustained excludes single studies and programmatic essays, requiring the institutional marks of a programme, shared instruments, cumulative findings, successors. Beyond the survey project concedes the one case that plainly qualifies and asks for a second. A critic confident in the charge should welcome the formulation, since it tells the school’s defenders exactly what would change the verdict and tells future researchers exactly what to look for in the archive of the school’s activities. A defender who finds the formulation unfair owes an alternative one, stated with equal precision, and the exchange of formulations would itself be progress of the kind the school existed to promote. Until either the programme is identified or a better test is proposed, the row stands, and it stands as the row with the widest consequences, because a school’s empirical record is where its scientific self image meets the world.

Taken together, the three internal rows describe a single underlying condition with three symptoms, and naming the condition plainly is this section’s remaining task. The condition is not dishonesty, bad faith or secret hostility to Marx, the charges of the pamphlets, all of which the record refutes. It is the ordinary, well documented tendency of intellectual groups to grade their own homework leniently at the points where their identity is invested, operating inside a group whose public mission was the abolition of exactly that tendency. The school’s distinction is not that it escaped the tendency, since it plainly did not, but that it built institutions, the seminar, the published reply, the revised edition, through which the tendency was repeatedly caught and corrected by its own members. The internal inconsistency is therefore best understood not as a scandal but as a measurement: of the distance between a standard proclaimed and a standard achieved, in the one modern school that gave its critics the instruments, the criteria and the documentary record required to take the measurement at all. Other schools have larger distances. They are harder to measure, because they kept fewer records and proclaimed vaguer standards, and the comparison should temper any reader’s temptation to treat this ledger as a ranking of virtue. It is a ranking of auditability, and on that ranking the school stands, paradoxically and creditably, first.

Sources, attribution and the ethics of criticising a school

A word about how the criticisms in this article were sourced is owed to the reader, because the school’s own standards include a rule about attribution that assessments of the school often violate. The rule is that a criticism should be attached to the strongest identifiable holder of it, stated in the form that holder would recognise, and answered in that form or not at all. Polemical writing on all sides of this quarrel violates the rule industriously, attributing to the school positions taken by its least careful popularisers, attributing to classical Marxism the slogans of its least careful defenders, and refuting the attributions with applause. This article has instead attached the individualism charge to the critics who argued it in books, the functional explanation charge to the member who proved it in print, the empirical charge to the documentary record of programmes built and not built, and the school’s replies to the texts in which its members actually made them. Where a charge circulates without a single canonical holder, as the labour process charge does across a whole tradition of workplace sociology, the article has said so and named the tradition rather than inventing a spokesman. Readers who check the attributions will find disagreements about emphasis. They should not find a criticism that nobody held, aimed at a position nobody defended, which is the standard this article set itself and the standard by which it asks to be judged in its turn.

The ethics of the exercise deserve one paragraph more, because criticising a school whose members are identifiable scholars, several of them still active in their disciplines during the period this series covers, imposes obligations that criticising a safely distant doctrine does not. The obligations observed here are three. Criticise the published work and the public record, not the persons, their motives or their private commitments, except where the public record itself makes biography the subject, as with the trajectory’s use by both camps. State the subject’s achievements in the same document and at comparable length, so that the criticism cannot be excerpted into a hatchet job without visible amputation. And attach defeaters, so that the criticised retain a specified route of reply that the critic has committed, in advance, to honour. Those obligations are not courtesies. They are the conditions under which criticism functions as a contribution to inquiry rather than as a factional instrument, and they are, once again, the school’s own conditions, learned from its practice at its best and applied to its practice at its worst. A reader who finishes this article and reaches for the school’s own texts to check its fairness will have understood it exactly as intended.

There remains the question of what a reader should do with an assessment of this kind, beyond agreeing or disagreeing with its rows. The article’s answer is implicit in its form. The ledger is an instrument, and instruments are for using. A student can use it to structure an essay and to learn the defeater habit. A teacher can use it to stage the seminar exercises described in the persona sections. A researcher can use its defeaters as a research agenda, since each one specifies a discovery that would change a standing verdict, and a verdict with a discovery attached is a project waiting for an owner. The empirical defeater is the most obviously ownable: a systematic census of the school’s empirical production, programmes, datasets, field studies and their successors, conducted to a stated standard, would either overturn the row or put it beyond further anecdotal dispute, and either outcome would be a contribution. The exchange defeater invites a formal project. The functional explanation row invites a philosophical one. Criticism that ends in projects has done its work, and a school that provokes such projects, even posthumously as a school, is not finished being useful.

Analytical Marxism and the philosophy of social science

The school’s quarrels radiated outward into the philosophy of social science, and the criticisms it received there form a dossier distinct from both the classical and the internal ones, since its judges in that arena shared its analytical commitments while disputing its applications. Philosophers of social science pressed the school on three fronts. The first concerned methodological individualism’s exact content: whether it is a claim about ontology, about explanation or about evidence, since the school’s writers slid among the three, defending the modest evidential version under pressure and wielding the imperial ontological version in polemic. The second concerned the status of its models: whether a model that deliberately falsifies, assuming perfect information, frictionless exchange or omniscient agents, can yield explanations of a world where the falsified features do causal work, a question the school’s economists answered with the standard instrumentalism of their discipline and its philosophers found less comfortable than the economists did. The third concerned functional explanation’s aftermath: whether the mechanism standard the school championed could itself be stated without remainder, since mechanisms are typically identified relative to a level of description, and the demand for mechanisms all the way down threatens a regress the school’s writers halted, in practice, at the level of individual action, a halt their critics called convenient rather than principled.

These philosophical criticisms matter to the assessment because they show the internal inconsistency was not a local accident of one founding text. The pattern, a standard announced in imperial form, defended in modest form, applied in the imperial form to opponents and in the modest form at home, recurs across the school’s methodological writings with a regularity that suggests a structural cause. The cause is not mysterious. A school defines itself against a tradition it regards as lax, and definition requires contrast, and contrast rewards the imperial form of every standard, since modest standards make modest contrasts. In the seminar, among friends, the modest forms do the work and the imperial forms are understood as banners. In print, against opponents, the banners are quoted. The school’s tragedy of unevenness is thus partly a tragedy of genre: it wrote manifestos like a movement and papers like a seminar, and its critics, reasonably enough, judged the papers by the manifestos. A school aware of the trap might have written its manifestos modestly. None does, which is why the trap is worth naming in an article that students of other movements will read: the standard by which a school is eventually judged is the standard of its banners, and the only protection is to make the banners true.

Set against the philosophical dossier, the school’s defenders can claim one clear victory that should be recorded at length. The mechanism standard, whatever its foundational puzzles, won. Across the social sciences, the demand that pattern claims be accompanied by process accounts moved from a school doctrine to a default expectation of referees, editors and examiners, and the school’s internal quarrel was the standard’s most visible advertisement. Traditions are sometimes most influential through their defeats, and Cohen’s founding text, by failing the standard conspicuously and honourably, did more to entrench the standard than a flawless text could have, since a flawless text would merely have complied. There is a lesson in that for the school’s critics as well. The classical tradition’s own founding texts are not short of conspicuous failures under their own criteria, and the failures have seldom been as productive, because the tradition lacked the institution, the internal referee with standing and a journal, through which a failure becomes a standard. Productivity of failure is a property of institutions, not of texts, and it is the school’s most transferable institutional lesson.

Feminist, postcolonial and other neighbouring criticisms

The classical and internal dossiers do not exhaust the criticisms the school received, and completeness requires registering the charges that arrived from neighbouring critical traditions, even though the brief for this article assigns them a smaller place. Feminist critics observed that the school’s agents were curiously ungendered: the endowed individuals of its models carried assets and preferences but no bodies, no caring obligations, no positions within households, and its class schemas located persons by their relation to production while treating the organisation of reproduction, the labour that produces and sustains labourers themselves, as background. The charge parallels the labour process charge in structure, an interior rendered invisible by the frame, with the household standing to the exchange model as the workshop stood to it, and it was pressed by scholars who noted that a tradition which had always struggled to see domestic labour had found, in analytical reconstruction, a method that made the invisibility rigorous. The school’s egalitarian wing eventually engaged questions of gender and the family with distinction in the normative period, Cohen’s and others’ writings on equality within households among them, but the engagement arrived as moral philosophy rather than as a repair of the class and exploitation frameworks, where the omission had done its work.

Critics writing from the study of empire and race lodged a cognate complaint. The school’s histories and models were built overwhelmingly on European materials and European transitions, its classes were the classes of the North Atlantic economies, and its theory of history, whatever its universal grammar, was calibrated on a provincial record. The tradition the school inherited had its own long, contested engagement with colonialism, and the school’s reconstruction largely set that engagement aside as belonging to the political rather than the theoretical Marx, a sorting decision that critics read as continuous with the method’s other exclusions: whatever the frame could not process was reassigned to a neighbouring genre where it would trouble the theory no longer. Defenders replied that a theory of capitalism’s logic is not obliged to double as a history of the world, and that the school’s concepts travelled to other settings in the hands of scholars from those settings, who adapted them as they saw fit. Both points have force, and the ledger does not give these charges rows of their own, because the brief’s architecture assigns the centre to the classical and internal dossiers. They are recorded here so that the article’s silence about them is a decision rather than an oversight, and so that readers pursuing them know the shape of the terrain: in each case a frame’s expressive limit, an omission unargued rather than defended, and a later partial repair conducted in a different genre from the one where the damage occurred. The recurrence of that shape across feminism, the study of empire and the labour process is itself a finding about the school, arguably the deepest the classical direction affords: its instruments decided, in advance and in silence, which interiors of social life would count as sites of theory, and the list of the excluded is longer than any of its critics initially said.

The school’s defenders at full strength

The article has quoted the school’s collective reply in its closing argument, but the individual defences mounted by sympathetic scholars deserve a section of their own, since an assessment is only as good as the opposition it survives. The strongest defence of the school’s method remains the one implicit in its results: point to the questions that could not be asked before it and cannot be unasked after it. Before the school, a Marxist could write a book about exploitation without stating what would count as its absence. After Roemer, that omission is visible as an omission, and every subsequent writer, including the classical critics in this article, writes in the space the school cleared, using distinctions, mechanism, definition, counterfactual, that the school either coined or made compulsory. The defence generalises the point into a claim about intellectual credit: a current should be judged by the quality of the discourse it makes possible, including the discourse of its enemies, and by that measure the school’s balance is strongly positive, since even the charges in this ledger are stated in a grammar the school taught its critics.

A second defence concerns the comparison class. The school is habitually judged against an idealised classical Marxism, rigorous, empirical, unified and wise, that never existed as a going concern. The actual tradition of the same decades contained its own asserted slogans, its own unexamined founding texts, its own empirical programmes that returned verdicts their sponsors declined to publish, and its own trajectories, of which the least said the better. Judged against the actual contemporaries rather than the ideal, the school’s profile changes: its internal criticism was more institutionalised, its revisions more candid, its standards more explicit, its empirical flagship more rigorous than the corresponding features of most rival currents. This defence does not touch the ledger’s rows, which grade the school against its own proclaimed standards rather than against rivals, and the article’s method was chosen partly to keep the comparison class honest. But the defence properly tempers the tone in which the rows are read. A school can fail its own high standards and still stand above the field it competed in, and both facts belong to a complete assessment, the first in the ledger and the second in this paragraph.

