The argument named by the phrase reform or revolution has organised socialist politics for well over a century, and it has done so in a peculiarly unhelpful form. Presented as a single choice between two roads, it invites loyalty rather than analysis. A participant declares for gradual change or for rupture, and the declaration is treated as a position, when in fact it bundles together at least four separate questions that have separate answers and separate bodies of evidence. Whether capitalism can stabilise itself is an economic question. Whether reforms can accumulate until the system has been transformed is a question about institutional change over time. Whether an elected government commands the power its office appears to confer is a question about the state. Whether reforms are worth winning for their own sake is a question about the value of improved lives under a system one hopes to replace. Nothing about answering one of these questions settles any of the others, and the history of the debate is largely the history of answers to one question being carried across, illegitimately, as answers to the rest.

Separating the questions is the work of this article, and separation produces a finding that the slogan version of the argument conceals. On most of the separated questions, the considered positions of serious reformists and serious revolutionaries converge far more than the polemic between them admits. Both sides hold that reforms are worth winning and that winning them improves lives. Both sides hold that elections can be won by parties committed to deep social change. The disagreement that survives separation is narrower, harder, and more precise than the famous quarrel suggests. It concerns one question above all: whether a government elected on a transformative programme can actually transform the ownership of productive wealth while working through the institutions it inherits, or whether those institutions, and the economic power standing behind them, will defeat, dilute, or expel any government that tries. That question is genuinely open. It has a large historical record bearing on it, and the record points in more than one direction. The three-out-of-four convergence, which names the pattern of agreement across the separated questions, is the claim this article sets out to demonstrate, and the demonstration requires giving each side its strongest form before any verdict is attempted on any part of the dispute.

A note on what this article is not. It is an analysis of an argument and its evidence, not a recommendation. The debate belongs to living politics, and readers will bring their own commitments to it, but the task here is cartography rather than recruitment: to draw the map of the argument accurately enough that a reader can locate any participant on it, identify which question a given claim answers, and judge what evidence would count for or against each position. Where verdicts are offered, they are confined to specific empirical sub-claims that history can test, such as whether a particular prediction about capitalist development was borne out. On the strategic question itself, which road a movement ought to take, this article delivers no instruction, because the evidence does not deliver one and because the choice depends on judgments about risk, value, and circumstance that no historical survey can make on a reader’s behalf.

Reform or revolution: one slogan, four questions

The slogan form of the argument was fixed in the late 1890s, when Eduard Bernstein, a leading theorist of German social democracy then living in exile in London, published a series of articles and then a book questioning central expectations of the Marxist tradition he had helped to codify, and Rosa Luxemburg, a young Polish revolutionary newly arrived in German politics, answered him in a pamphlet whose title supplied the enduring formula. Social Reform or Revolution appeared in 1899, and its title posed the choice as an either and or. Bernstein’s book of the same year, The Preconditions of Socialism, posed it from the other side. Between them they set the terms in which the argument has been conducted ever since, and the terms were rhetorically magnificent and analytically crude. Each writer was answering several questions at once, because the situation forced the questions together: a mass party, committed on paper to the revolutionary transformation of society, was in practice devoting its energies to elections, trade union organisation, and piecemeal reform, and somebody had to explain whether the practice betrayed the paper or the paper needed rewriting.

Can socialism be achieved through elections?

Reformists argue accumulated reforms can transform the system, while revolutionaries argue elected governments face structural constraints that prevent changes to ownership. Both accept elections can be won and reforms achieved. The live disagreement is whether an elected government can transform property relations without a rupture.

That short answer contains the whole argument of this article in embryo, and the sections that follow unpack it piece by piece. The unpacking matters because the bundled form of the question does active harm to clear thought about it. When a single slogan covers four questions, a debater can win the argument about one question and claim victory over all of them. The existence of durable welfare states is produced as a refutation of the revolutionary case, although no serious revolutionary denied that welfare states could exist, and the revolutionary question was never whether reforms were possible but whether reforms could sum to a change of system. Conversely, the failure of a particular elected government to carry out its programme is produced as a refutation of the reformist case, although the reformist claim was usually about the cumulative effect of many governments and many decades, not about the omnipotence of any single administration. Each side keeps a stock of victories won against positions the other side does not hold. The four questions grid, presented in full later in this article, is the instrument for ending that exchange of hollow victories, and every section before it supplies the material the grid organises.

There is a further reason the separation matters, and it is historical rather than logical. The participants in the original debate did not disagree about everything, and the record of their agreement is extensive and specific. Luxemburg insisted, in the very pamphlet that made her the emblem of the revolutionary position, that the daily struggle for reforms and democratic rights was indispensable, and that a party which neglected it would forfeit the working class it claimed to lead. Bernstein insisted, in the very book that made him the emblem of revisionism, that he remained a socialist committed to the transformation of society, and that his quarrel was with a theory of how that transformation would arrive, not with the destination. Karl Kautsky, the most authoritative Marxist theorist of the generation, occupied ground between them and spent a decade trying to hold a mass party together on a formula that postponed the choice. The polemic required each of them to describe the others in flattened form, and the flattening has been inherited by nearly every later summary. Recovering the unflattened positions is the first step, and it begins with the world in which the argument arose.

The setting: a mass party with a revolutionary programme and a reforming practice

German social democracy at the end of the nineteenth century was the largest and best organised socialist movement in the world, and its internal tension was visible in its founding documents. The Erfurt Programme of 1891, drafted with the involvement of both Kautsky and Bernstein, combined a theoretical first part, analysing capitalism as a system driving toward its own supersession, with a practical second part listing immediate demands: universal suffrage, labour legislation, free education, and the other furniture of democratic reform. The two parts sat together without any account of how the second would produce the first. In practice the party grew by the second part. Its vote rose election by election. Its trade union allies built funds, offices, and a professional apparatus. Its deputies sat in parliaments. Its members acquired, in short, a great deal to lose, and the question of what all this growth was for became steadily harder to dodge.

The party had also lived through persecution. Bismarck’s Anti-Socialist Laws, in force from 1878 to 1890, had banned its organisations while leaving its candidates able to stand for election, a combination that taught a generation of socialists two lessons at once: that the existing state would suppress them when it could, and that electoral work was the channel through which they survived and grew. Both lessons were carried into the revisionist debate. Bernstein drew on the second. His opponents drew on the first. The material circumstances of the debaters mattered as well. Bernstein spent his exile in England, observing the most gradualist labour movement in Europe, the strength of the Fabian belief in permeation and piecemeal engineering, and a capitalism that appeared to be delivering rising living standards to at least the upper layers of its working class. Luxemburg arrived in Germany from Zurich and Warsaw with the experience of clandestine politics under an autocracy, where no legal road existed at all and reformist patience looked like a fantasy available only to those whose rulers permitted parliaments. Geography did not determine their conclusions, but it stocked their imaginations with different evidence, and the difference shows in everything they wrote.

One more feature of the setting deserves emphasis, because it explains the debate’s peculiar intensity. This was not an argument between a party of reform and a party of revolution. It was an argument inside a single party that contained both, and the stakes were control of that party’s doctrine. Bernstein was not an outsider sniping at Marxism. He had been Engels’s literary executor’s close associate, a co-architect of the orthodoxy he now proposed to revise, and his heresy was dangerous precisely because it came from the centre. The party congresses that debated revisionism, at Stuttgart in 1898, Hanover in 1899, and Dresden in 1903, formally condemned it, and the condemnations changed remarkably little. The practice of the party continued on its reforming course while its programme continued to promise revolution, and that gap between doctrine and practice, rather than any individual’s ambition, is the engine of the whole controversy. Every later version of the argument, in every country, reproduces the same structure: a movement whose daily work is ameliorative and whose stated goal is transformative must eventually decide what connects the work to the goal, and the revisionist debate is the record of the first great attempt to force that decision into the open.

Bernstein’s case at full strength

Bernstein’s revisionism is usually introduced through its most quotable sentence and dismissed before its evidence is examined, which is a disservice to the reader and was a disservice to his original opponents, several of whom understood that the evidence was the dangerous part. The sentence came in a letter and was repeated in his book: the final goal, whatever it may be, is nothing, and the movement is everything. Read in context, the claim was not that socialism should be abandoned for aimless activism. It was that a movement’s daily work of democratisation and reform was not a mere prelude to transformation but was itself the substance of transformation arriving gradually, and that a theory which treated every reform as trivial beside a coming collapse had misunderstood how societies change. The polemical reading, in which Bernstein confessed to having no goal, was available to his critics and they used it for a century. The analytical reading is stronger and deserves the fuller hearing this section gives it.

What did Bernstein actually claim about capitalist development?

Bernstein claimed that capitalism was becoming more stable, more varied, and less crisis prone than the orthodox theory expected, that share ownership was spreading rather than narrowing, that small and middle enterprises were persisting beside large ones, and that these trends made gradual democratic transformation both possible and preferable to waiting for a collapse he thought would not arrive.

The empirical core of the case had several distinct components, and they have fared differently, which is why any honest audit must take them one at a time rather than delivering a single verdict on revisionism as a whole. The first component concerned living standards. The orthodox expectation, drawn from a reading of Marx’s account of accumulation, held that the condition of the working class would deteriorate as capitalism developed, whether absolutely, through falling real wages and deepening misery, or in some formulations relatively, through a shrinking share of a growing product. Bernstein pointed to the evidence available to him: real wages in the advanced economies had risen over the preceding decades, working hours had shortened under legislative and union pressure, and the consumption of the better organised workers visibly included goods that an earlier generation would have classed as luxuries. Whatever the long run held, the short and medium run visible from the 1890s did not look like progressive immiseration, and a party whose propaganda promised deepening misery was describing a world its own voters did not inhabit.

The second component concerned the structure of ownership and enterprise. The orthodox expectation was concentration: competition would eliminate small producers, capital would gather into fewer and larger hands, and society would simplify into two great camps confronting one another across an emptying middle. Bernstein answered with statistics and with distinctions. Joint stock companies, he argued, spread ownership of capital across a widening body of shareholders, so that the number of property owners grew even as individual enterprises grew larger. Small firms persisted in large numbers, sustained by new niches, by local markets, by the service trades that industrial growth itself created. Skilled workers, technicians, managers, and the expanding ranks of clerical and commercial employees formed intermediate strata that showed no sign of dissolving into the proletariat. A society that was becoming more differentiated, not less, could not be expected to polarise on schedule, and a strategy premised on polarisation would wait for a simplification that events were refusing to supply.

The third component concerned stability. Marx’s account of capitalist crisis, in which overproduction periodically convulsed the system, was for Bernstein a theory formed when capitalism was young, national, and unorganised. The capitalism of the 1890s possessed shock absorbers the earlier system lacked. Credit, flexibly extended, allowed firms to bridge difficulties that would once have destroyed them. Cartels and trade associations, whatever their other vices, could regulate output and damp the wildest swings of competition. Improved communications and a genuinely world market spread shocks across a wider surface, so that a local glut no longer had to become a general collapse. Communication, credit, and combination together suggested to Bernstein that general crises of the old type were becoming less likely, and that the catastrophic scenario on which revolutionary expectation rested, the breakdown that would throw state power into the street, was a receding prospect rather than an approaching one.