A third defence, the most philosophical, disputes the article’s central framing by denying that uneven application is the right name for what the record shows. On this view, standards in a living research programme are themselves under negotiation, and what looks like exemption is often the programme discovering, through hard cases, what its standards should be. The mechanism demand was sharpened by the functional explanation quarrel rather than merely violated by it. The method slogan was a hypothesis whose testing constituted much of the school’s work. The empirical shortfall reflects a judgement, defensible at the time, that the tradition’s confusions were conceptual and had to be cleared before measurement could mean anything. A programme, the defence concludes, is not a court, and grading its history as though its standards were statutes mistakes the nature of inquiry. The article’s answer has been given in pieces and is gathered here: the defence succeeds against a legalistic reading of the ledger and fails against the ledger as written, because the school itself insisted on the courtroom model when the defendants were classical. It demanded of the tradition exactly the statutory grading, claim by claim against announced criteria, that its defenders now call naive when the school stands in the dock. Symmetry is the article’s whole method, and the defence, to succeed, would have to show that the school graded the tradition by programme standards all along. The school’s polemical record shows the opposite, in quotable abundance, and so the framing stands, with the defence recorded as the strongest available and as a permanent caution against reading the ledger more legalistically than its own defeaters allow.

A note on the school’s texts and their afterlives

The criticisms in this article attach to texts, and the texts have had afterlives that complicate any simple verdict, since a book’s influence and its soundness are different variables that polemics habitually confuse. Cohen’s founding volume remains in print, assigned in courses on Marx, on explanation and on the philosophy of history, read by many who reject its theses because its arguments are the ones a serious reader must pass through. Elster’s reconstruction is still the first port of call for readers who want Marx’s claims sorted by a hostile intelligence of the first order, and its errors, where later scholarship has found them, are the errors of a strong reading rather than of carelessness. Roemer’s formal work entered textbooks of political economy where Marx’s name seldom appears, carrying the school’s questions into rooms its members never entered. Wright’s schema became an instrument of a discipline, used by researchers who could not name its author, which is either the final dissolution of Marxism into sociology or its most complete practical victory, depending on the reader’s theory of what survival means. An assessment of what the school got wrong should not be embarrassed by these afterlives. Soundness and influence part company often enough in intellectual history, and the ledger grades the first. But the afterlives do discipline the assessment’s rhetoric: whatever the school got wrong, it got it wrong at a level of quality that kept its books in the argument for decades, and a criticism that cannot explain the persistence of the criticised is missing something about its object.

The school’s own retrospective literature, memoirs, interviews, anniversary assessments and the autobiographical writings of its members, adds a final source of evidence that this article has used sparingly and should acknowledge. In retrospect the members were often their own best critics and occasionally their own worst, settling scores and polishing legends in the manner of all memoirists. The retrospections confirm the ledger’s central rows from the inside: the founding quarrel over functional explanation is remembered as the group’s making and its wound, the empirical shortfall is conceded with a frankness the polemical years seldom permitted, and the normative migration is described by its protagonists as a discovery about where their questions had been leading all along. They also supply the one criticism this article has not made and should record as the members’ own: that the school, for all its talk of mechanisms, never explained itself, never produced an account of its own formation, cohesion and dissolution that met its standards for explaining anything else. A group of theorists of collective action that could not theorise its own collectivity, of class formation that left its own formation unexamined, offers a last, gentle instance of the pattern the ledger documents, the instruments pointed everywhere except home. The members who noticed the omission treated it with the school’s characteristic remedy, by saying so in print, and the saying is where this article’s evidence properly ends: with a school that supplied, to the last, the materials for its own audit.

Case study: the incentives argument and the school’s self criticism

One episode from the school’s later, normative period deserves treatment as a case study in how its internal criticism worked when it worked best, because it shows the virtues and the limits of the pattern in a single exchange. Cohen’s critique of the incentives argument, the claim that talented people must be offered unequal rewards to perform socially necessary work, proceeded by distinguishing the senses in which such incentives might be necessary: necessary because the talented would otherwise refuse, necessary because no alternative motivation could be cultivated, necessary because an ethos of equality would cost more than it gained. The distinctions were the school’s method at its finest, and they were turned, in the ensuing literature, on Cohen’s own earlier positions with the same fineness, as critics asked whether his defence of equality relied on a picture of motivation his own distinctions should have interrogated, and whether the socialist ethos he recommended could survive the collective action analysis his school had taught everyone to apply. Cohen engaged the objections in print, revised the argument’s scope, and left the exchange in a state where every participant’s position had been improved by contact with the others. Nobody who reads that literature can doubt the school’s standards were real, operative and capable of wounding their holders.

The case study’s lesson for the ledger is double edged. On one side, it acquits the school of the pamphlet charge that its standards were a costume worn for opponents: here the standards governed an internal exchange among the school’s own leading figures, on the school’s own later ground, with revisions following. On the other side, it sharpens the question the internal rows ask, which is why the same discipline arrived so unevenly across the school’s history, rigorous in the normative seminar and lax in the founding manifesto, exacting about incentives and casual about the method slogan. The answer the record suggests is institutional rather than personal. Criticism flourished where the school’s members faced one another as equals on ground none of them owned, and faltered where a text or a slogan functioned as the group’s common property, defended by loyalty before it was examined by argument. Founding texts and badges are the common property of every movement, and every movement grades its common property leniently; the school’s distinction is that it eventually graded its own, in several cases definitively, and published the grades. The case study therefore ends where the article’s assessment ends, with the internal inconsistency real, documented and partly self corrected, and with the correction standing as the school’s best claim on the attention of readers who share none of its doctrines.

A smaller episode from the same period reinforces the point from the opposite direction. When Roemer’s market socialist designs were criticised for their information and incentive demands, the most effective critics included economists applying the calculation arguments that this series examined in the cluster on critiques of Marxism, and Roemer’s engagement with them was serious, technical and inconclusive in the honourable way. The school’s members, whatever their faults at home, never developed the besieged enclave’s habit of treating outside criticism as persecution. They answered referees in economics, philosophy and sociology on those referees’ own ground for four decades, and the habit is worth naming because it is rarer than it should be and because it explains the school’s influence better than any doctrine does. Doctrines date. Habits of engagement propagate, through students, journals and appointments, long after the doctrines that carried them have been abandoned by their own authors, and the school’s habits, clarity, mechanism, revision in public, are by that measure its most durable export, visible in corners of the social sciences whose inhabitants have never heard the September Group’s name.

Case study: class consciousness and the evidence that answered back

The survey project’s findings on class consciousness offer the cleanest case in the school’s record of evidence answering back, and the criticisms the episode generated illuminate both the school’s empirical credit and its characteristic limits. The classical expectation, inherited from the tradition’s theory of class formation, held that location within production relations would generate, through experience and organisation, a corresponding awareness of interests and opponents. The project’s instruments measured location with unprecedented care and awareness with the standard tools of survey research, and the associations returned were weaker, more variable across countries and more entangled with education, religion and national institutions than the expectation required. Critics of the school drew the moral that the entire location based approach had been a category error, mistaking a sociologist’s map for a mechanism of formation. Defenders, Wright foremost among them, drew a different moral: that formation is a process with mediating institutions, parties, unions, media, churches, whose strength varies and whose variation explains the pattern, so that the findings refined the theory by specifying its scope conditions rather than refuting it.

The exchange is a model of how empirical criticism should work and seldom does, because both morals were argued from the same published data by scholars who could inspect one another’s instruments. Its lesson for the ledger is twofold. First, the episode is the standing answer to any claim that the school feared evidence: its flagship programme built the instrument that embarrassed the tradition’s favourite expectation and published the embarrassment. Second, the episode exposes the limit that the internal empirical charge actually names. A programme of this quality existed in one corner of the school’s empire, around one member’s vocation, and its methods and morals were never generalised across the school’s other territories. No equivalent instrument was built for the theory of history, none for exploitation’s measurement, none for the formation of the preferences the models took as given. The school contained, in Wright’s practice, a complete demonstration of what its scientific self image required, and the demonstration remained a local exhibit. Institutionalising that practice, hiring for it, training students in it, building the shared instruments that let findings accumulate across a school rather than within a career, was the step the school never took, and the step’s absence is the precise content of the empirical row. Programmes are not sums of projects. They are projects plus succession, and succession is an institutional act that a dining circle of peers, however brilliant, is structurally unlikely to perform. The school’s sociology, criticised in an earlier section on other grounds, returns here as the explanation of its most consequential omission, and the explanation, once more, explains without acquitting.

What a fair syllabus would teach about the school

One practical test of an assessment is whether it can be converted into teaching without distortion, and the criticisms of the school convert unusually well, because their structure, two directions, a ledger, defeaters attached, is already pedagogical. A fair syllabus would begin with the school’s own self presentation, its manifestos and its founding text, read in the school’s own spirit, which means read for the arguments. It would proceed to the internal audit, Elster on Cohen, the mechanism standard generalised, the method slogan examined, since students grasp a school’s standards fastest when they watch the standards bite their authors. Only then would the classical dossier be introduced, in its refined forms, with the warning this article has repeated: that each classical charge identifies a real cost and claims, sometimes falsely, that the cost was decisive. The credit side would be taught as content and not as consolation, the clarity demand, the collective action problem, the normative reconstruction and the survey, each presented as a live instrument students can pick up. The trajectory would be taught last and as a lesson in evidence rather than in history, with the undecidability result derived in class from the two hypotheses, since a student who can derive it owns a piece of reasoning that transfers to every controversy where partisans quote a pattern at one another.

Such a syllabus would also teach the school’s failures of omission as methodologically instructive rather than merely regrettable. The labour process silence becomes a lesson in how frameworks select their phenomena. The preference formation debt becomes a lesson in where models hide their assumptions. The empirical thinness becomes a lesson in the difference between a standard and a programme. Students educated this way would be equipped to audit the next school that arrives announcing a reformation, including schools whose doctrines the teacher favours, which is the entire point of teaching method rather than allegiance. The school’s members, at their best, said they wanted exactly that outcome, a generation of readers who would apply the instruments without respect of persons. The deepest criticism available of the school is that it did not always want the outcome when the persons were its founders. The deepest compliment available is that its instruments survive the wanting, and that a syllabus built on them produces readers who can grade the school itself, this article included, by standards none of the parties can bend.

Objections to this article’s method, answered

An article that grades a school by its own standards should expect, and welcome, the objection that the device is rigged: that selecting the school’s criteria guarantees a verdict of inconsistency, since no movement in history has applied its criteria perfectly, and the interesting question is comparative, how the school fares against rivals graded by the same device. The objection is serious and the article answers it in three parts. First, the device was not selected by the critic for its yield; it was selected because the school made its criteria its public identity and its warrant for dismissing a tradition, so the criteria are the ones the school itself put into contention. Second, the article has in fact run the comparative check the objection demands, in the section on the school’s defenders, and reported the result: graded against contemporaries rather than against its own banners, the school’s record on internal criticism, candour and revision is stronger than most, a result that tempers the ledger without altering its rows. Third, perfect application is not the standard the ledger applies. Its rows record not the existence of gaps, which are universal, but their location and their fate: gaps at the founding text, the central slogan and the empirical core, persisting unexamined for years in a school whose mission was examination. Location and persistence, not mere existence, are the charges, and a rival school with gaps at its periphery, promptly audited, would receive a different ledger from the same device.

A second objection holds that the article’s severity toward the school’s rhetoric, its answering of crude charges in place of refined ones, its manifestos outrunning its papers, applies a standard of argumentative hygiene that no polemical literature meets and that the classical critics violated at least as often. The premise is true and the conclusion does not follow. The article grades the school’s hygiene by the school’s standard because the school claimed hygiene as its distinction; it grades the critics’ hygiene in the section on the polemical record, where the hostile literature’s errors are catalogued at length, and it declines to let either side’s lapses cancel the other’s, since cancellation is the arithmetic of camps and addition is the arithmetic of ledgers. A third objection, the most fundamental, questions whether the school’s standards were ever as unified as the article’s device requires: whether clarity, mechanism and testability formed one creed or a family of emphases that different members held in different strengths, so that the internal inconsistency dissolves into the ordinary diversity of a group. The record answers the objection in the school’s own voice. Its members presented the standards jointly, in shared volumes and collective self descriptions, and wielded them jointly against the tradition. A group may not claim the authority of a united creed in attack and the indulgence of diversity in defence. The device stands, and with it the article’s verdicts, each carrying the defeater by which a better reading of the record could overturn it.

The question of exploitation’s wrongness, revisited as criticism

The normative turn’s critics and defenders both appeal to a question the school forced into the open, and the question’s career illustrates the school’s mixed record in miniature. The classical tradition had treated exploitation as a scientific category whose condemnation required no separate argument: to show that surplus was extracted was, somehow, already to have shown the extraction unjust. The school’s reconstruction dissolved that somehow. Once exploitation was defined through asset distributions, its wrongness visibly depended on the justice of the distributions, and once it was defined through labour transfers, its wrongness depended on a theory of entitlement to one’s labour that the tradition had never stated. Critics of the school complained that the dissolution was a debunking trick, dissolving a moral certainty into a technicality. Defenders answered that a certainty which cannot survive the request for its grounds was never knowledge, and that the school’s service was to hand the tradition the bill for its moralising, itemised at last. The ledger’s position, developed in the article on whether exploitation’s wrongness is a moral matter or a technical one, is that the defenders are right about the bill and the critics are right about a residue the bill omits. The wrongness of exploitation, as workers and their movements have actually experienced and protested it, includes the wrong of subordination within production, of time commanded and dignity taxed, and that wrong is not a distributive judgement wearing overalls. A theory that prices the distribution and leaves the subordination unpriced has answered the question it knows how to ask. The school’s later egalitarians, to their credit, rebuilt accounts of domination and self ownership that priced more of the residue, and the rebuilding is recorded in the credit column. The criticism stands against the reconstruction as announced, which claimed to have preserved the concept’s critical content while discarding the part of the content that the concept’s users had always cared about most.

There is a broader pattern here that students of the school should learn to spot, because it recurs in every row where the grade is partly standing. The school’s clarifications characteristically proceeded by selecting the aspect of a concept that its instruments could process, demonstrating with genuine rigour that the selected aspect behaved differently from the tradition’s account, and allowing the demonstration to be received, by readers and sometimes by the authors, as a verdict on the concept entire. Exploitation became distribution. Class became location. History became a correspondence between levels. Consciousness became survey response. In each case the selection was defensible and the rigour real, and in each case the unselected remainder did not cease to exist merely because the instruments passed over it. The tradition’s defenders, sensing the remainder’s existence, responded with a fury disproportionate to their articulacy, since they could feel what had been dropped and could seldom say it in the new grammar. The school’s members, mistaking inarticulacy for emptiness, concluded the remainder was nothing. Both errors are instructive, and an assessment that wishes to be fair must hold the two sentences together: the selections were achievements, and the remainders were real. Much of the best subsequent work in this field, by scholars formed in the quarrel’s aftermath, consists of attempts to bring the remainders back under disciplined treatment, domination, formation, experience, without surrendering the clarity the school won. That work is the school’s unintended legacy, and it is a finer one than either camp’s official story allows.

Generations: the school’s students and the second wave

A school’s record includes the work of those it trained, and the criticisms directed at the school’s second generation complicate the picture in ways the ledger should acknowledge. The September Group’s members taught widely, and their students carried the analytical approach into political science, sociology and philosophy departments across several countries, where it met fields with their own standards and their own orthodoxies. The second wave’s work was often more empirical than the founders’, more institutionally embedded and less concerned with Marx as a text, a development the classical critics cited as the trajectory repeating itself in the next cohort and the school’s defenders cited as the approach maturing into normal science. Both readings inherit the undecidability this article has established for the original trajectory, and the same discipline applies: a pattern consistent with dissolution and with maturation is evidence for neither until discriminating evidence is produced, and the discriminating evidence in the second generation’s case would concern whether its empirical work continued to test Marxist claims or merely used Marxist vocabulary to decorate inquiries whose questions came from elsewhere. The record is mixed enough to resist summary, which is itself the honest finding.

Critics of the school have also pressed a charge about succession that parallels the empirical row. A school that valued institutions of criticism built few institutions of its own: no journal that outlasted its founding energy in the form its founders would have chosen, no training programme that reproduced its combination of textual knowledge and formal skill, no archive of its proceedings assembled to its own standards of evidence. The September Group remained what it began as, a circle of peers, and when the peers’ interests diverged the school did not so much end as evaporate, leaving its legacy in books and in other people’s departments. Whether that evaporation counts as a failure depends on a theory of what schools are for. If a school’s purpose is to perpetuate itself, the record is a failure plain and simple. If its purpose is to change the questions a field asks and then dissolve into the field it changed, the evaporation is the success’s signature, and the school’s habits, visible in disciplines that never cite it, are the proof. The article records the charge and the reply without a row, because the verdict turns on that theory of purpose, which the school never stated and its critics never defended. It is fitting that the school’s assessment should contain one question its own method cannot settle from the record, as a reminder that the ledger’s completeness is a discipline and not a metaphysical guarantee.

The school’s Marx: criticisms of its reading of the texts

A further classical dossier, quieter than the methodological one, concerns the school’s reading of Marx himself. The school’s members were, several of them, formidable readers, and Cohen’s and Elster’s books are dense with textual argument. Yet classical scholars charged the readings with a characteristic deformation: Marx was treated as a colleague in the seminar, his claims extracted from their polemical and historical settings, standardised into propositions and graded, while the features of his writing that resist that treatment, the irony, the rhetorical staging, the deliberate ventriloquism of opponents’ categories in the critique of political economy, were discounted as noise. The first chapter of Capital, on this charge, is not a set of propositions about value awaiting a referee; it is a performance in which the categories of bourgeois economy are made to indict themselves, and a reading that extracts the propositions and discards the performance has not clarified the chapter but replaced it. The school’s reply was its usual one and not weak: that a performance whose propositional content cannot be stated cannot be believed either, only admired, and that Marx himself, in his scientific moods, demanded to be read as making claims about the world that stand or fall by evidence. Both positions capture something real about a writer who was genuinely both a system builder and a rhetorician of genius, and the ledger grades the textual charge as contested, with a rider that matters more than the grade. The school’s readings, even where they deformed, were answerable, because they stated what they took the claims to be, and the classical readings, even where they preserved the performance, were often unanswerable for the opposite reason. Answerability is not the whole of fidelity, but it is the part of fidelity that lets a tradition correct its readings, and the school’s Marx, whatever its losses, could be argued with in a way the tradition’s Marx frequently could not.

The textual dossier contains one charge with a sharper edge, concerning selection. The school’s canon within Marx was heavily weighted toward the writings that suited its questions: the historical materialist prefaces, the value theory, the class fragments, while the writings on Ireland, on India, on the commune, on the state in its Bonapartist moods, received comparatively little analytical attention, although they contain some of Marx’s most concrete theorising about politics and empire. A reconstruction of a thinker’s doctrine from a selection of his texts will find in the selection the doctrine the selector brought to it, and critics argued that the school’s Marx, systematic, economistic, curiously apolitical in his theory of the state, was partly an artefact of the shelf the school chose to read. The school’s defenders could reply that every reader selects and that their selections were argued for, on grounds of centrality and influence, rather than smuggled. The reply holds for the practice and falters for the presentation, since the reconstructions were typically offered as accounts of Marx’s theory rather than of a Marx, and the indefinite article, dropped in the titles and the manifestos, is exactly the sort of small inexactness the school existed to catch in others. The rider generalises beyond this dossier. Again and again, the school’s substantive work survives its critics, and its self presentation does not, and the distance between the two is the internal inconsistency wearing its most everyday clothes.

Comparisons across the same decades: the fairness protocol in practice

The brief for this article requires that comparisons between a small school and the wider tradition be drawn over the same decades, on stated dimensions, and the requirement deserves a section demonstrating it, since most published comparisons violate it in the school’s disfavour or in its favour according to the author’s camp. Compare empirical production across the decades of the school’s activity, and the tradition’s column fills with the work of historians of labour, sociologists of work, students of peasantries and revolutions, geographers and anthropologists who identified with Marxism in some register and produced evidence in industrial quantities. Against that column the school’s single survey programme looks slight, and the internal row stands. Compare theoretical clarification across the same decades, and the columns reverse: the tradition’s theoretical production of the period is vast and largely unclarified, while the school’s is compact and dense with distinctions that everyone, including the tradition’s own clearer heads, came to use. Compare institutionalised self criticism, and the school’s column, the published internal audit, the revised editions, the conceded points, outweighs the tradition’s record of the same decades, in which internal critics were more often answered with organisational measures than with arguments. Compare connection to movements of the people the theory describes, and the tradition’s column fills again, with party schools, union education and popular publishing in which the school, a professors’ circle, barely figures.

The exercise’s point is not to tot up a score, since the dimensions are incommensurable and any total would be an artefact of the weights. The point is to show what the ledger’s severally graded rows already imply: that the question whether the school succeeded is ill posed until the dimension is named, and that most polemics on both sides proceed by naming the dimension silently, in the author’s favour, and reporting the result as a verdict on the whole. Classical critics who lead with the labour process and the movements are not lying; they are selecting. Defenders who lead with clarity and self criticism are not lying either. The article’s method, dimension by dimension with the dimension declared, is the only honest alternative to selection, and it yields the assessment’s characteristic shape: a school that failed the dimensions it did not value, excelled at the dimensions it did, and stands condemned chiefly where the dimensions it failed were ones its own creed obliged it to value, evidence and self audit among them. That shape is less satisfying than a verdict of triumph or dissolution. It has the compensating virtue of being true to the record, and of teaching the reader a protocol, declare the dimension, fix the decades, state the defeater, that will serve in controversies where the parties are less obliging than this school in publishing the evidence of their own shortcomings.

The state, politics and the school’s relative silence

Another omission charged against the school, related to the labour process row but distinct enough to merit separate treatment, concerns the state and political power. The classical tradition, for all its theoretical untidiness about the state, possessed a rich repertoire of concrete political analysis: Bonapartism, the exceptional regimes, the party forms, the long arguments about reform and revolution that filled its journals. The analytical school’s political theory, by contrast, concentrated on the design of just institutions and the logic of collective action, and had comparatively little to say about the state as an apparatus with its own personnel, inertia and interests, or about the strategic dilemmas of movements confronting it. Critics argued that the omission followed from the same individualist premises as the others: a state modelled as an aggregation of agents’ choices, or as a neutral arbiter in a constitutional design problem, loses the feature the classical analyses existed to capture, its relative autonomy and its material density, the way an apparatus staffed and funded in particular ways bends every intention that passes through it. The school’s work on constitutions and precommitment, Elster’s above all, did engage the bending problem from the side of design, asking how agents who foresee their own future weakness bind themselves in advance, and that work is political theory of a high order. It is nonetheless a theory of politics viewed from the constitution maker’s desk, and the view from the ministry corridor, the barracks and the picket line, the tradition’s native vantage points, remained largely unwritten by the school.

The charge’s grading follows the pattern of the labour process row with one qualification. As an observation about the corpus, it stands: the school wrote little political analysis of the classical kind, and its members’ interventions in the strategic debates of their time were individual rather than programmatic. As a claim about necessity, it overreaches in the familiar way, since nothing in analytical method forbids the study of apparatuses, and neighbouring analytical work on bureaucracy, parties and state formation shows the tools travelling to the territory without difficulty. The qualification concerns the school’s defence. Where the labour process omission could plead a division of labour with a flourishing neighbouring tradition, the state omission faced no such neighbour in good health during the school’s decades, the classical theory of the state being itself in a condition the school’s members had helped to diagnose as confused. To diagnose a field’s confusion and then decline to rebuild it, while rebuilding exploitation, class and history on either side of it, is a choice that invites the critic’s question what principle governed the selection. No principle was ever stated. The selection followed the members’ competences and the neighbouring disciplines’ fashions, which is to say it followed the sociology of the seminar rather than the logic of the doctrine, and the ledger records the omission as a further instance of the school’s agenda being set by formation rather than by argument, the very failing its method existed to prevent in the tradition it audited.