The fourth component was political, and in some ways the most consequential. Democracy, Bernstein argued, was not a façade. Where the franchise widened and parliaments gained real authority, the working class acquired an instrument through which its numbers could be converted, law by law, into social power. Trade unions were not merely defensive bodies but schools of administration and levers of industrial democracy. Municipal government, which socialists were beginning to win, demonstrated that public ownership of utilities could be built by election and run competently. The state, on this view, was not a monolith permanently fixed in the service of one class. It was a set of institutions whose character changed as the balance of social forces acting on it changed, and sustained democratic pressure could change it progressively. From these premises the strategy followed without strain: pursue reforms energetically, treat each as a permanent acquisition and a platform for the next, extend democracy from the political sphere into the economic, and allow socialism to arrive as the cumulative result of organised pressure rather than as the prize of a single insurrectionary moment. Evolutionary socialism, the name the position acquired, was socialism reimagined as a direction of travel rather than a destination seized at a stroke.

Two features of this case are regularly lost in summary and must be restored for the debate to make sense. The first is that Bernstein presented himself as correcting Marx by Marx’s own method. He charged the orthodoxy with freezing a provisional nineteenth century analysis into dogma, with retaining a Hegelian dialectical scaffolding that forced the evidence into a predetermined pattern of negation and collapse, and with a Blanquist residue, an inheritance from an era of conspiracies, that overrated political seizure and underrated economic development. Revision, in his presentation, was loyalty to the scientific spirit of the doctrine against its letter. The second feature is the ethical supplement. Bernstein, influenced by the neo-Kantian currents then moving through German philosophy, held that socialism needed an ought as well as an is: a moral case for a fairer society that did not depend on predictions about collapse coming true. This mattered strategically as well as philosophically. A movement whose claim rested on justice could argue for each reform on its merits, in public, to voters who shared no theory of history. A movement whose claim rested on prophecy had to keep explaining why the prophecy was late.

The empirical audit: what the following century did to Bernstein’s claims

Because Bernstein staked his case on evidence, evidence is the fair tribunal, and the audit is more interesting than either side’s mythology suggests. His critics in the movement declared him refuted in his lifetime, and later revolutionary tradition treated the refutation as complete. His defenders reply that events vindicated him. Both judgments are too coarse. Taken claim by claim, the record is mixed in instructive ways, and the mixture explains why the debate never closed.

On living standards, Bernstein read his moment correctly and the longer record mostly extended his point for the advanced economies across much of the twentieth century. Real incomes rose across generations, social insurance removed the terror of sickness, unemployment, and old age for hundreds of millions, and the absolute immiseration of the industrial working class in the rich countries did not occur. Qualifications pile up around that sentence and a careful audit states them. The gains were unevenly shared between countries and within them, they depended on political organisation rather than arriving as capitalism’s automatic gift, they coexisted with severe insecurity at the margins of labour markets, and periods of stagnation in median wages in some leading economies late in the century complicated any simple story of continuous ascent. The relative form of the immiseration claim, concerning shares of income and the distance between top and bottom, had a more complicated career, narrowing in the middle decades of the century under the pressure of war, taxation, and union strength, then widening again. Bernstein’s claim as he made it, that misery was not increasing in the manner the propaganda assumed, stands up well. A stronger claim sometimes attributed to him, that the distribution question would solve itself, was not his and would not have survived.

On the middle strata, the audit also favours him in substance while correcting his picture. The independent small producer did decline over the long run in most advanced economies, though more slowly and unevenly than the polarisation thesis expected, and self employment never vanished. What Bernstein grasped, and what proved durable, was that the space between capital and manual labour kept refilling with new occupants: salaried managers, professionals, technicians, administrators, and the vast clerical and service employments of the later economy. The middle did not disappear. It changed its composition, trading the workshop owner for the salaried employee, economic independence for organisational employment, and in doing so it changed its politics as well, becoming a constituency that every electoral strategy had to address and none could take for granted. The simplified two camp society, on which so much nineteenth century strategic arithmetic rested, did not arrive.

On concentration, the audit splits, and the split is the most instructive in the whole examination. Bernstein was right that the number of owners of capital, in the formal sense of shareholders and later of pension beneficiaries and fund holders, multiplied enormously, so that ownership in the legal sense diffused. He was wrong, or at least incomplete, about control. The separation of ownership from control, analysed a generation later by Berle and Means and debated ever since, meant that diffused shareholding could coexist with highly concentrated decision making in the hands of managerial and financial elites. In industry after industry, moreover, the large firm did grow larger, merger waves did consolidate sectors, and in the digital economy of a later age a handful of corporations achieved market weights that would have startled even the orthodox. The honest summary is that capitalism concentrated control while dispersing title, and that each side of the original debate fastened on the half of that sentence that suited it. Bernstein’s statistics answered a claim about the disappearance of small property. They did not answer the claim about the concentration of economic power, and the difference between those claims is one of the permanent lessons of the exchange.

On stability, the audit turns against him, and his opponents were entitled to their satisfaction. The credit system and the cartels did not abolish the general crisis. Within a generation of his book, the world economy suffered the most violent peacetime convulsion in its history, and the crash that began in 1929 and the depression that followed displayed every feature of the old crisis theory at magnified scale, including the inability of combination and credit to contain the collapse and their active role in transmitting it. Later generations added further entries to the register: the stagflation of the 1970s, the financial crises that punctuated the following decades, and the global banking collapse of 2008, in which the credit mechanism Bernstein had cast as a shock absorber functioned instead as a shock amplifier on a planetary scale. Luxemburg’s counter argument, that credit stretches the elastic of the system only to make its eventual recoil sharper, and that cartels export anarchy to the unorganised spaces of the world market rather than abolishing it, reads against this record as the better prophecy. Yet the audit must record the other half as well. The crises came, repeatedly and destructively, and the political sequel Bernstein’s opponents expected from them did not come with them. The great depression produced reforming governments, dictatorships, and eventually war, but in the advanced capitalist countries it did not produce the revolutionary supersession of the system, which emerged from mid century with its property relations not merely intact but reorganised and relegitimated. Bernstein was wrong that the system had learned stability. His opponents were wrong that instability would deliver transformation. The two errors are the twin pillars on which the whole subsequent debate unsteadily stands.

Luxemburg’s reply: the means, the end, and the wage relation

Luxemburg’s answer to Bernstein, published in newspaper articles in 1898 and 1899 and collected as Social Reform or Revolution, is treated elsewhere in this series as a text in its own right, and readers who want its architecture chapter by chapter should go to the reform or revolution readers guide. What belongs here is the argument, because the argument is one pole of the debate this article owns. Luxemburg’s central move was to refuse Bernstein’s separation of the movement from its goal. A movement for reforms, she held, derives its socialist character entirely from the end toward which the reforms are directed. Trade union struggle and parliamentary action are the means by which the working class organises, learns, and accumulates strength. The conquest of political power and the abolition of wage labour are the end that makes the means socialist rather than merely ameliorative. Sever the means from the end, elevate the movement into everything, and what remains is not a maturer socialism but a particularly energetic liberalism, improving capitalism’s conditions while surrendering the claim to replace it.

How did Luxemburg distinguish means from ends in the struggle for reforms?

Luxemburg distinguished reforms as the means, the daily work of organising, legislating, and defending living standards, from revolution as the end, the conquest of political power that alone could abolish the wage relation itself, arguing that reforms pursued without that end lose their socialist character and become mere management of the existing order.

The distinction carried an economic argument inside it, and the economic argument is the part of her reply that still does work. Reforms, however numerous, operate within the wage relation. A law shortening the working day, a union contract raising wages, a factory act limiting abuse: each improves the terms on which labour power is sold, and none touches the fact that labour power is sold, that production is organised by owners of the means of production for profit, and that the worker’s dependence on the wage is reproduced every morning the improvement is enjoyed. Accumulation cannot cross that line by addition. A thousand improvements to the wage relation sum to a better wage relation, not to its abolition, in the same way that a thousand improvements to a tenancy sum to a better tenancy and not to ownership of the house. The quantity of reform and the quality of transformation belong to different registers, and a strategy that confuses them will mistake its own success at the first for progress at the second. Bernstein’s gradualism, on this account, was not a slower road to the same destination. It was a different destination wearing the same name, since a capitalism indefinitely improved is still capitalism by every criterion the tradition itself had supplied.

Luxemburg pressed the economic argument at the specific points where Bernstein had claimed novelty. Credit, in her analysis, was no sedative. It was the mechanism by which capital overleapt its own limits most violently, binding enterprises into chains of mutual dependence along which failure travelled faster, and sharpening the crises it was supposed to soften. Cartels did not plan production for society. They planned restriction for their members, raised prices against consumers and unorganised competitors, and displaced competition onto the world market, where it returned as international antagonism. Democracy she valued as highly as Bernstein did and differently. Political rights were for her the arena in which the working class formed itself into a class capable of ruling, not a machine that automatically converted votes into social transformation. A parliament could register a balance of forces. It could not substitute for the balance, and where the real forces were unequal, paper majorities would discover the limits of paper.

The misreading Luxemburg has suffered ever since should be corrected with emphasis, because it inverts her position. She did not oppose reforms, despise trade unions, or counsel waiting in purity for the great day. She led agitation for democratic rights, defended union organisation against its detractors, and treated the daily struggle as the indispensable school and substance of the movement. Her objection was never to the reform but to the theory that made the reform self sufficient. Indeed her argument contains a warning aimed at her own side as much as at Bernstein’s: a party that treats reforms as mere recruiting devices, valuable only as preludes, will slight the real needs of real people and lose them. The difference between the two positions is therefore not activity versus inactivity, nor gradual work versus dramatic gesture. Both camps worked for reforms with energy. The difference is a claim about what reforms can add up to, and about where the power to change the property structure of society can come from. Keeping that difference in focus, and refusing the caricatures in which reformists want only comfort and revolutionaries want only drama, is the precondition for everything the four questions grid will later formalise.

Her second great contribution to the argument came a few years later, in the pamphlet on the mass strike written after the Russian revolution of 1905, and it answered gradualism from an unexpected direction. The mass strike, as she analysed it, was not a tactic to be switched on by a committee but the characteristic form of a revolutionary period itself, in which economic and political struggles fused, partial strikes for wages became strikes for rights, defensive actions became offensive ones, and the boundary between reformist demand and revolutionary situation dissolved in practice. The significance for the debate is that she refused its premise at a deeper level than the pamphlet of 1899 had done. If movements develop in waves in which modest demands escalate by their own logic, then the choice between reform and revolution is not a menu from which leaders select in advance. It is a rhythm imposed by events, and the task of a party is to be present, organised, and clear about its end when the rhythm changes. Gradualism’s error, on this account, is doubled: it misunderstands not only what reforms sum to but how history moves, supposing a smooth ascent where the record shows long plateaus broken by convulsions in which everything suddenly becomes possible and perishable at once.