Technology, ecology and the forces of production reconsidered

The school’s founding theory assigned the development of the productive forces the leading role in history, and later critics, including some formed within analytical philosophy, pressed difficulties about that assignment which belong in a complete register of the criticisms. The development thesis requires the forces to be identifiable and their growth measurable independently of the relations they are said to explain, and critics argued that the school’s writings never supplied the required metric: forces as disparate as a plough, a power grid, a body of chemical knowledge and a form of work organisation do not compose a single magnitude whose rise and fall across millennia could carry the explanatory weight assigned to it. Cohen’s replies treated the thesis as a claim about a tendency grounded in human rationality and scarcity rather than in a metric, rational beings under conditions of want will generally seize improvements in their productive powers, and the reply has the school’s characteristic strengths and weaknesses in concentrated form. As psychology it is plausible. As the foundation of a theory of history it relocates the explanatory burden onto a premise about rational adoption that the school’s own later work on collective action and institutional inertia showed to be anything but automatic, since improvements are routinely blocked, forgotten, suppressed and misapplied by the very agents whose rationality was to guarantee their adoption. The founder’s premise and the school’s own subsequent findings stand in a tension that the school never formally acknowledged, and the tension is a small, perfect instance of the article’s theme: a programme whose parts were each rigorously maintained and whose joints were nobody’s responsibility.

Ecological critics added a further count that the school’s founding generation had not anticipated and its later members engaged only partially. A theory that explains relations by their tendency to develop the forces treats the growth of productive power as history’s engine and, implicitly, as its justification, and critics writing after the ecological costs of that growth became undeniable charged the school’s historical materialism with sharing the productivism of the civilisation it analysed. The charge is partly anachronistic when aimed at texts of the late nineteen seventies, and the article records it with that caution. It is not wholly anachronistic as a criticism of the school’s failure to revisit the premise in later decades, when the costs were in plain view and the school’s normative wing was rebuilding its egalitarianism on other foundations. A theory of history whose engine is the expansion of production, held by philosophers who were simultaneously theorising the limits of consumption and the claims of future persons, contained an unexamined contradiction between its historical and its moral wings, and the examination, when it came, came from outside the school. The ledger’s rows do not include a productivism charge, since the brief assigns the centre elsewhere, but the article’s catalogue of omissions would be incomplete without it, and its structure will be familiar by this point in the reading: a premise inherited, a tension unexamined, a correction imported. The school’s history suggests that its instruments, applied to the premise in time, would have found the tension without assistance. The criticism is that the application, once again, waited for outsiders to demand it.

On slogans: no bullshit and the rhetoric of reform

The school acquired in its later years a slogan, associated with Cohen’s teaching and his valedictory writings, that its enterprise could be summed up as a refusal of bullshit: of language designed to impress rather than to state, of concepts that gesture at depth to conceal vacancy, of the whole economy of prestige through which weak thought protects itself in the humanities. The slogan captured something true about the school’s practice and about the gratitude of generations of students who learned from its members that a sentence owes its reader a meaning. It also attracted a criticism that the article should record, because it is the internal inconsistency translated into the key of rhetoric. A slogan against slogans is still a slogan, and the school’s badge phrases, no distinctive method, no bullshit, bullshit itself being a term its popularisers wielded with more enthusiasm than definition, functioned in the school’s public life exactly as badges function everywhere: as membership signals that exempted their bearers from the examination they demanded of other people’s badges. Students carried the phrases into seminars as verdicts rather than as questions. Popular accounts of the school repeated them as though repetition were the argument the school had neglected to supply. The school’s senior members were mostly innocent of the vulgarisation and occasionally complicit in it, enjoying, as founders do, the warmth of a phrase that travels.

The criticism is minor in the ledger’s scale and is entered here at the end of the catalogue for a reason beyond completeness. It shows the internal inconsistency operating at the smallest possible scale, in a handful of words, which suggests the phenomenon is not a correctable oversight but a standing temptation of intellectual identity itself. Any group that defines itself by a virtue will convert the virtue’s name into a possession, and possessions are defended rather than practised. The school’s distinction, for the final time, is not that it escaped the conversion but that it noticed the conversion in others with unmatched clarity, practised the virtue at a level its rivals seldom matched, and left, in its internal critics, the instruments by which its own lapses could be, and in this article have been, measured. A reader who finds that combination, high practice, imperfect consistency, exemplary self correction, insufficient as a defence has understood the ledger correctly and weighted its rows differently from its author, which is the outcome the defeater discipline exists to make possible.

The ledger’s omissions: a confession

An article organised around a ledger should confess the ledger’s own limits before closing its argument, since the form that disciplines the assessment also deforms it in ways a reader should be told. A ledger requires rows, and rows require criticisms discrete enough to grade, which biases the assessment toward the school’s arguable propositions and away from its atmospheres: the feel of its prose, the temper of its seminars, the experience of being its student or its target, all of which shaped its influence as surely as any theorem and none of which a row can hold. A ledger requires verdicts, and verdicts freeze exchanges that in life remained fluid, with positions softening across decades in ways a single grade can only approximate by the device of dating the ambition against which the grade is given. A ledger requires defeaters, and defeaters privilege the kinds of evidence that can be specified in advance, documentary and formal evidence, over the slower kinds, the shifts of judgement by which a field quietly revalues a body of work across a generation. The school’s reputation has undergone several such shifts already, and the ledger’s rows, however carefully graded, are a snapshot of an assessment that the field will continue to revise without asking the ledger’s permission.

The confession is not a retraction. Every form of assessment deforms, and the alternatives to the ledger, the narrative verdict, the camp summary, the trajectory quoted as proof, deform worse and conceal their deformations. The ledger at least publishes its exclusions, in this section, and its conditions of defeat, in every row. It is, in that respect, the school’s own preferred instrument turned on the school’s assessment itself, and the recursion is deliberate. This article has argued that the school failed, in specified ways, to apply its standards at home. The article’s defence against the same charge is the one the school never quite made for itself: that the standards were stated in advance, the evidence was shown, the defeaters were attached and the omissions were confessed in the document rather than discovered by its critics. Readers who find omissions beyond those confessed are invited to name them with sources, in the school’s manner, and the invitation is not a formality. It is the article’s final act of fidelity to the method it has spent its length both using and auditing, a method whose whole point is that no text, this one included, is the last word its subject will receive.

A worked example: how one charge travels through the grades

To make the ledger’s method concrete, this section follows a single charge, the oldest in the classical dossier, through every grade it has received in the literature, showing how the article’s grading was reached and how a reader might reach a different one honestly. The charge that the school’s individualism begged the question against Marx began, in its earliest published forms, as an accusation of ideological contamination: the method was bourgeois, its practitioners were liberals in disguise, and its results could be discounted by its origins. Graded against the record, that form fails entirely, since the practitioners’ socialist commitments are documented and the genetic fallacy is a fallacy whatever camp commits it. The charge’s second form was methodological: individualism as a doctrine about explanation was said to be false, because structures possess causal powers irreducible to the actions of their occupants. That form generated a genuine philosophical literature, in which the school’s modest wing conceded the furniture of structures while defending a tracing constraint, and the charge’s holders conceded that many allegedly structural explanations did dissolve, under pressure, into stories about persons. The exchange ended in a narrowing on both sides that the polemics of the period barely registered. The charge’s third form, the one this article grades as partly standing, is neither ideological nor metaphysical but bookkeeping based: whatever the doctrine’s truth, the school’s practice incurred a debt, the formation of the preferences and beliefs its models consumed, and the debt was serviced by one member’s specialism rather than by the programme. A reader can grade that form differently from this article by disputing the debt’s size, by pointing to the school’s writings on ideology and adaptive preference and judging them a programme rather than a specialism. The disagreement would then be about evidence and emphasis, conducted in the open, which is all the ledger ever asks. What the worked example shows is that a charge is not one thing with one fate. It is a family of claims of decreasing crudity, and the history of a controversy is largely the history of the family’s refinement under fire. Assessments that grade the family by its crudest member, in either direction, are the standard failure this article was written to correct, in the school’s critics and in the school alike.

A worked example from the other direction: the mechanism standard’s career

Symmetry demands a worked example from the credit side, following an achievement through the grades with the same care. The mechanism standard began as a weapon in an internal quarrel, Elster’s insistence that functional explanations in social science specify the feedback through which an effect sustains its cause. Its first grade, within the school, was as a successful criticism: Cohen’s revisions conceded its force in part, and the philosophy of social science took the quarrel up as a live problem about explanation rather than as a family dispute. Its second grade came from the standard’s migration. Sociologists building analytical programmes of their own adopted the demand as a founding rule, distinguishing mechanisms from correlations and from covering laws, and a literature grew up classifying mechanism types, aggregation, situational and transformational, whose practitioners often had no interest in Marx and sometimes no knowledge of the standard’s origin in a Marxist quarrel. Its third grade is the one a fair assessment must also record: the standard, in careless hands, hardened into a ritual, in which the word mechanism is pronounced over a narrative as a blessing, and in which explanations are rejected for lacking a level of process detail that the rejecter’s own favoured explanations do not supply either. The school’s internal inconsistency, it turns out, is contagious: a standard exported becomes a badge elsewhere, and badges exempt their holders. The example’s lesson completes the article’s argument from the far side. The school’s standards were genuine achievements whose fate in the world repeated the school’s own fate in miniature, proclaimed in imperial form, practised in modest form, audited by nobody at home. That the pattern recurs wherever standards travel is the strongest reason for treating the school’s unevenness as a general phenomenon to be guarded against rather than a scandal to be savoured, and the strongest reason for the defeater discipline, which is the one device in this article designed to survive its own export: a verdict that publishes the evidence of its own defeat cannot harden into a badge without ceasing to be itself.

The school in the classroom wars: pedagogy as a battlefield of the criticisms

The criticisms of the school were fought out not only in journals but in classrooms, and the pedagogical front deserves a record because it shaped how two generations of students met Marx at all. In departments where the school’s members and students taught, Marx entered the syllabus as a set of claims to be examined, excerpted into articles and chapters whose arguments could be isolated, numbered and disputed in the seminar’s weekly hour. In departments formed by the classical tradition, Marx entered as a corpus to be inhabited, read whole and in sequence, with the seminar’s task understood as the recovery of a meaning that paraphrase would betray. Each pedagogy produced readers the other regarded as deformed. The analytical classroom’s graduates could dismantle an argument in minutes and sometimes could not say what it felt like to read Capital as its author arranged it, a loss their teachers regarded as the price of rigour and their critics as rigour’s reductio. The classical classroom’s graduates carried the corpus in their bones and sometimes could not state its central claims in a form an outsider could test, a loss their teachers regarded as fidelity and their critics as the original sin the school had been founded to correct. The quarrel was never resolved, because it was never only about Marx. It was about what a university owes a difficult text, immersion or inspection, and the school’s role was to have forced the question into the open in department after department, usually by the scandal of a reading list on which Marx appeared between an economist and a game theorist rather than between two historians.

The pedagogical record also supplies evidence relevant to the ledger’s rows, of a kind the bibliographies cannot. The school’s teaching was famously generative: its members supervised students who disagreed with them, examined theses that attacked the school’s positions, and wrote references for scholars whose work dismantled their own, a pattern attested across the memoir literature and entirely consistent with the internal audit culture this article has documented. A school confident enough to train its own critics has a claim on the virtue of intellectual generosity that sits awkwardly beside the pamphlet portrait of a clique. At the same time, the classroom was where the school’s unevenness reproduced itself most efficiently, since students absorb a teacher’s badges faster than a teacher’s arguments, and the slogans, no distinctive method, no bullshit, travelled through cohorts as passwords long after the arguments that might have earned them had faded from the carriers’ memory. Every teacher of a strong school faces that transmission loss. The school’s members, who theorised the mechanisms by which beliefs propagate through populations, might have predicted it of themselves, and the few who remarked on it in retrospect did so with the rueful air of men discovering their own theory’s application at home, a discovery this article has had repeated occasion to record in larger matters.

For the reader assembling a judgement, the classroom evidence cuts in the school’s favour on the question of character and is neutral on the questions of doctrine, which is where the ledger wants it. Character evidence matters because several of the hostile charges are, at bottom, character charges in methodological dress: that the school was a fraud, a career vehicle, a solvent applied in bad faith. Against those, the testimony of the taught, including the taught who became the school’s critics, is decisive, and the article has graded accordingly. Doctrine is another matter. A generous school can hold a false doctrine and an ungenerous one a true doctrine, and the ledger’s rows, which concern the doctrines and the standards, stand or fall on the published arguments alone. Holding character and doctrine apart, in both directions, is the last discrimination this article asks of its readers, and it is the one the polemical literature, on every side of this quarrel, has been least willing to perform.

Translations, receptions and the school abroad

The school’s reception outside the Anglophone world supplies a further body of criticism, and of evidence, that a complete assessment should register. In the German speaking lands, where Marxist scholarship possessed its own formidable traditions of textual and historical work, the school was received with a mixture of respect and bewilderment: respect for the rigour, bewilderment at the confidence with which its members announced discoveries, about exploitation’s logic or history’s structure, that German scholarship believed it had been debating, in other vocabularies, for generations. The complaint of provincialism was returned with interest. A school that read little of the continental literature, its critics there observed, had mistaken the confusions of its local tradition for the confusions of Marxism as such, and had rebuilt, at great expense of ingenuity, positions that the unconsulted literatures had already mapped, criticised and in places surpassed. The charge parallels the article’s other omission charges in structure, an unexamined inheritance treated as a blank page, and it has documentary support in the school’s bibliographies, whose linguistic range is narrow for a movement theorising a German thinker. The school’s defenders replied that argument travels by quality rather than by passport and that the continental literatures had their own unexamined inheritances, including a reverence for the founding texts that the school’s irreverence usefully disturbed. Both points stand, and the exchange’s residue is a caution about the geography of rigour: standards are universal in aspiration and local in formation, and a school auditing a world tradition from one province of it will find, as this one did, that some of its discoveries are parochial and some of its parochialisms are discoveries.

In Scandinavia the reception took a different shape, since Elster’s own base made the school a domestic affair, debated in the region’s journals as a contribution to social science generally rather than as a Marxist event. In France, where a rival analytical tradition of a very different temper had preceded it, the school arrived late and partially, its rational choice wing read against indigenous debates its members seldom engaged. In the Marxist cultures of Asia and Latin America, where the tradition’s centre of gravity was political and organisational, the school was often received as a curiosity of the metropolitan academy, interesting where its egalitarian philosophy spoke to distributive questions and remote where its methodological quarrels presupposed seminar conditions that barely existed locally. That reception history matters to the assessment for one precise reason. The school claimed to audit Marxism, a world tradition. Its audit was conducted, in the event, on the tradition’s Anglophone academic province, with the other provinces appearing mainly as data in the theory of history or not at all. The claim and the conduct diverge in exactly the manner of the method slogan, a total scope announced and a partial scope performed, and the divergence is the gentlest and most general form of the internal inconsistency this article has documented in sharper cases. Gentleness should not be mistaken for triviality. A school’s sense of the size of its subject shapes every judgement it makes, and the habit of mistaking one’s province for the world is the besetting sin of metropolitan intellectual life, of which the school, for all its self scrutiny, did not cure itself.

The school’s economics: criticisms from within the discipline it borrowed

The school borrowed its economics from the mainstream of its day, and a strand of criticism worth recording came from economists who asked whether the borrowing had been critical enough. The models the school built for exploitation and class assumed competitive markets, well defined property rights and agents whose optimisation could be taken at face value, at a moment when the mainstream itself was complicating each of those assumptions with work on imperfect information, incomplete contracts and endogenous institutions. Critics charged that the school had imported a simplified economics, already dated in the departments down the corridor, because the simplified version was the one in which Marxist questions could be posed cleanly. The charge has a precise technical form. In a world of incomplete contracts, the employment relation is a governance structure in which authority does real work, and the classical claim that the wage relation involves domination ceases to be a metaphysical residue and becomes a theorem about residual control rights. Analytical tools existed, in other words, to state the classical point rigorously, and the school’s early frameworks, built on the cleaner economics, had defined the point away before the better tools were consulted. Later work in and around the school absorbed the contract theoretic perspective, and the absorption, once again, moved the school’s economics toward the relational phenomena its first models had excluded.

The episode generalises into one of the article’s most portable lessons. A tradition importing a formal apparatus imports a vintage as well as a logic: the state of the lending discipline at the moment of borrowing, with its simplifications, its fashions and its blind spots, frozen into the borrowing tradition’s foundations. The lending discipline moves on. The borrowing tradition, whose identity is invested in the borrowed tools, often does not, and finds itself defending the mainstream of the borrowing decade against its successor, in the name of rigour. The school’s history shows the pattern in economics and, in the other direction, in philosophy, where its members were themselves the moving frontier and the tradition that borrowed from them inherited their vintage in turn. Critics of the school who treat its economics as a timeless essence make the mirror error. The ledger’s grades are dated for this reason, explicitly or by context, to the versions of the frameworks their holders actually published, and readers applying the article’s verdicts to the school’s later, contract aware work will find several rows softened, which is as it should be. An assessment that cannot soften is a pamphlet with footnotes.

Liberty, community and the values the frameworks could not hold

A criticism from within the school’s own normative turn deserves registration, because it shows the internal audit operating in the school’s final period and reaching a characteristically mixed result. As Cohen’s philosophy matured, its centre shifted from the distribution of goods to the quality of relations: to community, to the ethos by which citizens regard one another, to a socialism defined less by an allocation than by a way of living together in which market motives are displaced by mutual regard. Critics within egalitarian philosophy, several of them Cohen’s own students, asked whether the later work’s central notions could bear the analytical weight the school’s standards required. Community, ethos, mutual regard: these are exactly the kind of terms the young school would have sent back for definition, and the later Cohen, by his own cheerful admission in the camping trip essay and its companion pieces, was content to let them work partly by evocation, trusting the reader’s recognition over the definer’s precision. Admirers called the shift wisdom, the discovery that some goods are damaged by the demand for exact statement. Critics called it the school’s final irony, its founder ending where the tradition began, with a vision whose power depends on the unanalysed remainder. The article records the exchange without a row, because it concerns one man’s late philosophy rather than the school’s programme, and because both descriptions are true of the same pages. It belongs in the assessment as a coda to the internal story: the standards the school proclaimed were, in the end, relaxed by their own author at the point where his deepest commitments lay, which is the most human instance in the entire record of the pattern this article has traced, and perhaps the one that should most temper the severity with which the pattern is judged in everyone else.

What would change the verdicts: the defeaters gathered

The article has attached defeaters to its verdicts as it went, and gathering them in one place serves readers who want the assessment’s research agenda without the argument. The empirical verdict falls if a further programme of empirical research, substantial in scale and apart from the survey project, is identified that the school itself created and maintained. The functional explanation verdict falls if the founding defence is shown, from its text, to have satisfied the mechanism test its own school imposed on rival explanations, a showing that would have to reckon with the author’s own subsequent revisions. The method slogan verdict softens if a full dress argument for the claim, proportioned to its scope, is produced from the school’s writings rather than assembled by charity from its case work. The exchange verdict falls if the frame is shown to state the constitution of agents by relations from its own resources. The labour process and state verdicts fall if bodies of school work on those interiors are produced that the bibliographies have somehow hidden. The trajectory verdict, that the pattern is evidence for neither camp, falls if discriminating evidence is produced, evidence about the method’s handling of cases independently settled, of the kind the article specified. Each defeater is stated so that meeting it is a discovery and failing to meet it, after a serious search, is a confirmation, and the list as a whole is the article’s answer to the reader who asks what its author could possibly accept as refutation. Assessments that cannot answer that question are not assessments. They are allegiances with bibliography, and the school, whatever its failures, taught its readers to demand better of every text that claims their agreement, this one emphatically included.

The article’s route through the series

Readers who have followed the argument this far will want the routes onward stated in one place, since the assessment touches nearly every article in this cluster and its neighbours. The school’s own story, its founding, membership and contributions, is told in the the complete guide to the school, and the present article should be read as its critical companion rather than its replacement. The systematic comparison with the tradition it broke from is in the comparison with classical Marxism, which supplies the paired exposition that the classical charges presuppose. The exploitation reformulation at the centre of the exchange and domination rows is reconstructed in the reconstruction of Roemer’s exploitation theory, and the class framework and survey programme behind two further rows are treated in the account of Wright’s class analysis. The founding quarrel over functional explanation has its own full treatment in the full treatment of Cohen on functional explanation, where the mechanism standard’s career can be followed in detail, and the question that powered the normative turn, whether exploitation’s wrongness is moral or technical, is opened up in the article on the moral status of exploitation. A reader who works through that route with the ledger in hand will hold the cluster’s whole argument, the school’s construction and its audit, in a single view, which is the condition the series exists to create.

The measurement of a school: influence, soundness and the confusion of the two

Public discussion of the school habitually confuses three quantities that this article has tried to keep apart, and a section naming them may prevent the confusion from reimporting itself into the reader’s verdict. The first quantity is soundness: whether the school’s claims, graded against the evidence and its own standards, survive. The second is influence: whether other scholars took the claims up, built on them, taught them and cited them, a quantity measurable in syllabuses, citations and the migration of distinctions into common use. The third is fertility: whether the claims, true or false, generated further work of quality, including refutations. The school’s profile across the three quantities is uneven in an instructive pattern. Its soundness, on this article’s grading, is mixed, strong on the critical standards, weak on several constructive frameworks, weakest where its self audit lagged. Its influence is high and documented across disciplines that never shared its doctrines. Its fertility is very high, since its errors were the kind that organise research: a founding text whose flaw defined a standard, a class schema whose revisions mapped a field’s learning, a set of definitions whose counterexamples filled journals. Polemics proceed by substituting one quantity for another. Defenders cite influence when soundness is challenged, as though a widely used error were less erroneous. Critics cite unsoundness when influence is conceded, as though a refuted framework could not have taught a field its business. The ledger’s discipline is to grade soundness and let the other quantities be what they are, considerable in this case, and the reader who keeps the three apart will find that many famous quarrels about the school dissolve into parties reporting different quantities in the same vocabulary.

The distinction also clarifies the school’s place in the longer history this series recounts. Marxism’s schools are usually remembered by their doctrines, and their doctrines are usually the least durable part of them. What survives of the school examined here is unlikely to be its theory of history, its definitions or its schema, each of which the record has already partly overtaken. What survives is a set of habits and standards, the mechanism demand, the definition before the dispute, the revision in public, the defeater attached to the verdict, which have passed into the common practice of fields far from Marxism and which operate there anonymously, as good tools do. A school’s final grade, if the phrase means anything, is the quality of the practice it leaves running in other people’s work. By that grade the school stands high, and the height is compatible with every row of the ledger, since the practice it left includes, as its centrepiece, the practice of auditing schools. This article is itself an instance of the survival it describes, an audit conducted with the school’s instruments on the school’s record, and its author is content that the school’s defenders should cite that fact, as they are entitled to, in the school’s defence.