Kautsky and the space between: the strategy of attrition

Between Bernstein’s revision and Luxemburg’s reply stood the figure who did most to shape what the movement actually did, and no account of the debate is complete without him. Karl Kautsky was the acknowledged guardian of Marxist orthodoxy, the editor of its leading theoretical journal, and the co-author of the Erfurt settlement that held the party’s two halves together. His response to Bernstein was to defend the doctrine: he wrote the systematic refutation, published in 1899 as Bernstein and the Social Democratic Programme, reasserting the concentration of capital, the inevitability of crisis, and the revolutionary goal. His response to the practical question was more elusive, and the elusiveness was itself a position. The party, he held, was a revolutionary party that could not make a revolution. Revolutions arrive from the conditions of society, not from committee decisions. The task of social democracy was therefore to organise, educate, and accumulate forces until the correlation of strength permitted the conquest of power, a posture that came to be called the strategy of attrition: wear the enemy down, avoid premature confrontation, preserve the organisation intact for a decisive moment whose timing history would announce.

Where does Kautsky sit between revision and rupture?

Kautsky sits between them by defending the revolutionary goal against Bernstein while refusing, against Luxemburg, to treat any present action as its trigger, holding that the party should accumulate organised strength, win democratic ground, and await the conditions in which power could be taken, a posture later named the strategy of attrition.

The attrition formula deserves respect as an attempt to solve a real problem rather than dismissal as mere fence sitting. A mass legal party with millions of voters, vast funds, newspapers, and affiliated unions could not adopt the posture of a conspiracy without destroying itself, and could not adopt Bernstein’s candour without destroying its doctrine and, many feared, its unity. Kautsky’s middle way preserved both the organisation and the goal by assigning the goal to the future and the organisation to the present. Its cost only became visible over time: it trained a generation of functionaries in postponement, made the preservation of the apparatus an end that could override the purposes the apparatus existed to serve, and left the party rhetorically committed to a transformation it had made no plans to enact. When the decisive moment arrived in Germany in 1918 and 1919, in circumstances no strategist had scripted, the party of attrition discovered that it had become a party of government by default and of revolution not at all. The dispute between Kautsky and Luxemburg over the mass strike in 1910, in which she accused him of reducing a living movement to a parade that marches only when ordered, was the moment the middle position’s tensions broke the surface in public. The later fratricidal argument between Kautsky and Lenin, conducted after 1917 over democracy and dictatorship, belongs to the parallel dispute examined in the lenin vs kautsky debate, but its roots lie here, in the unresolved question of what an organised movement does with power when history hands it over unscheduled.

The three positions can now be stated with the precision the rest of this article requires. Bernstein: the goal dissolves into the movement, because capitalist development has made transformation a gradual and cumulative process available to democratic politics. Luxemburg: the goal governs the movement, because reforms cannot sum to transformation and power must be conquered in a rupture when conditions ripen. Kautsky: the goal stands, the movement proceeds, and the connection between them is deferred to a future conjuncture. Nearly every subsequent participant in the argument, however original the vocabulary, can be located by reference to these three coordinates, and the four questions grid explains why: each position is a bundle of answers to the four questions, and the bundles differ chiefly in which answers they combine.

The four questions grid

The separation can now be made explicit. Bundled inside the single slogan are four questions, each with its own subject matter, its own evidence, and its own characteristic confusions. The first asks whether capitalism can stabilise itself: whether the system’s crisis tendencies can be managed, damped, or indefinitely postponed by credit, combination, state intervention, and the other adaptive devices Bernstein catalogued. The second asks whether reforms can accumulate into transformation: whether enough changes within the property order can, at some threshold, amount to a change of the property order. The third asks what the state does to a reforming government: whether winning office delivers command over the apparatus of rule and over the economic power that stands outside the state altogether. The fourth asks whether reforms are worth pursuing regardless of the answers to the other three: whether shorter hours, safer workplaces, insured sickness, and educated children need any further justification than themselves. The four questions grid arranges the debate on those lines, and its final column carries the finding this article exists to establish.

What separates the four questions the debate bundles together?

The four questions are separated by their objects and their evidence: stability is judged by the record of crises, accumulation by whether reforms ever change ownership, office and power by what elected governments actually commanded, and the worth of reforms by their effect on lives, so evidence for one cannot settle another.

Question	Reformist answer	Revolutionary answer	Evidence bearing on it	Do the two sides actually disagree?
Can capitalism stabilise itself?	Largely yes, through credit, combination, and state management of demand and crisis	No, adaptation displaces and sharpens crisis rather than abolishing it	The recurrence of general crises across a century and a half, alongside long stretches of managed growth	Yes, in part: the record supports recurrent crisis and also long stabilisations, so the strong forms of both answers fail
Can reforms accumulate into transformation?	Yes, each reform shifts the balance and prepares the next until the system is changed in substance	No, reforms improve the wage relation without abolishing it, and quantity does not become quality by addition	Welfare states transformed security and services while leaving ownership of productive assets broadly intact	Yes: this is the sharpest surviving disagreement, and the record shows transformation of conditions without transformation of ownership
Does winning office deliver power?	Substantially yes, where democratic institutions are real and a government is determined	No, office brings responsibility for an apparatus and an economy the government does not control	Elected governments repeatedly constrained by capital flight, administrative resistance, military autonomy, and constitutional vetoes	Largely no: serious reformists concede the constraints and dispute only whether persistence and mandate can overcome them
Are reforms worth pursuing regardless?	Yes, improved lives are good in themselves and build the movement	Yes, the daily struggle is indispensable and neglect of it forfeits the class	Both traditions organised for reforms continuously and defended them against reaction	No: the agreement is nearly complete, and the polemic that denies it on either side is caricature

Reading the grid vertically teaches the first lesson: each row is a genuine question with evidence that bears on it specifically, and no row’s evidence can be borrowed to settle another row. Reading it horizontally teaches the second: the two great traditions of the argument are not two monoliths opposed at every point. They are two bundles of answers, agreeing outright on the fourth question, substantially agreed on the third once caricatures are removed, divided on the first in ways the evidence itself divides, and opposed at full strength only on the second. The three-out-of-four convergence is therefore not a rhetorical softening of a hard argument. It is the accurate description of where the argument actually lives once its components are separated. Three of the four questions produce agreement, near agreement, or a shared dependence on mixed evidence. One question, the accumulation question, carries the whole surviving weight of the dispute, and it is there, on whether reforms can sum to a transformation of ownership, that serious argument should concentrate. The sections that follow walk the grid row by row.

Question one: can capitalism stabilise itself?

The stability question is the most purely economic of the four and the one on which the evidence is richest, because the system has run the experiment repeatedly under observation. Bernstein’s stabilisers were credit, cartels, communications, and the world market. The twentieth century added larger instruments to the list: central banking with lender of last resort powers, deposit insurance, countercyclical public spending, welfare transfers that automatically sustain demand in downturns, international monetary institutions, and the whole apparatus of macroeconomic management that governments acquired after the catastrophe of the 1930s. Judged by results, the apparatus works, in the limited sense that depressions of the old depth became rarer in the economies that built it, recoveries were organised rather than awaited, and long periods of stable growth were delivered in the decades after the Second World War. A defender of Bernstein can fairly claim that adaptive capacity turned out to be greater than the collapse theory allowed, and that each crisis produced not the system’s end but a new layer of stabilising machinery built from the wreckage of the last.

The revolutionary answer survives the same record in a different register. The crises were managed; they were not abolished. The register of general convulsions continued to fill: the crash of 1929 standing at its head, the monetary and inflationary disorders of the 1970s, the debt crises that rolled through the developing world in the 1980s, the crashes of 1987, 1997, 2000, and above all 2008, when the financial system of the leading economies required rescue by the very states whose redundancy to the economy had been proclaimed. Each rescue, moreover, was purchased by socialising losses, expanding public debt, or displacing instability onto populations and regions with less protective machinery, so that stability in one place and period repeatedly proved to be instability stored or exported. The stabilisation thesis and the crisis thesis both capture something the record contains, which is that the system is simultaneously more adaptable and more crisis prone than either camp’s strong claim requires. For the purposes of the larger debate, the important point is narrower. Even a complete victory for Bernstein on this row would not settle the argument, because stability is a question about the system’s health, not about its transformability. A capitalism that had achieved perfect stability would still leave open whether it could be voted out of existence, and a capitalism in permanent crisis would still leave open whether crisis delivers its successor or merely its casualties. Participants who treat the crisis record as the whole argument are confusing the first row of the grid with the entire table.

Question two: can reforms accumulate into transformation?

The accumulation question is the heart of the surviving dispute, and it deserves the closest statement each side can give it. The reformist version, in its strongest modern form, does not claim that any single reform transforms the system. It claims that transformation is a matter of thresholds and ratchets. Each reform changes the distribution of resources and rights from which the next campaign begins. Social insurance removes survival from the market’s discretion and thereby changes what workers can refuse. Public education alters the capacities of the next generation. Union rights change the balance of force inside every firm. Codetermination places workers inside the governance of enterprises. Public ownership of utilities and transport removes whole sectors from profit’s command. None of these is socialism. Their sum, the argument runs, shifts the centre of gravity of the economy by degrees, and at no single point need there be a dramatic rupture for the destination to differ fundamentally from the starting point, any more than a coastline needs a single dramatic moment to change its shape. The metaphor of accumulation does quiet work here, borrowing the prestige of Marx’s own account of how capital accumulates, and turning it against his political conclusions.

The revolutionary answer denies the metaphor at its root. Ownership, on this account, is not a quantity that reforms incrementally redistribute but a relation that either governs production or does not. A worker whose sickness is insured, whose children are educated at public expense, whose union is strong and whose pension is secure, still sells labour power to an employer who owns the workplace, directs the labour, and appropriates the surplus. The reforms have transformed the worker’s security, bargaining power, and life chances, all real and all worth defending, and they have left the constitutive relation of the system exactly where it stood. Indeed the revolutionary adds a darker corollary that the reformist must answer: reforms that stabilise the relation may entrench it. A capitalism that insures its workers against catastrophe is harder to indict than a capitalism that starves them, and the welfare state can be read, and has been read by conservative statesmen who built parts of it, as the price of the property order’s survival. Bismarck, who paired social insurance with the Anti-Socialist Laws, understood that reading perfectly.