Counterfactuals: the school that might have been

A disciplined counterfactual can test an assessment, and the criticisms in this article imply one worth stating. Suppose the school had applied its standards at home from the first day: the founding text’s functional claims published with their mechanism debts itemised, the method slogan issued as a research hypothesis with its evidence to date, an empirical programme founded alongside the theoretical one with the survey’s methods generalised, the models’ expressive inventories printed in their prefaces beside the results. Would that school have achieved more or less than the actual one? The ledger’s method suggests a divided answer. It would have achieved less provocation, since the founding text’s overclaim and the slogans’ sweep were engines of the attention that made the school a force, and a school that itemised its debts in advance might have been filed, by a busy profession, as one more cautious research programme. It would have achieved more durability, since its frameworks would have been built with their remainders acknowledged and their successors trained, and its dissolution, when interests diverged, might have been a succession instead. And it would have achieved, most speculatively, a different relation to the tradition it audited, since a school visibly grading its own homework first earns the standing to grade others, and the classical tradition’s defensive fury, which consumed so much of the actual school’s energy, might have been disarmed by the spectacle of an auditor who began at home. The counterfactual cannot be run, and the article offers it not as a verdict but as a diagnostic: the distance between the actual school and the counterfactual one is the precise size of the internal inconsistency, measured not in sins but in forfeited standing. Schools that come after this one, in whatever tradition, can read the distance as a price list. The items are cheap individually, an itemised debt here, a modest slogan there, a programme founded beside the theory. Their compound interest, across four decades, was the difference between the school that was and the school its own standards described.

The school’s critics in the labour movement and the parties

One constituency of critics remains unrecorded, and its testimony, though seldom written in the journals this article has drawn on, shaped the school’s public meaning more than any review. The organised labour movement and the parties of the left received the school, where they received it at all, as a distant academic event whose conclusions reached them filtered through journalism and polemic, usually in the form that another group of professors had found Marx wanting. Trade union educators and party intellectuals who had sustained a practical Marxism through difficult decades resented the filtration and, often, the school itself, whose members appeared in the filtered version as the tradition’s undertakers. The resentment was unfair to the members’ intentions and commitments, several of whom had records of movement involvement the filtration erased. It was not baseless as a response to the school’s public effect, since a school is responsible in part for the foreseeable career of its conclusions in a world that reads headlines, and the school’s members, sophisticated theorists of how beliefs propagate, might have foreseen that a decade of publicised abandonments would function politically as a permission structure for abandoning more than theses. The classical journals made the point in a coarser idiom, speaking of objective service to reaction, a phrase this article will not adopt, since intentions and arguments cannot be graded by their echoes. The narrower point survives the idiom’s rejection. A current that understands mechanisms owes some attention to the mechanisms of its own reception, and the school’s attention to its reception was, by the standard of its other work, cursory: it published, it argued, and it left the filtration to chance. Whether scholars owe their publics that attention is a genuine question on which reasonable positions differ, and the ledger records the movement critics’ charge as belonging to that question rather than to the audit of arguments, noted here so that the catalogue of criticisms is complete and so that readers from the movement tradition find their objection stated in a form they can recognise as their own.

Final accounting: the rows summed without a gestalt

The article promised at its outset that its overall assessment would be nothing more than the sum of its rows, and the promise should be kept explicitly, since the temptation to round a ledger into a slogan is the temptation the whole document exists to resist. The classical rows sum as follows. The school’s frameworks converted relations into distributions in case after case, and the conversion cost content the tradition valued: domination within exploitation, constitution within class, the interior of production, the density of the state. The school’s replies defeated the crude forms of the charges and, in Wright’s case, answered a major charge by revision toward the critics’ ground. The refined forms stand, as statements about expressive limits and unargued omissions, and they identify the school’s characteristic intellectual shape more accurately than any summary in terms of doctrines abandoned. The internal rows sum more severely and more creditably at once. The school’s founding explanation failed its own mechanism standard, its central slogan outran its argument, its scientific posture outran its empirical production, and each failure is documented from the school’s own publications. The same publications document the corrections, partial, public and in several cases definitive, which no rival current of the period matched. The credit rows sum without strain: clarity, mechanisms, collective action, normative theory, the survey, and the culture of revision are permanent acquisitions, held by the field in common, owing their existence to this school’s work. The trajectory row sums to nothing, deliberately, and the article’s refusal to let it sum to anything is among its findings rather than a gap in them.

A reader who totals those sums will arrive at an assessment with a shape rather than a slogan: a school strong where it audited and weak where it built, except when it built instruments for auditing; a school whose frameworks clarified the surfaces of its subject and thinned at its interiors; a school whose deepest failing was the unevenness of a virtue it genuinely possessed. That shape will satisfy neither camp, and the article regards the dissatisfaction as evidence in the shape’s favour, since the camps’ satisfactions are manufactured by selection and the record, read whole, selects nothing. What the school got wrong is the article’s title and its question, and the answer, in the end, is narrower than the pamphlets say and more interesting: it got wrong the application of its own best idea to itself, for a while, in the places where its identity was invested, and then, uniquely among the schools of its tradition, it published the correction. Both halves of that sentence are the verdict. Neither half may be quoted without the other by anyone claiming this article’s authority, and the defeaters stand open, in the ledger and in the gathered list, for the reader who can do better with the evidence than the article has done.

The problem of the school’s name and the criticisms it invited

Even the school’s name attracted criticism, and the apparently trivial quarrel over labels encodes a substantive one worth recording. Analytical Marxism was a label its members accepted unevenly: some embraced it as a banner, others preferred descriptions of their work that omitted Marx altogether, and the wider world, including the September Group’s joking alternative names for itself, treated the label as a convenience rather than a creed. Classical critics seized on the naming instability as a symptom. A current unsure whether it is a Marxism, a method or a dining club, they argued, will predictably produce work whose relation to Marx is negotiable, and the trajectory merely ratified the label’s original vagueness. The school’s defenders answered that names in intellectual history are assigned by librarians and enemies as often as by founders, and that the work’s relation to Marx is a question about arguments, settled text by text, not about badges. The article grades the naming charge as rhetoric rather than argument, and records it for a different reason: the quarrel illustrates how much of the school’s reception was conducted at the level of identity, where its members were least defended and least interested in defending themselves. A school of argument, confronted by an argument about identity, tended to change the subject back to the claims, which was principled and ineffective in equal measure, since publics, including academic publics, grade identities first and claims afterward. The lesson generalises to any reader assessing a school from the outside: the volume of a controversy is a measure of the identities at stake, not of the arguments’ quality, and the ledger’s quiet rows are a deliberate refusal to let volume set the grades.

Reading Marx after the school: the audit’s permanent deposit

The criticisms examined in this article leave a residue in the practice of reading Marx that survives every verdict, and a section on the residue may help readers who came for the school and stayed for the author. After the school, certain questions are asked of any Marxist claim as a matter of course, by readers who could not name their source. What would count against this claim? Through what mechanism does the alleged cause produce its effect? Is the concept doing distributive, relational or moral work, and has the writer noticed which? Are the agents’ preferences being treated as given, and if so, whose problem is their formation? Is the verdict carrying its defeater, or is it a badge? These questions are the school’s permanent deposit in the common reader’s account, and they are asked, in seminars the school’s members never visited, of texts the school never discussed. A tradition is changed less by the answers it receives than by the questions it learns to ask, and by that measure the school’s intervention was among the most consequential in the tradition’s modern history, which is precisely why its own failures to ask those questions at home matter enough to fill an article. The deposit also sets the standard by which the article itself should be read. Its claims about the school are claims, with sources named and defeaters attached, and a reader who applies the school’s questions to the article, who asks what would count against its rows and whether its concepts are doing the work they announce, is not refuting the article but completing it. The audit, to be worth anything, must be auditable, and the school’s final gift to its critics is the standard by which the critics can be criticised in their turn.

A brief history of the criticisms themselves

The criticisms in this article have a chronology, and tracing it briefly protects the reader against a distortion the ledger’s synchronic form can create, as though all the charges arrived at once against a finished body of work. The first wave, in the years immediately after the founding texts, was dominated by the internal quarrel over functional explanation and by classical responses to Cohen’s theory of history, conducted in the journals of philosophy and in the tradition’s own reviews with a speed that testified to the founding text’s importance. The second wave, through the following decade, widened to the frameworks: Roemer’s definitions, Wright’s schema and Elster’s reconstruction each drew their dossiers, and the charges of conversion, relations into distributions, were formulated in the refined versions this article grades, largely by critics who had absorbed the school’s standards well enough to aim them precisely. The third wave accompanied the normative turn and was the most public, since the migration of the school’s leaders into egalitarian philosophy drew fire from the tradition’s journals and praise from philosophy’s, in a crossfire that fixed the dissolution narrative in the wider culture. The fourth wave, retrospective and still continuing in the memoir and anniversary literature, is the one this article chiefly belongs to, in which the school’s own members joined the assessment and the trajectory argument reached its stalemate. Each wave corrected the previous one’s crudities and introduced its own, and the ledger’s rows are best read as the sediment of that forty year process, with the grades reflecting the refined forms that survived the waves rather than the slogans that rode them. A reader who understands the chronology will also understand why the article has been so insistent on versions: in a controversy of this length, the version is the argument, and grading a charge without dating its form is like grading a framework without naming its edition, an error the school’s critics and the school committed against each other with perfect symmetry for decades.

Why this assessment matters for the cluster that follows

This article closes the cluster’s examination of the analytical school, and its method sets the terms for the cultural theory cluster that follows in the series, where the objects change and the audit’s questions remain. The schools of Marxist cultural theory and literary criticism that the next cluster examines made claims of a different kind from the analytical school’s, claims about texts, forms and consciousness that resist the mechanism standard’s simplest applications, and the temptation, for a series formed by the analytical example, will be to grade those claims by criteria their makers never professed. The lesson of the internal inconsistency cuts against that temptation with precision. A school is most damagingly and most fairly criticised by the standards it sets itself, and the cultural theorists set standards of their own, of textual attention, of historical placement, of fidelity to the work’s form, by which their practice can be audited as strictly as the analytical school’s was audited here. The transferable discipline is not any single standard. It is the procedure: find the standards the subject proclaims, state them in the subject’s own strongest form, measure the practice against them with the evidence shown, attach the defeaters, and record what the subject got right at the length its achievements deserve. Readers who carry that procedure into the next cluster will find it works on material as different from game theory as material can be, which is the final evidence this article can offer for its central contention, that the school’s method, turned impartially on any school including its own, is the most valuable thing the school produced.

The ethics of revision: what the school’s corrections teach

The school’s practice of public revision, praised at several points in this article, deserves one examination from the critical side before the assessment closes, because revision has its own pathologies and the school’s record exhibits mild forms of them that a complete catalogue should not spare. A revised position can be presented in ways that obscure the size of the change: a definition quietly narrowed between editions, a claim’s scope reduced in a footnote, a founding text’s later printings carrying qualifications the original’s readers never saw, so that the public record shows a smoother continuity than the argumentative history warrants. Critics charged the school’s members, Cohen in particular, with revisions of this quiet kind, in which the theory of history that emerged from a decade of criticism was substantially weaker than the theory that had been defended, while the weakening was distributed across so many essays and editions that no single document announced it. The charge is the mirror image of the school’s complaint against the tradition’s unmarked shifts of meaning, and it has enough documentary support to be entered in the ledger’s margin. Its force is limited by the school’s countervailing practice, the explicit retrospectives in which members did announce their changes and itemise them, and the balance between the quiet and the announced revisions is a matter on which the school’s bibliographers, rather than its polemicists, will eventually rule. The article records the charge because symmetry requires it and because it completes the portrait of revision as a practice with standards of its own: a correction should be as visible as the claim it corrects, in the same venue and at the same prominence, a standard the school met in its best moments and missed in its ordinary ones, which is to say it behaved, in revision as in everything else, like a school of human beings whose virtue was real and whose consistency was not.