Between those statements lie the genuinely hard cases, and the hardest is the one that tests whether the accumulation thesis has a limiting instance. In Sweden in the 1970s, the trade union movement adopted the Meidner plan, a proposal under which a portion of company profits would be paid, not in cash, but in shares to collective wage earner funds, so that ownership of Swedish industry would pass gradually, by an automatic annual mechanism, from private shareholders to the employees collectively. Here at last was a reform designed precisely to cross the line the revolutionary answer declares uncrossable by reforms: not another improvement in conditions but a slow transfer of ownership itself, legislated, cumulative, and irreversible by design. The fate of the plan is therefore evidence of the first importance for the grid’s second row. The proposal was diluted under intense employer mobilisation, electoral pressure, and internal hesitations within the labour movement itself, enacted in a drastically weakened form in the 1980s, and later dismantled. Defenders of the accumulation thesis read the episode as a contingent political defeat, the product of a particular campaign and particular compromises, from which the lesson is to build wider alliances and firmer mandates. Defenders of the revolutionary answer read it as the limiting case revealing the rule: a reform that touches ownership itself triggers a mobilisation of the owning class that ordinary reforms never provoke, and the closer accumulation approaches its goal, the more the terrain ceases to be the smooth slope the metaphor requires. The evidence does not decide between those readings by itself, which is exactly why the second row of the grid remains the live centre of the argument. What the Meidner episode does establish, beyond either reading, is that the accumulation question is empirical as well as theoretical, and that its answer will be written by future attempts as much as by past texts.

Question three: does winning office deliver power?

The third question separates two things that electoral language habitually fuses: office and power. Office is occupancy of the formal positions of government, won by votes and recognised by constitutions. Power is the capacity to make one’s decisions effective in the world, and it depends on much that elections do not distribute. A newly elected government inherits a civil service whose habits and loyalties were formed under its predecessors, courts whose composition it did not choose and whose doctrines may veto its programme, central banks and treasuries with institutional commitments to policies it was elected to reverse, armed forces with their own corporate sense of where the nation’s interest lies, and, standing behind all of these, an economy whose daily functioning depends on the confidence and cooperation of the owners of capital, who can express their verdict on a government not by voting but by withholding investment, moving funds abroad, and allowing unemployment to deliver the argument. The systematic treatment of these mechanisms belongs to the marxist theory of the state guide, which examines how the form of the state constrains any government that operates it. Here the point is what the constraint does to the debate, and the answer is that it dissolves most of the apparent disagreement.

What happens when a reforming government meets the state apparatus?

A reforming government meets an apparatus it did not build and does not fully command: officials formed under other policies, courts guarding inherited doctrines, finances dependent on the confidence of owners of capital, and coercive forces with their own loyalties, so that office arrives with responsibility first and power only as far as those constraints permit.

No serious reformist tradition has ever denied the constraints in principle, whatever campaign rhetoric suggests. The literature of democratic socialist government is filled with the memoirs of ministers who discovered the treasury view, the hostile briefing, the investment strike, and the constitutional tripwire, and the theoretical wing of reformism, from the Webbs through Crosland to the Eurocommunist strategists, built its plans around the expectation of resistance and the need for a broad social bloc, administrative preparation, and international room for manoeuvre. No serious revolutionary tradition, conversely, has denied that office can be won and used for real gains, since its own history includes municipal governments, parliamentary fractions, and wartime administrations it regards with pride. What the two sides dispute on this row is a matter of degree and conclusion: whether the constraints are obstacles that determination, preparation, and popular mobilisation can progressively overcome, or whether they are constitutive, such that a government which accepts the apparatus accepts its limits and a government which defies them must expect to be broken or to break them, with all that breaking entails. That is a real disagreement, but it is narrower than the slogan of the parliamentary road versus the revolutionary road suggests, and both sides argue it in the same vocabulary of structural constraint that the grid’s third row imposes.

One case towers over this row of the grid, and discipline requires that it be linked rather than argued here, because it has an article of its own and because its evidentiary weight is so great that a paragraph would trivialise it. The Chilean experience of 1970 to 1973, in which a socialist government elected on a programme of democratic transformation attempted to carry that programme through the inherited institutions and was destroyed by a military coup, is examined in full at allende chile democratic road. Every participant in the debate must pass through that case. Reformists must explain what was contingent in it, the narrowness of the electoral mandate, the specific international hostility, the errors of sequencing and alliance, and why a different government in different conditions could expect a different end. Revolutionaries must explain what general lesson a single case, however terrible, can bear, and why the same apparatus that crushed an elected government would permit the revolutionary alternative they propose to be built beside it. The case disciplines both sides, which is the most that any single historical episode can do, and it leaves the third row where the grid places it: a question on which the facts are largely agreed and the conclusions from the facts are not.

A distinction from the strategic literature sharpens the row further. Winning an election, holding office, and exercising power form a sequence in which each stage can fail independently. Parties have won elections and been denied office by constitutional manoeuvre. Parties have held office for years while exercising power only within boundaries drawn elsewhere, administering budgets whose parameters were set by markets and treaties they had campaigned to escape. And parties have occasionally exercised more power than their office seemed to promise, when mass mobilisation outside parliament shifted the limits of the possible faster than institutions could adjust. The third row’s real question is therefore not whether elections matter, since both sides grant that they do, but which stage of the sequence is the bottleneck for transformation, and the honest answer the record suggests is that the bottleneck sits between office and power, in the terrain of apparatus, economy, and organised force that the state theory tradition maps.

Question four: are reforms worth pursuing regardless?

The fourth question is the shortest to state and the one on which the grid records its most complete agreement, and the agreement matters because the polemical tradition of both camps denies it exists. The revolutionary of caricature wants only the final confrontation and regards every amelioration as a narcotic delaying it. The reformist of caricature has traded the goal for the amelioration and regards every reminder of transformation as an embarrassment. Neither figure survives contact with the record of what the traditions actually did. The revolutionary parties and movements of the Marxist tradition built their strength through relentless organisation around immediate demands: the eight hour day, the franchise, insurance against sickness and unemployment, the right to combine, the defence of living standards against every attempt to lower them. Luxemburg’s own argument, examined above, made that daily work the very medium in which a revolutionary class forms. The reformist parties, for their part, officially retained transformative goals for generations after Bernstein, and even those that formally renounced them continued to justify reforms as steps in a continuing democratisation of economic life rather than as mere housekeeping. On the question whether reforms are worth winning, the two traditions speak with one voice, and the voice says yes, for two reasons that are also shared: because the improvements are goods in themselves to the people who live them, and because the struggle for them is the school in which collective capacity is built.

Why do reforms both improve lives and stabilise the order they amend?

Reforms improve lives by removing specific insecurities and indignities from the people who receive them, and they stabilise the order they amend by relieving the pressures that might otherwise gather against it, and both effects are real at once, which is why the same reform can be honestly celebrated and honestly suspected by different judges.

The duality stated in that answer is not a paradox to be resolved but a fact to be kept in view, because each side of the debate habitually seizes one effect and denies the other. The reform that relieves pressure does tend to legitimate the arrangement within which the pressure arose; the history of social insurance begins with a chancellor who intended exactly that. The reform that improves lives does also tend to organise and embolden its beneficiaries; the history of the franchise, of union rights, and of public education is filled with concessions that created the constituencies which demanded the next concession. Which tendency dominates in a given case depends on the reform’s design, on who controls its administration, on whether it divides its beneficiaries or unites them, on whether it creates rights that can be built upon or charities that can be withdrawn. That dependence is the doorway through which the most interesting attempt to escape the whole dichotomy enters, André Gorz’s distinction between reforms that stabilise and reforms that shift power, which receives its own section below. For the grid’s purposes, the fourth row closes with a verdict both camps can sign: reforms are worth pursuing regardless of the answers to the other three questions, and the argument about them is never whether to pursue them but what they can be made to add up to and who they empower while they accumulate. With that agreement registered, the debate’s centre of gravity moves decisively to the two places the grid has marked as live: the accumulation question of the second row, and the structural analysis that explains why parties keep failing to answer it in practice, which is the subject of the next section but one.

The structural dilemma: Przeworski and the arithmetic of the electoral road

The most consequential contribution to the debate since its founders came not from a pamphleteer but from a political scientist writing with the tools of formal analysis, and its power lies in converting a century of accusations about betrayal into a problem of arithmetic. Adam Przeworski, in Capitalism and Social Democracy, published in 1985, and in the companion study Paper Stones, written with John Sprague and published in 1986, asked what follows from the brute demographic fact that manual workers, the class to which socialist parties historically appealed, have never constituted a majority of the electorate in any advanced capitalist democracy. The electoral arithmetic underneath that fact, and the question of who the potential majority consists of, connects this debate to the wider inquiry pursued in does the working class still exist, since a party’s strategy depends on where it believes its voters are to be found. Przeworski’s dilemma follows from the arithmetic with uncomfortable directness. A socialist party that appeals to workers alone, on a programme of undiluted transformation, assembles a loyal minority and loses. A socialist party that broadens its appeal to the salaried employees, professionals, small proprietors, and other strata whose support a majority requires must moderate its programme to what those allies will accept, since allies are not recruited to projects that threaten their own positions. Electoral success and programmatic purity therefore stand in structural tension, not because leaders are cowards or traitors, but because the composition of the electorate makes the combination of a majority and an uncompromised transformative mandate arithmetically unavailable in the normal case.

How does Przeworski model the electoral dilemma of socialist parties?

Przeworski models the dilemma as a trade off built into the electorate’s composition: workers alone are never a majority, so a socialist party must either keep its transformative programme and lose, or broaden its appeal to other strata and dilute the programme, making electoral success and undiluted transformation structurally difficult to combine.

The force of the model is that it explains the pattern of social democratic moderation without recourse to anyone’s character. Generation after generation of revolutionary critics surveyed the record of socialist parties in office and found the same outcome: programmes trimmed, transformations postponed, capitalism administered with a human face and then, in time, with an increasingly familiar one. The critics explained the pattern morally, as betrayal, or organisationally, as the capture of parties by a bureaucracy with interests of its own. Przeworski’s model explains it positionally. Any party that wishes to govern through elections must solve the majority problem, and the solution exacts the moderation as its price, whoever leads the party and however sincere its founding commitments. Leaders who refuse the price are replaced by the electorate’s arithmetic with leaders who will pay it, or their parties are replaced by parties that already have. The dilemma also runs in the other direction, and Przeworski was careful to state the second horn with equal force. A socialist government that wins office on a broad and moderate mandate and then attempts transformation beyond that mandate confronts the structural power of capital in its purest form: investment is withheld, currency flees, and the government faces a choice between retreat and an economic disruption for which it will be blamed by the very voters whose support it needs. The dilemma is therefore double. Dilution is the price of election, and economic retaliation is the price of exceeding the diluted mandate. Between the two prices, the space for an electoral road to a transformed property order becomes very narrow, and the model predicts that it will be found, if anywhere, only under unusual conditions of crisis, unity, and international shelter.