On the uses of a defeated school

There is a question beneath the assessment that the article should surface before ending, because it explains why a series of this length devotes a full article to a school’s failures. Defeated and dissolved schools are the normal fate of intellectual movements, and most are forgotten without injustice. The analytical school repays its audit because its defeat was of the instructive kind, in which the instruments survived the doctrines and the manner of losing taught more than most victories do. Its members set out to save a tradition by examining it, examined it past the point where the tradition’s custodians could follow, and recorded the whole passage in documents that let later readers re walk the route and grade every step. A reader can disagree with the school at every row of the ledger and still owe it the ledger. That combination, maximal auditability, is vanishingly rare in the history of doctrine, whose schools generally arrange to be remembered by their slogans and forgotten in their arguments. The series this article belongs to has examined many schools and movements, and the contrast is instructive: for most of them, the criticisms must be reconstructed from hostile fragments and apologetic summaries, while for this one the criticisms lie in the open, in the school’s own journals and editions, awaiting only arrangement. Arrangement is what this article has supplied, together with the grades and the defeaters, and if the arrangement is judged severe, the severity should be credited where it belongs, to the school that made severity possible by keeping the evidence. There are worse epitaphs for an intellectual movement than that its critics needed nothing but its own books.

A last discrimination: failing a standard and failing a subject

One final discrimination protects the assessment from its most likely misreading, and the article ends its argument by making it. To fail a standard is not to fail a subject. The school’s founding text failed the mechanism standard, and the theory of history it defended remains a live option whose merits the failure did not settle, since an explanation can be true before its mechanisms are known, as the school’s founder insisted with the biological analogy, and as the history of science confirms in cases the school’s critics honour. The school’s frameworks failed to capture relations, and the phenomena of class and exploitation remain where they were, awaiting vocabularies adequate to them, some of which the school’s revisions began to supply. The school’s empirical record failed its self image, and the questions its theories posed remain open, awaiting the programme the defeater describes. In each case the ledger grades the school’s performance against its promises, and the grading leaves the underlying subject, Marx’s questions about capitalism, class and history, exactly as standing as it was before the school convened. Readers who take the article as a verdict on those questions have mistaken its jurisdiction. Its jurisdiction is a school’s consistency, and within that jurisdiction its verdicts are as firm as the documents allow. Beyond it lie the questions themselves, which the series’ other articles pursue, and which no audit of any school, however rigorous, has ever closed.

The school and its neighbours in the academy

The school did not work in isolation, and some criticisms of it are really criticisms of the company it kept, which a fair assessment should disentangle. Its closest neighbours were analytical philosophy, mainstream economics and the analytical wing of sociology, and its members’ standing in those fields was both its resource and its temptation. The resource was real: referees who owed Marxism nothing tested the school’s arguments as arguments, and the arguments that survived were stronger for the testing. The temptation was subtler: the neighbours’ questions gradually became the school’s questions, not by conspiracy but by the ordinary gravity of seminars, appointments and journals, until the distance between the school’s agenda and its parent disciplines’ agendas was hard to detect. Critics who call that gravity assimilation and defenders who call it integration are describing the same process with different valences, and the article’s standing discipline applies: the process is a fact, its evaluation depends on premises about what the school was for, and the premises were never agreed even among the members. What can be graded without those premises is the school’s awareness of the process, and there the record is mixed in the familiar way. Individual members wrote searchingly about disciplinary drift and about the costs of respectability. The school as a collectivity never made the drift an object of its method, never asked, with its own instruments, what mechanisms were selecting its questions, although it possessed, in its theories of institutions and incentives, every tool the question required. A school of mechanism that did not turn the mechanism question on its own situation exhibits, in its purest form, the asymmetry this article has documented from the founding text onward, and the purity of the instance is why the article mentions it last among the omissions: it is the omission that explains the others, the unexamined situation from which the unexamined interiors, the unbuilt programmes and the unargued slogans all followed, as naturally as the school’s own theories would have predicted, had anyone in the founding circle thought to apply them at home.

How to argue with this article

The article closes its argumentative sections with instructions for its own refutation, in keeping with the discipline it has praised and practised. A reader who believes a row is misgraded should begin by identifying the version of the charge the row grades, since most disagreements in this literature are version disagreements in disguise, and should then produce the evidence the row’s defeater specifies, or argue that the defeater itself is wrongly drawn and supply a better one. A reader who believes a charge is missing should name its holder, state it in the form the holder would recognise, and show that it engages premises the school professed or facts the school’s record contains, since charges meeting neither condition belong to the polemical register this article has deliberately set aside. A reader who believes the credit side is understated should name the achievement, its documentary basis and the test it passes, remembering that influence and soundness are different quantities and that the ledger grades the second. And a reader who believes the whole device, grading by professed standards, is misconceived should say what device would grade the school more fairly and apply it, in full, to the tradition the school audited as well, since a device that only ever acquits is as rigged as a device that only ever convicts. Arguments conducted on those terms are the school’s element and this article’s hope. Arguments conducted on other terms are available in abundance elsewhere, and the reader is free to prefer them, with the article’s blessing and its prediction that they will change fewer minds.

Envoi: the audit continues

The school’s story has no ending of the kind its first members imagined, neither the vindication of a rescued doctrine nor the clean dissolution its enemies announced, and the absence of an ending is itself the assessment’s last datum. Its questions are in other hands, its standards are in other fields’ refereeing, its errors are in the textbooks as warnings and its achievements as tools, and its members’ books continue to be argued with by readers who were not born when the September Group first met. An audit article in a series like this one is a small part of that continuation, and it should end by acknowledging the continuation’s terms. The verdicts recorded here will be revised by the evidence their defeaters describe, or they will harden into the kind of badge the article has warned against, and which of those fates befalls them depends on readers the author will never meet. The school’s founders, who built their enterprise on the proposition that claims should be exposed to exactly that uncertainty, would not have wished the article a safer fate, and the article does not wish itself one.

The school’s treatment of ideology: the charge that completes the formation debt

The preference formation debt recorded in the individualism row has a named address in the classical tradition, the theory of ideology, and the school’s handling of that theory attracted criticisms that complete the dossier on its omissions. Classical Marxism’s account of ideology, for all its notorious difficulties, addressed the question the school’s models bracketed: how the beliefs and wants of agents are shaped by the relations they inhabit, so that domination reproduces itself inside the preferences of the dominated. The school’s engagement with the topic was real but oblique. Elster wrote penetratingly on the mechanisms by which wants adapt to circumstances, the sour grapes phenomenon and its relatives, and Cohen examined the ways beliefs about legitimacy stabilise arrangements. Yet the school never assembled these fragments into the systematic account its own framework required, an account that would have connected the formation of preferences to the structures of class and property its models described. Critics argued that the omission was not accidental but architectural. A theory of ideology worthy of the name must treat consciousness as an effect of relations, and a method committed to taking agents as the unmoved movers of explanation experiences that treatment as a threat to its foundations, since the movers turn out to be moved. The school’s fragments on adaptive preference show the threat being managed piecemeal, one mechanism at a time, each safely local, while the general claim, that the givenness of preferences is itself a social product requiring theoretical centre stage, remained unassimilated.

The grading follows the article’s established pattern. As an observation about the corpus, the charge stands: ideology is the least developed major topic in the school’s work relative to its importance in the tradition and to its necessity within the school’s own explanatory scheme. As a claim about the method’s incapacity, it overreaches, since nothing in the analytical toolkit forbids modelling formation, and the fragments prove the tools can bite on the material when a member chose to aim them there. The interesting question is the one the school’s sociology again answers. Ideology was the tradition’s most jealously guarded territory, the ground on which the classical custodians felt strongest and the school’s founders felt least equipped, and a circle that chose its battles by its strengths left the territory alone. The ledger does not add a row, since the charge is the individualism row’s debt viewed from the creditor’s side, but readers should register that the debt had a name and an address all along, and that the school’s mail, on this subject, was largely left unopened.

On judging a school by its questions

A final consideration may reconcile readers troubled by the distance between this article’s severe rows and its warm credits, since the distance dissolves once the right unit of assessment is chosen. Doctrines are the wrong unit. A school’s doctrines are hostages to evidence and fashion, and most doctrines of most schools are, in the long run, abandoned, qualified or absorbed, as the analytical school’s were with unusual speed and documentation. Questions are the durable unit. A school that teaches a field to ask better questions has contributed something that the abandonment of its own answers cannot revoke, and the analytical school’s questions, about mechanisms, definitions, defeaters, formation and the price of clarity, are asked wherever social theory is seriously practised, by scholars who would mostly be surprised to learn their provenance. Judged by its doctrines, the school is a mixed and partly defeated enterprise, and the ledger has said so without flinching. Judged by its questions, it is among the most successful interventions in its tradition’s modern history, and the article has said that too, in the credit sections, at the length the record warrants. The two judgements are not in tension, because they concern different objects, and much of the confusion in the literature this article surveys comes from forcing one unit’s verdict to do the other’s work. Critics who proved the doctrines wanting announced that the school had failed. Defenders who pointed to the questions’ triumph announced that the criticisms had failed. Both announcements mistake the bookkeeping. The school failed at several of the things it built and succeeded at the thing it taught, and an assessment that keeps the units separate can honour the success without discounting a single row of the failure, which is the balance this article has attempted from its first page to its last.

The criticisms in one paragraph, for the reader in a hurry

A reader who needs the whole assessment compressed can take it in this form, with the understanding that compression is what the remaining sections exist to correct. The school was charged from outside with building its conclusions into its models: agents with given wants, exchange as the elementary act, class as a portfolio, exploitation without domination, a workplace left unexamined and a destination in liberal philosophy misdescribed as a homecoming. It was charged from inside with failing its own examinations: a founding explanation short of the mechanism test, a methodological badge worn in place of its argument, and a scientific banner flown over a largely untested body of theory. It answered the crude versions of the external charges, absorbed several refined ones by moving toward its critics, and never answered the internal ones except by the corrections its members published, which is to say it answered them in deed where it could not in word. It got right the demand that claims be stated so they can fail, the insistence that patterns produce their processes, the problem of collective action, the need for a moral theory its tradition had dodged, and the one great survey that showed what its standards could build. Its history of abandoned doctrines proves nothing for either camp, since honest inquiry and rigged method predict the same retreat. The rest is rows, and the rows are the article.

The assessment’s balance sheet in the profession’s terms

Professions keep their own informal accounts of schools, and the analytical school’s account in the wider profession offers a last external check on this article’s grades. In philosophy, the school’s members are remembered as the people who made Marx a presence in the discipline’s central conversations about justice, equality and explanation, and the memory is warm even among those who reject the doctrines. In economics, the memory is thinner and more technical, a set of definitions and a market socialist design that specialists respect and few teach. In sociology, the memory runs through instruments and datasets, the class schema’s long career in stratification research, where the school’s name attaches to a tool rather than to a doctrine. In history, the memory is the most distant, a theory debated for a season and set aside for want of archives. The pattern of the memories matches the ledger’s shape with embarrassing fidelity: strongest where the school’s contribution was a standard or a question, weakest where it promised a substantive science of society and delivered frameworks. Professions are careless bookkeepers, swayed by fashion and friendship, and their accounts would not convict anyone in this article’s court. As corroboration, though, the convergence is worth the paragraph it occupies. Independent audiences, grading by their own unrelated criteria across four decades, arrived at the distribution of strengths and weaknesses that the internal audit, conducted on documents and defeaters, reached by another road. Convergence of independent measures is the closest thing assessment has to confirmation, and the article rests its confidence there rather than on any single row.