The model’s implications cut against both camps in ways each finds uncomfortable, which is the mark of its neutrality and the source of its value to this article. Against the naive form of reformism, it denies that the electoral road is a smooth ascent obstructed only by failures of will. The obstruction is structural, and will is precisely what the structure taxes. Against the moralised form of the revolutionary critique, it denies that the record of moderation proves corruption. The same record would be produced by honest leaders facing the same arithmetic, and the accusation of betrayal, applied universally, explains nothing because it is compatible with everything. What the model leaves to the revolutionary is the question his tradition has always pressed: if the electoral road narrows as Przeworski describes, what road remains, and what is its record when subjected to the same unsparing analysis? What it leaves to the reformist is the burden of identifying the exceptional conditions under which the dilemma relaxes, and of showing that a strategy can be built to wait for, recognise, and use them. Neither burden is discharged by quoting the model. Both are clarified by it, because the model tells each side exactly which claim it must now defend.

Qualifications belong in the record. The dilemma is a model, and models simplify. Parties are not the only vehicles of transformation, and the electorate is not a fixed census but a shifting field that organisation, events, and ideas can reshape; the boundary of the working class itself is one of the contested variables, as the linked inquiry into its existence shows. Historical episodes exist in which socialist parties won large mandates on comparatively radical programmes, most instructively where war or collapse had discredited the old order and temporarily rearranged what majorities would accept, and episodes exist in which broad coalitions sustained reforms deeper than the model would predict. Przeworski himself drew a conclusion more austere than many of his readers: that the historical experience of social democracy showed parties choosing, at each fork, the option that preserved their electoral viability, and that the cumulative result was a capitalism humanised in many respects and transcended in none. Whether that verdict is a refutation of the reformist road or a description of its actual historical function, valuable on its own terms, is precisely the judgment the two camps will continue to dispute. The model’s achievement is to have made the dispute exact.

The dilemma’s reach extends beyond the European cases from which it was built, and testing it against other terrains is a useful exercise in what kind of claim it is. In settler democracies with strong labour parties, in the Latin American republics where urban workers formed larger electoral blocs, and in the postcolonial states where socialist governments were elected amid the prestige of liberation movements, the same sequence recurs with local variations: a mandate assembled from a coalition wider than the transformative core, a programme trimmed at the coalition’s edges, an administration judged by an economy its enemies substantially owned, and an eventual reckoning in which the movement’s own historians divide between the vocabularies of betrayal and of structure. The variations matter as much as the recurrence. Where the state itself was newly built by the movement’s allies, where land reform could be carried against a defeated oligarchy, or where a commodity boom filled the treasury, governments travelled further along their programmes before the constraints closed, which is what a structural model predicts: structures set the gradient, and circumstances decide how far along it a given attempt slides before it stops. A model that explains both the recurrence and the variations, without needing a villain in either, has earned its place at the centre of the modern debate, displacing the founders’ polemics not by refuting them but by stating the problem they were both, from opposite ends, describing.

The strongest objection, and the strongest reply

Every debate has a point at which the hardest argument must be faced without flinching, and for the reformist road that point is the objection from the universal pattern. Stated at full strength, the objection runs as follows. The parliamentary road to socialism has been attempted, in one form or another, by socialist and labour parties across a century, in countries rich and poor, in times of boom and crisis, under leaders of every temperament from the cautious to the fiery. The result is invariant. No party that took office through elections has used that office to transform the ownership of the productive economy in the direction of socialism and completed the transformation. Every one of them, without exception, ended by administering the capitalist order: some humanely and reformingly, some timidly, some in the end as the executors of policies their founders would have denounced, but all within the property relations they were founded to abolish. A pattern that universal, across that many independent trials, cannot be explained by the failings of individuals, which vary, or by accidents of circumstance, which cancel out. It requires a structural explanation, and the structure is not hard to find: the state a government inherits, the economy it must keep running, the electorate it must keep persuading, and the party apparatus it must keep fed all select for administration and against transformation, so reliably that the outcome might as well be a law. On this objection, the reformist road is not a road that has been travelled badly. It is a road that does not reach its destination, and a century of arrivals at the wrong place is the proof.

The reply must be given at equal strength, and it has three parts. The first part concedes the pattern and disputes its interpretation. The universal failure to transform ownership is real, but the reformist asks what the counterfactual is. The revolutionary road has also been attempted across the same century, in its own set of countries and conditions, and its record, measured by the values both traditions profess, is not better. Where revolutionary governments took and held power, they transformed ownership indeed, and the transformation was accompanied, in case after case, by the suppression of the democratic rights and independent working class organisations in whose name the transformation was made, and in the largest cases by economic results that eventually collapsed back toward the market. A choice between roads is a choice between records, and the reformist record of transformation forgone must be weighed against the revolutionary record of transformation achieved at prices its own tradition’s founders would have recognised as defeat. The second part of the reply redistributes the pattern’s meaning. What the electoral road reliably produced was not nothing. It produced the welfare states, the labour rights, the public health and education systems, and the taming of the market’s cruellest exposures that constitute, for hundreds of millions of people, the difference between security and destitution. A tradition that counts those as failures because ownership persisted has defined success so that only its own road can achieve it, which is a definition, not a finding. The third part accepts Przeworski’s structural explanation and draws the opposite moral from it. The dilemma is genuine, and precisely because it is genuine it acquits the generations who faced it of the charge of betrayal, and reframes the strategic question honestly: not why did they sell out, but what conditions, coalitions, and sequences might relax a dilemma that is structural rather than moral, and whether preparing for those conditions is not exactly the work the daily struggle performs. The objection demanded a structural explanation. The dilemma supplies one. The objection’s further claim, that the structure settles the matter against reform forever, is the part the reply refuses, on the ground that structures constrain and select but have never yet relieved human beings of the need to test their limits in practice.

Gorz and the attempt to dissolve the dichotomy

If the four questions grid is the analytical instrument for separating the debate, André Gorz supplied the most influential attempt to escape the debate’s practical dilemma from within. In Strategy for Labor, published in French in 1964 and in English in 1967, the Austrian born French theorist proposed a distinction between two kinds of reform. A reformist reform, in his vocabulary, is one whose criteria are set by the system’s own requirements: it improves conditions in ways that leave the distribution of power untouched or strengthened on the side of capital, it can be granted from above without danger to the order, and its beneficiaries receive it as clients. A non-reformist reform, by contrast, is judged by criteria independent of the system’s logic: it is won rather than granted, it creates new centres of power and new rights that their holders can defend and extend, it raises demands the system cannot absorb without changing, and it is designed as one step in a sequence whose direction its participants understand. The distinction relocates the entire argument. The question is no longer the false one of reform versus revolution as rival timetables. It is a question about the design and direction of each particular reform, about whether a given measure demobilises its beneficiaries into gratitude or organises them into a force capable of demanding the next measure.

The proposal’s attractions are obvious, and so are the difficulties, which a fair account must state in the order Gorz’s critics pressed them. The first difficulty is classification. Whether a reform shifts power or stabilises the system is rarely legible in advance, and the same measure can do both at different horizons: social insurance stabilised Bismarck’s Reich and also created the material security from which stronger unions grew; public education serves the economy’s need for skilled labour and also produces citizens harder to deceive. Gorz’s criterion may therefore be less a test that sorts reforms beforehand than a discipline of intention, valuable for forcing movements to ask what a demand does to the balance of forces, without promising a reliable answer. The second difficulty is absorption. A system as adaptive as the stability debate showed capitalism to be can absorb many power shifting reforms by conceding the form and recapturing the substance, funding the new rights while reorganising around them, so that yesterday’s non-reformist reform becomes the platform of a new stability. The third difficulty is the agent. A sequence of power shifting reforms requires a movement coherent enough to hold the sequence in mind across defeats and decades, which is precisely the organisational achievement the dilemma literature casts in doubt. None of this refutes Gorz. His distinction survives as the most useful question a strategist can ask of any particular proposal, and its survival in that modest form illustrates the general pattern this article keeps finding: the debate’s durable products are instruments for thinking, sharper questions and better distinctions, rather than verdicts that close it.

Eurocommunism: the democratic road restated with the state in view

The debate’s most systematic later restatement came in the 1970s, when the communist parties of Italy, Spain, and France, the mass parties of the Western European left that stood outside the social democratic tradition, converged on a strategy that accepted the electoral road while refusing Bernstein’s conclusions. Eurocommunism, as the tendency was labelled by journalists and half accepted by its protagonists, held that in societies with dense civil societies, entrenched democratic institutions, and working classes integrated into national life, socialism could only be built by a long democratic march through the institutions, carried by a broad bloc of social forces extending well beyond the industrial proletariat, and legitimated at every stage by majority consent. Santiago Carrillo’s Eurocommunism and the State, published in 1977, gave the tendency its theoretical statement on the decisive point: the state apparatus inherited from capitalism could not simply be wielded for new purposes, as a naive instrumentalism supposed, nor did it have to be smashed in a single act, as the classical revolutionary formula required. It had to be transformed from within and without at once, democratised piece by piece under the pressure of a society whose organisations were themselves being democratised, in a process for which the Gramscian vocabulary of hegemony, absorbed through the Italian party’s culture, supplied the grammar.

Eurocommunism matters to the four questions grid because it was, in effect, an attempt to answer the grid’s third row as a practical programme. Its authors had studied the Chilean catastrophe at close quarters, and Enrico Berlinguer’s reflection on it, the strategy of the historic compromise, was drawn directly from the lesson that a government of the left with a bare majority, facing a hostile apparatus and an alarmed middle class, would be destroyed; therefore the bloc had to be wider, the consent deeper, and the transformation slower than impatience wished. Critics from the revolutionary left answered that the strategy conceded the dilemma’s second horn in advance, accepting constraints so completely that transformation dissolved into administration by another name, and the subsequent history gave those critics material: the French party’s decline after its alliance and rupture with the socialists, the Spanish party’s marginalisation after the transition it had helped to negotiate, the Italian party’s long evolution ending, after 1989, in its own dissolution into a post communist formation. Defenders reply that the counterfactual is again unkind, that the revolutionary alternatives in those same societies produced nothing but sects, and that the Eurocommunist parties at least preserved democratic commitments their tradition had historically lacked, a gain whose value the century’s record makes it impossible to dismiss. The episode closes, for this article’s purposes, with the observation the grid predicts: Eurocommunism agreed with reformism on the fourth question, with the revolutionary critique on the third question’s diagnosis, and staked everything on the second, the possibility that democratising reforms could accumulate across the state and civil society into a change of system. It is the accumulation question that judged it, as it judges every strategy in this debate.

The high water mark: a reforming government in Britain after the war

Abstract argument about what elected governments can do benefits from one fully described instance, and the British Labour government of 1945 to 1951 is the instance the debate returns to because it is the strongest case the reformist record contains. Elected in the landslide that followed the Second World War, with a clear majority and a programme its leaders had prepared in detail, the Attlee government proceeded to build more of a new order in six years than most reforming governments attempt in a generation. It created the National Health Service, removing medical care from the market and from charitable discretion alike and establishing treatment free at the point of use as a universal right of citizenship. It nationalised the Bank of England, the coal industry, the railways and much of inland transport, electricity and gas supply, and iron and steel, bringing roughly a fifth of the economy into public ownership. It completed a comprehensive system of social insurance, built houses on a mass scale, extended secondary education, and maintained full employment as the standing objective of economic policy. The transformation of British life was profound and, in the case of the health service, proved among the most durable institutional creations of the century in any country.

The same instance, examined from the other side of the grid, marks the reforming road’s limits with equal clarity. The nationalisations were conducted on the model of the public corporation, compensating owners generously and reproducing inside the public sector the managerial hierarchies of the private one, so that the miner’s relation to the coal board differed less from his relation to the coal company than the movement’s rhetoric had promised. The commanding heights beyond the nationalised fifth, finance above all in its private reaches, manufacturing, land, food distribution, remained in private hands, and the government, dependent on American credit and committed to restoring exports, governed the mixed economy in close negotiation with the interests it had been elected to supersede. Private wealth, private schools, and the social order they sustained emerged from the six years substantially intact. When the long postwar boom later faltered and a hostile government after 1979 set about reversing the settlement, the reversal found much to work with: the privatisation programme of the 1980s and 1990s returned most of the nationalised industries to private ownership, labour law was rewritten against the unions, and the parts of the 1945 creation that survived intact were those, preeminently the health service, that had built a universal constituency too broad to attack frontally. Reformists draw from the episode the lesson that deep change by election is possible and that its durability is a function of how completely a reform embeds itself in the expectations of the whole population. Revolutionaries draw the lesson that even the strongest reforming government, with war’s prestige, a landslide, and a disciplined party, transformed the conditions of the working class and the services of the state while leaving the structure of ownership and class power available for restoration the moment political weather changed. Both lessons are drawn from the same six years, and the grid exists precisely to show that both can be true, because they answer different rows: the fourth row records an unambiguous triumph, the second row an unambiguous shortfall, and any verdict that collapses those rows into one word, success or failure, has stopped analysing.

The historical record of both roads

A full comparative history of the two roads would fill a library, but the contours relevant to the grid can be drawn honestly in a few strokes, provided the comparison is run in both directions and neither tradition is judged by its aspirations while the other is judged by its outcomes. The comparison between Marxism as a revolutionary tradition and social democracy as a governing one is pursued as a study of two traditions in marxism vs social democracy; what follows is the narrower evidentiary summary the debate requires.

The reformist road’s positive record consists of the societies of the postwar West at their best: the Scandinavian countries above all, where social democratic governments, resting on powerful union movements, built universal welfare states, compressed income differences to the narrowest ranges recorded in capitalist societies, maintained high employment, and sustained levels of social security, public health, and education that stood for decades as the practical benchmark of what democratic politics could extract from a market economy. Around that core stands the wider record of Western Europe and its offshoots: labour law, public housing, mass secondary and then higher education, the civilising of industrial relations, and the conversion of the working class from a propertyless caste into a citizenry with stakes, savings, and expectations. Against that record stand the entries already examined: ownership untransformed, settlements reversed when the political weather turned, colonial and postcolonial violence in which social democratic governments were sometimes complicit, and the long late century drift in which party after party of the tradition administered, and in some cases actively conducted, the market liberalising programmes of the era, hollowing the tradition’s distinctiveness until its own voters struggled to state it. The tradition’s defenders and its critics are both describing the same century, selecting different ledgers.

The revolutionary road’s record must be drawn with the same double entry discipline. Its positive entries are real and are sometimes written out of summaries compiled by its opponents. Revolutionary governments broke the power of landed oligarchies that no electoral process had been able to touch, carried through land reforms, literacy campaigns, and public health drives that transformed the life chances of millions, industrialised backward economies at speeds the market had not achieved there, and in the Soviet case supplied the manpower and sacrifice that broke the armed force of Nazism, at a cost its peoples paid in full. Its negative entries are graver than any in the reformist ledger, and honesty forbids euphemism here. The one party states that revolutionary governments built suppressed independent unions, free expression, and political pluralism as a matter of system rather than emergency; the largest of them conducted terror against their own populations on a scale that stands among the century’s defining crimes; their economies, after initial feats of mobilisation, stagnated into shortages and privilege, and between 1989 and 1991 the European members of the system collapsed, mostly without defenders, and restored private property with a speed that suggested how little of the transformation had rooted itself in the allegiance of the people who lived it. China and Vietnam, retaining the party state, restored the market economy under its direction. No revolutionary government created the combination of social ownership with democratic liberty that the tradition’s founding texts had projected as socialism’s meaning, and most ceased to claim the combination was near.

The comparative conclusion the evidence supports is uncomfortable for every participant, which is usually the sign of an honest one. Measured by transformation of ownership, the revolutionary road did what the reformist road never did, and the price recorded against the transformation includes the liberties in whose name it was undertaken. Measured by the security, health, education, and political freedom of ordinary people sustained over generations, the reformist road’s best societies outperform every revolutionary state, and the transformation of ownership in those societies is precisely nil. The grid’s second row, the accumulation question, is where the reformist road fails its own founding test. The revolutionary road fails a test its founders set with equal explicitness, the test of emancipation, since a transformation of property accomplished by extinguishing the self organisation of the class it claimed to liberate reproduces, in the state, the separation between rulers and ruled that the doctrine existed to abolish. Any participant who quotes only one side of that double failure is campaigning, not analysing, and the reader now possesses the instrument for detecting the omission: ask which row of the grid the quoted record answers, and which rows have been left in darkness.

The question of time: horizons hidden inside the argument

Beneath the four questions runs a fifth disagreement that the grid does not tabulate because it concerns not evidence but time, and naming it explains much of the debate’s emotional temperature. The reformist road is a claim about long durations. Its unit of account is the generation: rights accumulate across decades, expectations shift across lifetimes, and the transformation it promises is the work of a century conducted by people who will not live to see its completion. The revolutionary road is a claim about concentrated moments. Its unit of account is the conjuncture: the crisis, the collapse of legitimacy, the weeks in which the possible expands violently and then contracts, possibly for another lifetime. Each temporality breeds its characteristic virtue and its characteristic corruption. Patience sustains institutions and delivers compound gains, and it corrupts into the postponement that made attrition a habit and the goal a decoration, since a destination always one generation away exerts no discipline on the present whatsoever. Urgency seizes the moment that patience would miss and honours the truth that some doors open once, and it corrupts into the permanent emergency in which any liberty may be suspended and any crime excused by the decisiveness of the hour, since a movement that lives in the exception soon builds its home there.

The two temporalities also distribute risk differently across the people in whose name both speak, and a complete analysis registers the distribution. The costs of the reformist road are paid in the currency of lives lived under conditions that reform reaches slowly or never: the generation that grows up before the housing is built, the workers whose industry is gone before the transition arrives, the colonised and the bombed for whom the metropolis’s gradual enlightenment came too late to matter. The costs of the revolutionary road are paid in the currency of the moment itself and its aftermath: the violence of the rupture, the civil wars that so often attend it, and the long discipline of the states built to defend it, whose bills are presented to populations that did not sign them. Neither currency is imaginary, and the choice between the roads, where a real choice exists, is partly a choice about which costs a movement is willing to impose and on whom, which is a moral and political judgment that no grid can settle. What analysis can do is prevent the costs from being hidden. A reformism that speaks only of its gains has concealed its price in other people’s lifetimes. A revolutionism that speaks only of the old order’s crimes has concealed its price in the new order’s prisons. The debate conducted honestly is, among other things, an exchange of invoices, and each side’s willingness to present its own is the truest available measure of its seriousness.

Misreadings that keep the argument crude

Certain errors recur so predictably in summaries of the debate that correcting them is a service in itself, and each error corresponds to a row of the grid collapsed or a position flattened. The first error holds that revolutionaries oppose reforms. Luxemburg’s texts refute it directly, her career refutes it in practice, and the organisational history of every revolutionary movement refutes it structurally, since such movements grew by fighting for immediate demands and withered when they disdained them. The error survives because it is useful: it lets the reformist present every revolutionary as a prophet of misery who needs conditions to worsen, a figure who exists in rhetoric more often than in the record. The second error, its mirror, holds that reformists deny class conflict and imagine society as a harmony awaiting administration. Bernstein’s pages acknowledge antagonism throughout; his claim was that antagonism could be processed by institutions rather than that it had evaporated, and the union based reformism of the Scandinavian type was built by people who expected conflict permanently and organised for it. The error survives because it lets the revolutionary present every reformist as a dupe of the system’s self image.

The third error holds that the existence of the welfare state settles the argument, in one direction or the other. For the triumphant reformist, the welfare state is produced as proof that the system was transformable all along and the revolutionary warning empty. For the revolutionary, the same welfare state is produced as proof of the system’s cunning, a bribe that purchased the property order’s survival. The grid shows why both productions fail: the welfare state answers the fourth question triumphantly and the first question partially, and it is nearly silent on the second, which asks about ownership, and on the third, which asks about the power that permitted it and later trimmed it. An institution can be a genuine conquest and an insufficient one in the same breath, and the refusal to hold both judgments at once is the refusal of analysis itself. The fourth error is lexical. Revisionism in the polemical vocabulary of the movement means betrayal, a sellout dressed as theory, and the word does the work of refutation without performing it. In analytical use, revision names what Bernstein actually did and what every living doctrine periodically requires: the comparison of inherited predictions with arrived evidence and the amendment of the doctrine where the two diverge. Whether his particular revisions were right is the empirical question this article has audited claim by claim. That revision as such is treason is a proposition no science and no serious politics can accept, and the parties that enforced it most rigorously were obliged, in the end, to revise in secret what they forbade in public, at the cost of the candour that makes revision corrigible.

The fifth error holds that Bernstein was refuted by events, full stop, a judgment repeated so often that it has acquired the texture of fact. The audit conducted earlier in this article shows the actual distribution: his claims about living standards and the middle strata were substantially borne out for the cases he examined, his claim about stability was destroyed by the century’s crises, his claim about ownership confused the diffusion of title with the dispersal of control, and his political wager, that democracy would progressively socialise the state, received the mixed verdict that the state question always delivers. A thinker with that score sheet was neither refuted nor vindicated, and the insistence on a single word verdict is itself one of the habits the four questions exist to break. The sixth and final error is the framing error with which this article began, the treatment of the whole matter as a single question with a single answer, and it is the only error whose correction the article claims as its own contribution. Everything else, the evidence, the texts, the records, was already available. What was missing was the separation.