Coda: severity and gratitude

The two sentiments this article has kept in suspension, severity toward the school’s record and gratitude for the candour that documented it, are not reconciled in a formula, and the coda will not attempt one. Severity was owed to the evidence, because a series that grades other schools cannot soften its instruments for the school that invented them. Gratitude was owed to the same evidence, because without the school’s habit of publishing its defeats, its revisions and its members’ second thoughts, the severity would have had nothing to work on except rumour and the recollections of enemies. A reader who finds the combination uncomfortable has located the article’s intended effect. The schools that deserve the longest audits are the ones that make them possible, and the audit’s length is, in the end, a form of respect that the pamphlets on both sides, with their easy verdicts and their unexamined trajectories, never paid. The ledger closes with its defeaters open and its rows available for correction by any reader willing to do the work the school taught: state the claim, name the source, show the evidence, and say what would change the verdict.

For students and exam candidates

Students should organise revision around the two directions, since examination questions on this school reward the distinction above all else. Learn the classical charges as a connected family: individualism, exchange, exploitation, class, labour process and the normative turn all express the single complaint that relations were converted into distributions. Learn the internal charges as a different family with a different logic: the founding explanation failed the mechanism test, the method slogan outran its argument, and the empirical record stayed thin beside the survey project. The internal inconsistency is the phrase examiners hope to see used accurately, meaning not hypocrisy but uneven application of professed standards. Practise attaching a defeater to every verdict you write, since a judgement that names its own condition of defeat reads as evaluation rather than summary. The comparison article in this series supplies the paired material most questions presuppose, and a VaultBook card reproducing the criticisms ledger, one row per charge with its verdict and defeater, converts this article into a revision instrument. Avoid the two symmetric errors: reciting the trajectory as proof of anything, and treating the school’s self criticism as a reason to dismiss its achievements.

For teachers and seminar leaders

The internal inconsistency teaches a general lesson that transfers to any syllabus: a method’s first honest application is to its own founding texts, and a movement should be graded on whether it permits that application. Seminar leaders can stage the point directly by setting Elster’s critique of Cohen beside the standards Cohen’s book announced, and asking students to find a premise in the critique that the target had not already endorsed. A second exercise uses the ledger: assign each row to a pair of students, one to prosecute and one to defend, with the rule that the defence must answer the refined version of the charge rather than the crude one, since answering the wrong version is the failure this article documents most often. A third exercise concerns the trajectory: present the sequence of abandonments without attribution and require students to state what additional evidence would be needed before the sequence could support any conclusion about the method. Discussion should close on the credit side, the clarity demand, the mechanism standard and the collective action problem, so that criticism of the school is learned as an achievement of the school’s own culture rather than as an external sport.

For researchers, historians, and journalists

Researchers will find the live questions in the gaps this article has marked rather than in its verdicts. The preference formation debt remains unpaid across the social sciences, and the school’s failure to pay it is an invitation: a research programme that modelled how relations shape wants, using the school’s own standards of mechanism and evidence, would answer the deepest classical charge from within the analytical camp. The general problem of formalisation, pursued further in the article on Marxism and microfoundations, awaits its comparative history, in which the analytical Marxist episode is the best documented case and economics itself the necessary second. Historians of the school have rich primary material, since the participants published their disagreements, their revisions and their retrospectives with unusual completeness, and the defeater attached to the empirical charge in the ledger specifies exactly the discovery, a sustained empirical programme beyond the survey project, that would most change the assessment offered here. Journalists covering the afterlife of Marxist ideas should resist the dissolution headline the trajectory tempts them toward; the accurate story is narrower and more interesting, a school that audited a tradition, was audited by its own audit, and left tools behind that its critics still use.

Frequently Asked Questions

Q: What did analytical Marxism get wrong?

By the grading this article applies, three things stand out. It applied its own standards unevenly, so that its founding defence of functional explanation, its central methodological slogan and its empirical ambitions all escaped the audit it imposed on the tradition. It kept poor expressive accounts for its models, offering partial vocabularies in the guise of complete replacements. And it announced a scientific reformation while delivering, apart from the class survey project, a body of theory largely untested against evidence. Each charge rests on the school’s own published record and is stated with the defeater that would overturn it, in the criticisms ledger above.

Q: What do classical Marxists object to in the school?

Classical critics object, in a connected family of charges, that the school’s method converts relations into distributions. Methodological individualism takes persons and preferences as given where Marx treats both as socially formed. Exchange models represent the wage relation as a transaction between independent owners. The reformulated exploitation concept cannot distinguish domination from inequality. The asset based class schema turns a relation into a possession. The labour process, Marx’s central site of analysis, disappears from view. And the normative turn, whatever its merits, departs from the tradition’s questions while claiming to complete them. Critics such as Callinicos and Wood pressed versions of these charges, and the article grades each in its refined form rather than its crudest.

Q: What fails by the school’s own standards?

Three commitments fail the audit the school set for everyone else. The founding defence of functional explanation offered consequence explanations without the feedback mechanisms the school’s criteria required, and a member, Elster, demonstrated the failure within four years. The claim that Marxism possesses no distinctive method of its own was asserted and repeated far more than it was argued to the standard the slogan’s scope demanded. And the school’s scientific self presentation was not matched by empirical production, the international class survey excepted. These are internal failures, engaging premises the school itself professed, which is why they bind more tightly than any charge arriving from an opposing tradition.

Q: Does the school ignore the labour process?

Substantially, yes. The school’s characteristic settings are markets, distributions and strategic games, while the interior of production, where Marx located the extraction of surplus and the formation of class experience, appears mainly as a black box converting purchased labour power into performed labour. Defenders plead specialisation among researchers, and the labour process tradition that descends from Braverman did flourish elsewhere in Marxism. The reply explains the omission without dissolving it, because the omitted material bears directly on the school’s own constructions of exploitation, class authority and collective action. The charge stands, explained by the school’s composition and tools rather than by any argued conclusion that the topic lacked importance.

Q: Was the normative turn a departure?

Yes, as a matter of subject matter, and the school’s honest members admitted as much. Inquiry migrated from the workings of capitalism and the direction of history toward the requirements of justice, the defensibility of distributions and the debts individuals owe one another. Cohen’s egalitarian philosophy and Roemer’s institutional designs proceed by the methods of liberal political philosophy and reach largely kindred conclusions, and their link to Marxism runs substantially through biography and motivation. Where critics overreach is in calling the turn a betrayal. It was declared, argued for and defended openly, and the school could reply that its own findings about exploitation forced the questions of justice it then pursued. Departure is the accurate verdict; condemnation requires further premises.

Q: Did the school produce empirical work?

It produced one substantial empirical enterprise and a great deal of programmatic statements about the value of evidence. The enterprise is the coordinated class survey programme that Wright built and led, which surveyed populations in several countries using the school’s own class schema, published its data and entered mainstream stratification research. Beyond that project, the record is thin: models were seldom tested against data, the theory of history received no comparable testing programme, and no data infrastructure outlived individual projects. The ledger’s verdict on this charge carries a stated defeater: it falls if a substantial programme of empirical research, beyond the survey project, that the school itself generated and sustained is ever identified, and no such identification has been made.

Q: What did analytical Marxism get right?

Five achievements survive both the classical and the internal audits. The demand for clarity deprived weak arguments of the shelter that specialised vocabulary had provided for decades. The critique of functional explanation established the mechanism standard that social science broadly absorbed. The collective action problem converted class organisation from a presumption into a research question. The insistence that Marxism needed a normative theory exposed a real incoherence and produced philosophy of lasting influence. And the class survey project stands as a genuine empirical achievement. To these should be added the school’s culture of public self correction, in which members revised their own claims in print as criticisms accumulated.

Q: Can the criticisms be answered?

Some can, some partly, and some have not been. Charges that target clarity as such are answered completely by the school’s reply that no tradition can plead vagueness as its defence. The modest version of methodological individualism answers the metaphysical form of the individualism charge. Wright’s revisions answer much of the class charge by moving the schema toward relations. The refined charges survive their replies: the expressive limits of the exchange frame, the labour process silence, the unargued scope of the method slogan and the thin empirical record all stand in at least substantial part. The ledger grades each row separately, because the criticisms do not share a single fate.

Q: Did the school dissolve Marxism?

No, and the metaphor misleads in both directions. A seminar circle lacks the power to dissolve a tradition carried by parties, unions, publishers and universities across many countries, and the school’s members neither intended dissolution nor performed it. What they did was examine central Marxist claims with instruments that found some claims indefensible, some concepts in need of replacement and some questions in need of new owners. Where their conclusions converged with liberal egalitarianism, the convergence records the path of their arguments, not an act of sabotage. Marxism continued, and continues in forms the school influenced even where it is rejected, since its standards of clarity were absorbed by the very traditions that opposed it.

Q: Is the trajectory evidence for the critics?

No, because the trajectory is equally consistent with the rival explanation and therefore discriminates between nothing. A method rigged against its object would yield a steady sequence of abandoned theses. So would honest inquiry directed at claims that were largely indefensible. Critics who quote the sequence as proof of bias, and defenders who quote it as proof of integrity, are both presenting an outcome that both hypotheses predict as though it favoured their own. Evidence that would genuinely discriminate exists in principle: the method’s handling of cases where truth is independently settled, or documents showing conclusions fixed in advance. Until such evidence is produced and weighed, the trajectory should be retired from the argument.

Q: How does the internal inconsistency grade the school by its own standards?

The internal inconsistency is the article’s central grading device. It names the fact that the founding defence of functional explanation the school published would not have passed the tests of clarity, mechanism and evidence that the school applied to classical Marxism, which is why a member dismantled it within four years using only shared premises. Grading by its own standards means every criterion invoked, explicitness, specified mechanisms, argued rather than asserted claims, testability, was first proclaimed by the school itself. A failure on home criteria cannot be dismissed as a foreign misunderstanding, and a school cannot decline the audit without surrendering the standards that constituted its claim to attention.

Q: What should a student separate between a criticism from outside the school and one from its own premises?

A student should separate the premises each criticism needs. An external criticism, such as the charge that exchange models presume voluntariness, depends on claims a classical Marxist accepts and an analytical Marxist may reject, so answering it requires defending contested ground. An internal criticism, such as the mechanism objection to functional explanation, depends only on premises the school itself published, so it binds regardless of which tradition the reader inhabits. In examination answers the separation earns its marks when it is applied row by row: state the criticism, name its source, identify the premises it requires, and only then grade it. Verdicts that skip the premise step read as allegiance rather than evaluation.

Q: Why did the school produce so little empirical research?

Several causes combined, and the article weights none of them as a complete explanation. The school’s personnel were philosophers, economists of theory and political scientists, trained and rewarded for the monograph and the article rather than the dataset. Its disciplines separated theoretical from empirical crafts. Its questions, how explanation should proceed, what exploitation coherently means and how class concepts should be built, were ones it judged prior to measurement. And as members’ conclusions drifted from Marxist theses, the motive to build testing programmes for those theses weakened. The class survey project shows what the school could do when one member made empirical work a vocation; the surrounding thinness shows that the vocation was individual rather than institutional.

Q: How does the criticisms ledger order the charges?

The ledger orders charges to make their binding force visible rather than to follow the chronology of the polemics. Each row fixes a criticism to a named source, a direction, classical or internal, the best reply the school can offer and a verdict carrying its defeater. Classical rows come first in the table because they are older and more numerous, but the internal rows weigh more in the assessment, since they rest on premises the school publicly professed and therefore allow the school no jurisdictional escape. Reading the ledger row by row, instead of adopting a camp’s summary of it, is the use for which the table was built, in study, in teaching and in further research.

Q: What separates a model that assumes its conclusions from one that derives them?

The separation lies in where the contested feature enters. A model assumes its conclusion when the feature under dispute is installed in the setup: agents arrive with given preferences, exchange is the elementary act, class is an asset portfolio, and the results then report what the setup contained. A model derives its conclusion when the contested feature emerges from elements that do not presuppose it, so that a different outcome was representable within the same frame. The test a reader can apply to any construction in this literature is to ask whether the frame could have produced the opposite verdict. Where it could not, the model has illustrated a premise rather than tested a claim.