What would settle the argument, and why it stays open

A debate that has run for well over a century invites the suspicion that it is not really about evidence at all, and the suspicion deserves a direct answer. Parts of the argument are evidence responsive and have visibly moved under evidence’s pressure. Bernstein’s empirical challenge forced every subsequent Marxist to argue about concentration, living standards, and the middle strata with data rather than incantation, and the strong collapse theory he attacked is defended in its original form by almost no one. The Chilean catastrophe permanently changed what democratic road strategists had to answer, and Eurocommunism’s shape is unintelligible without it. The collapse of the Soviet system between 1989 and 1991 settled, for all but the most committed, the specific claim that the one party planned economy was a viable and attractive form of socialism, and in settling it redrew the entire debate’s map, since the revolutionary pole lost its standing exhibit. Arguments that move under evidence are not theology, however slowly they move.

The accumulation question, the grid’s second row, stays open for a specifiable reason: the decisive experiment has never been run in its clean form. The clean form would require a government elected on an explicit programme of transferring ownership, holding a durable majority and a united movement, enjoying international conditions that neither strangle it nor rescue it, and persisting across enough electoral cycles for the transfer to consolidate, and then the record of what happened to the programme, the economy, and the democracy would answer the question for that class of cases. History has supplied approximations, each ruined for experimental purposes by a confounding variable a partisan can cite: mandates too narrow, movements divided, superpowers intervening, wars interrupting, leaders hesitating. Sweden’s funds were diluted before they bit. Chile was destroyed from outside and within. Britain in 1945 never aimed at general ownership transfer in the first place. The absence of a clean run is itself a datum the two camps read differently, the reformist as contingency, the revolutionary as the structure revealing itself precisely in the universality of the confounders, since a world in which the experiment can never be run clean may be a world in which the obstacles are not accidents but features. That meta disagreement, about what the absence of evidence is evidence of, is the deepest layer of the dispute and the least likely to be settled by any single future event.

What can be said at this point in the analysis is where a reader should now direct attention when the argument next presents itself, as it will in every generation’s vocabulary and under every era’s party names. Refuse the bundle. Ask which of the four questions any speaker is answering, and decline to let an answer to one stand in for answers to the rest. Expect agreement, once the questions are separated, on the worth of reforms and substantial agreement on the constraints that office imposes. Concentrate scrutiny on the accumulation claim, in whichever direction it is made, and ask for its mechanism: by what steps, through which institutions, against which predictable resistance, does this reform sum with others into a change of who owns the productive wealth of society, or by what necessity does every such sum fall short? And hold both traditions to their full records, the security built and the ownership untouched on one side, the ownership transformed and the emancipation forgone on the other. The debate survives because the question at its centre is real and the evidence is honestly mixed. It deserves better than its slogans, and its slogans, separated, turn out to contain a great deal of agreement and one precise, difficult, still open question.

That precision is itself a historical achievement worth recording. The founders argued with prophecies, and their successors argued with memories, and the argument improved each time it was forced onto narrower ground: from the fate of a system to the fate of its predictions, from the character of leaders to the arithmetic of coalitions, from the justice of a cause to the design of the institutions that would have to carry it. A reader who follows the narrowing has watched an argument grow up, and the growing is not finished, because each generation’s circumstances pose the four questions in materials the founders never handled, digital platforms and financialised housing, climate transition and care economies, and the answers earned on the old materials transfer only as method, never as verdicts.

The debate becomes a split: war, collapse, and two revolutions

For fifteen years after Dresden the argument remained, formally, an argument within one movement, conducted in journals and congresses by people who expected to die in the same party. The First World War destroyed that container. In August 1914 the parliamentary fraction of German social democracy voted the credits that funded the Kaiser’s war, and the Second International, whose congresses had pledged the workers of Europe to resist exactly that act, collapsed in a weekend. The relevance to the debate is not merely chronological. The war vote was the reform or revolution question posed in its purest practical form: a party built by decades of legal, electoral, ameliorative work discovered what that work had made of it when the state demanded its loyalty against its doctrine, and the apparatus chose the state. Bernstein, to his credit and his critics’ confusion, opposed the war and joined the anti war minority, a reminder that the camps of 1899 did not map simply onto the camps of 1914, and that gradualism and nationalism, however often they travelled together, were distinct commitments. Kautsky manoeuvred for the centre and lost both wings. Luxemburg, imprisoned for much of the war, wrote from her cell the pamphlet that indicted the party she had served for the catastrophe of its surrender.

What followed the war posed the question as action rather than text. The Russian revolutions of 1917 demonstrated that a revolutionary road existed in fact and not only in pamphlets, and simultaneously transformed its meaning, since the party that took power in October proceeded to build a model whose relationship to the older Marxist programme became the century’s most fought over question, treated in this series where it belongs. The German collapse of November 1918 handed the question to the debate’s surviving founders in person. The party of attrition found itself, by the abdication of the Kaiser and the disintegration of the old order, in possession of a government it had not planned to take, facing a choice between convening a national assembly on the parliamentary road and building a republic of councils on the revolutionary one, with armed force, economic breakdown, and a mutinous soldiery deciding the tempo. It chose the assembly, allied with the old officer corps against the revolutionary left, and in January 1919 Luxemburg was murdered by soldiers operating under that alliance’s protection, weeks after founding a communist party whose first programme she had written with warnings against premature insurrection that her killers did not stay to read. The debate’s two great antagonists thus ended the story asymmetrically: Bernstein lived to serve as an elder of the republic his gradualism had imagined, in a state whose socialists governed exactly as he had advised and transformed exactly as little as Luxemburg had predicted; Luxemburg died at the hands of the order her old party chose to defend, having foreseen the choice and named its price. Neither outcome settles the argument. Both outcomes are data for its rows, and the German row of the ledger is the one every later participant consults first, because it is the rare case in which the same movement, in the same country, within the same months, was offered both roads in reality and the consequences of the one taken can be traced in documents rather than imagined.

Luxemburg on democracy: the revolutionary case against the revolutionary state

One document complicates every simple map of this debate and deserves its prominence: the manuscript Luxemburg wrote in prison in 1918 on the Russian revolution, published after her death, in which the leading revolutionary critic of reformism turned her criticism on the revolutionary government she otherwise defended. She supported the Bolsheviks’ seizure of power as an act of historical necessity and praised their courage in terms her social democratic opponents found scandalous, and in the same pages she attacked, with precision and foresight, their suppression of the constituent assembly, their restrictions on press and association, and their substitution of party rule for the broadest public life. Her reasoning was not liberal ornament. It was the theory of this debate applied to its newest case: without the unrestricted political activity of the broadest masses, she argued, public institutions die, bureaucracy becomes the only active element, and the dictatorship of a class degenerates into the dictatorship of a committee over the class. Her sentence on freedom, that it is always the freedom of the one who thinks differently, has been quoted wherever dissent needs a patron, and it carries ten words of a manuscript that runs to a systematic case.

The manuscript matters to the grid because it refuses the allocation of virtues on which both camps’ propaganda depends. The reformist tradition does not own democracy in this debate, since its German custodians in 1919 defended their road with the help of irregular troops against elected workers’ councils; and the revolutionary tradition does not own the critique of bureaucracy, since its most penetrating early statement came from inside the revolutionary camp at the moment of its triumph. Luxemburg’s position also supplies the revolutionary answer to the strongest objection when that objection is redirected: if the record of revolutionary states is the refutation of the revolutionary road, her manuscript is evidence that the record was not the unfolding of a single logic but the outcome of choices she identified and opposed from within, in real time, at the cost of her standing with the government she defended. Whether her alternative, soviet democracy with full political freedom, was available in the Russia of 1918, blockaded, invaded, and collapsing, is a question historians argue with the documents open. That the argument exists at that level, conducted by the debate’s own protagonists over the debate’s own values, is what separates the real argument of this article’s title from the cartoon of it.

The apparatus problem: Michels and the sociology of the party itself

Between Przeworski’s arithmetic and the founders’ polemics stands a third kind of explanation for the pattern both describe, and it came from inside the movement’s own house. Robert Michels, a sociologist who had been a member of the German party and moved in its intellectual circles, published Political Parties in 1911, a study of the socialist parties of Europe, above all the German, which concluded that organisation itself generates oligarchy as reliably as any law of politics he could formulate. A mass movement requires officials. Officials acquire skills, information, salaries, and a corporate interest in the survival and respectability of the organisation that employs them, interests which diverge from the transformative goals in whose name the organisation recruits. Leaders become a stratum, the stratum becomes cautious, caution becomes doctrine, and the party founded to conquer the existing order develops a vested interest in the tranquillity of the order that guarantees its offices, newspapers, and funds. Michels called the tendency the iron law of oligarchy, and drew from it a pessimism about democracy in general that eventually carried him to destinations the socialists who had hosted him found repellent.

For the reform or revolution debate, Michels supplies the mechanism that connects its levels. Przeworski explains why a vote seeking party moderates its programme. Michels explains why a party may continue moderating after the votes have been won, why the apparatus can resist a radicalised membership as stubbornly as it resists the state, and why the German party of 1914 behaved as an institution defending its assets rather than as a movement executing its programme. His law also reaches the revolutionary side of the street, where its application is, if anything, more damaging: the parties built to escape reformist oligarchy by discipline and centralism reproduced oligarchy in concentrated form, with the apparatus commanding not merely newspapers and funds but, after power, the state itself. A law that explains both camps’ degenerations cannot be the property of either, and its presence in the analysis disciplines the objection from the universal pattern in a useful way. The pattern of moderation in office may owe as much to the sociology of organisation, which afflicts every large body human beings build, as to the specific structure of electoral politics or of capital. Strategy argued at the level of roads and destinations must eventually descend to the level of institutions and incentives, and Michels is the descent’s unavoidable station.

The trade union front of the argument

The debate was never only about parties and parliaments, and its trade union front clarifies positions that the political front blurs. Bernstein valued the unions as the economic arm of democratisation, schools in which workers learned administration, negotiation, and the habits of incremental advance, and he resisted the subordination of union judgment to party doctrine with a consistency his opponents found exasperating. Luxemburg’s position was again more dialectical than its caricature. She defended the unions’ daily wage struggle as indispensable and attacked, in the mass strike pamphlet and elsewhere, the emerging theory and practice of a separate union bureaucracy that treated its funds and its neutrality as ends in themselves, hoarded its strength against political use, and in 1914 delivered its members to the war with the same institutional reflex the party displayed. The economic struggle itself she analysed as bounded by laws it could strain but not repeal: within a functioning labour market, unions can redistribute the terms of the wage bargain, defend real wages against encroachment, and organise the class, while the wage relation itself and the unemployment that disciplines it persist beyond the reach of bargaining as such. Union success, on her account, resembles the labour of Sisyphus only if it is mistaken for the goal; as the means by which a class maintains itself in fighting condition, it is beyond price.

Later history wrote large entries on this front of the grid. The union movements of the postwar decades were the load bearing institution of the reformist achievement, delivering through collective bargaining much of what legislation promised, and their decline across the advanced economies from the 1970s onward tracks the reversal of that achievement with uncomfortable fidelity, an association both camps can cite: reformists, as evidence that the ameliorative road was real while its organisations stood; revolutionaries, as evidence that gains held by organisation rather than by transformation of ownership last exactly as long as the organisation does. The union front also generated the debate’s most durable hybrid, the syndicalist and later the council traditions, which proposed to locate transformation neither in parliament nor in the party but in the organs of workplace self organisation themselves, federating the means of production under their producers. Those traditions belong to their own articles in this series, but their existence is a standing reminder that the two roads of the slogan were never the only paths the movement imagined, and that the debate’s binary form owes as much to the accident of two famous titles as to the logic of the choices themselves.

What the founders left unsaid: Marx and Engels on legal and peaceful roads

Both sides in 1899 claimed the founders, and both could quote them, because Marx and Engels had left the strategic question genuinely underdetermined by their texts. Marx’s own statements point in more than one direction, and the directions correspond to a distinction his followers often flattened: between countries with developed democratic institutions and countries without them. In a speech at Amsterdam in 1872, closing the Hague Congress of the First International, he allowed that in some countries, naming America, England, and possibly Holland, the workers might attain their goal by peaceful means, while in most of continental Europe, where despotism or its residues ruled, force would be the lever. The concession was conditional and geographical, not a doctrine of the parliamentary road, and Marx elsewhere wrote of the Paris Commune that the working class could not simply lay hold of the ready made state machinery and wield it for its own purposes, the sentence on which every later theory of the state’s resistance to its new occupants rests. Engels, in the last year of his life, wrote an introduction to a new edition of Marx’s Class Struggles in France that surveyed the movement’s electoral growth with undisguised satisfaction, argued that the era of barricade fighting by minorities had been made obsolete by modern armies and modern streets, and presented legal agitation and the ballot as the weapons under which the party had grown irresistible. Bernstein quoted the introduction as a testament. The party leadership, alarmed at its use by the Prussian authorities as evidence and by reformists as a licence, published it in softened form, and Engels protested in private letters that his tactical advice for a specific German situation was being converted into a universal recipe for legality at any price.

The founders’ ambiguity was not carelessness. It reflected a theory in which the form of transition depends on the institutional terrain a movement confronts, and the terrain varies by country and period in ways no text written in London exile could legislate for Warsaw, Paris, and New York at once. What the texts did not contain was any account of the mechanism by which reforms accumulate into transformation, because Marx’s critique of the Gotha Programme, the closest he came to addressing a party’s practical platform, concerned the distribution of the social product in a first phase of a society already transformed, not the ladder by which an untransformed society climbs. The revisionist debate can therefore be described, without disrespect to any participant, as the argument the founders made necessary and impossible to settle by citation. Every generation since has repeated the structure: an ambiguous inheritance, a changed terrain, and two camps each holding genuine quotations. The grid’s contribution is to relocate the argument from the founders’ intentions, where it is undecidable, to the four questions, where evidence can reach it.

Millerand and the ministerial question: may a socialist sit in a cabinet of the old order?

A companion controversy of the same years posed the office question in miniature and with useful concreteness. In 1899, the year of Bernstein’s book and Luxemburg’s reply, the French socialist Alexandre Millerand accepted a portfolio in the cabinet of Waldeck-Rousseau, a government of republican defence formed during the Dreyfus crisis, in which he sat alongside the general who had suppressed the Paris Commune. The case for acceptance was the case for office in its most seductive form: the republic itself was in danger from a nationalist and clerical right, a socialist minister could protect the movement’s legal existence, deliver tangible reforms, and demonstrate governing competence to a watching electorate. Millerand did deliver reforms in labour law and hours, and his defenders, led in the French movement by Jean Jaurès, argued that refusing power when it is offered, on grounds of doctrinal purity, abandons the field to worse hands and confuses the movement’s principles with its comfort. The case against, pressed by Jules Guesde and adopted by the international movement in resolutions that grew stricter from Paris in 1900 to Amsterdam in 1904, was the case the grid’s third row formalises: a socialist minister in a government he does not control becomes responsible for that government’s acts, lends it his movement’s credibility, and discovers that the portfolio confers duties more reliably than power, while the cabinet’s collective discipline forbids him the independence his voters imagined they were purchasing.

The sequels gave the argument its evidence. Millerand’s own career carried him steadily rightward, through the presidency of the republic, ending as a figure of the conservative order his younger self had entered a cabinet to resist, a trajectory his critics read as the biography of a mechanism rather than of a man. The ministerial question recurred wherever the electoral road prospered, in the wartime coalitions of 1914, in the popular front governments of the 1930s, in the coalition arithmetic of the postwar decades, and the pattern the Guesdists predicted and the Jaurèsians disputed repeated with variations: participation delivered real but bounded goods, cost the participating party a portion of its identity and eventually of its electorate, and never once served as the bridge to transformation its defenders hoped or its opponents feared. The episode’s analytical value is that it isolates a variable the larger debate entangles. Holding a ministry is office at its most personal and testable, and the century’s verdict on ministerialism is the third row’s verdict in small: office is real, its goods are real, its constraints are structural, and nobody has shown the mechanism by which a seat in someone else’s government becomes command over the state.

The agrarian complication: the peasantry inside the electoral arithmetic

Bernstein’s statistics concerned the town, but the electoral arithmetic of the countryside pressed the same challenge in a form Kautsky himself had to answer. The peasantry of Europe, small proprietors working family farms, was the largest single class in most countries the movement hoped to win, and the orthodox expectation treated it as a residue: capital would grind the peasant between the millstones of competition and debt until the holding dissolved and its owner descended into the proletariat, simplifying the social map in the countryside as in the factory. Kautsky’s Agrarian Question, published in the debate’s central year of 1899, gave the orthodox answer its most rigorous form, arguing that large scale production’s technical superiority would indeed subordinate agriculture, while conceding, in the analysis’s honest detail, how slowly and unevenly the subordination proceeded and how tenaciously the family farm survived by self exploitation, working longer for less than any wage labourer would accept. Bernstein fastened on the concession: a class that survives by overwork does not vanish on schedule, and a party that awaits its vanishing forfeits the countryside for generations to parties of order and altar.

The political consequences ramified through the whole debate. A movement that needs peasant votes must moderate its programme on land, since expropriation threatens the proprietor the party is courting, and Przeworski’s dilemma thereby acquires a rural form decades before its formal statement: the broader the coalition, the more of the founding programme is mortgaged to allies whose property the programme exists to abolish in its urban form. Revolutionary strategy faced the mirror problem, discovering in Russia and later in China that a revolution in a peasant country must either carry the peasantry with promises of land that create a new class of proprietors, or collide with the class it has armed. The agrarian complication thus belongs inside the grid rather than beside it. It is the second row’s accumulation problem and the dilemma’s majority problem wearing rural dress, and it explains why the debate’s map was always drawn most confidently in the industrial cities where both camps’ facts were thinnest on the ground beyond them.

The council form: an institutional answer to the office question

One tradition inside the wider movement proposed to escape the office question by changing the institution through which power is exercised, and its proposal recurs too persistently to omit. The workers’ council, appearing in Russia in 1905 and again in 1917, and across central Europe in 1918 and 1919, in Turin in 1920, and in later echoes elsewhere, is a body of delegates elected in workplaces, recallable by their electors, combining legislative and executive functions, and rooted in production rather than in territorial constituencies. Its theorists argued that parliamentarism’s separation of the voter from power, exercised once every few years through representatives the voter cannot recall and an apparatus the representatives cannot command, was not an accident of bad constitutions but the form of a state built to be operated by others, and that a republic of councils would keep power inside the class that makes it, dissolving the gap between office and power by redesigning office itself. Luxemburg’s democratic version of the form, with full political freedom and competing tendencies inside it, has already been examined in this article. The council idea in its broader career supplied the debate with its only genuinely institutional third option, neither Bernstein’s parliament gradually socialised nor the party state that the Russian road in fact produced.

The record of the councils is a record of brilliance under siege conditions and of defeat or absorption everywhere they contended with states that retained armies, treasuries, and time. Councils arose in collapse, when the old power had already fallen, and their problem was never inventing a form but surviving the interval in which the form had to feed cities, run railways, and face enemies organised on the old lines, tasks at which improvised federations of delegates repeatedly faltered or were displaced, by parliamentary republics in Germany, by the party state in Russia, by military reaction in Italy’s aftermath. Defenders answer that the councils were never tried in the conditions their theory required, a functioning economy and a breathing space, and that their suppression by both of the debate’s great camps, social democratic governments and Bolshevik centralisers alike, testifies to the threat the form posed to every apparatus. Critics answer that a form which appears only in collapse and survives only in the absence of enemies has specified the conditions of its own impossibility. The council tradition’s permanent contribution is diagnostic rather than programmatic: it located the debate’s hardest problem exactly where the grid’s third row puts it, in the design of institutions rather than in the intentions of the people who staff them.

The national frame: reform inside a world market

Every strategy in this debate was framed inside a nation, and every one of them confronted, sooner or later, the fact that capital is not. The constraint operates through channels the founders catalogued in other contexts and which the twentieth century widened into motorways: a government whose programme alarms the owners of mobile wealth watches its currency sold, its borrowing costs rise, and its investment fall, not because any committee of capitalists has voted against it but because thousands of portfolio decisions aggregate into a verdict no election can repeal. Luxemburg’s own economics placed the world market at the centre of the system’s dynamics, arguing that capitalism’s need for non capitalist markets drove its expansion and its violence, and Bernstein’s stabilisers included the world market’s shock absorbing breadth, so both founders of the debate understood that the national frame was a simplification. Their strategies nevertheless proceeded nationally, because elections, parties, unions, and states are national, and the mismatch between a national instrument and an international object became one of the reforming road’s defining experiences.

The experiences accumulated into a literature of their own. Governments of the left in France in the 1930s and again in the early 1980s met the constraint in its classical form, capital flight and currency crisis forcing programmatic retreat within years or months of victory, and the architects of the second episode drew the conclusion that the national road had narrowed past usability inside an integrated European economy, redirecting their tradition’s hopes toward building the missing frame at continental scale. Revolutionary governments met the same constraint from the other side, as blockade, intervention, and the discipline of trading with a hostile world market on its terms. The grid absorbs the point without needing a fifth row: the world market is the largest single entry in the third row’s column of constraints on office, and its growth since the founders wrote strengthens the evidence that office and power are different things, while leaving the second row’s question, whether coordinated reform across a sufficiently large frame could accumulate into transformation, exactly as open as the national experiments left it, and considerably harder to test.

The ratchet and its reverse gear: how settlements are undone

Reformist strategy relies, usually without saying so, on a ratchet: the assumption that gains, once won, are locked in, so that the movement’s task is always to add and never to defend from zero. Revolutionary criticism replies that the ratchet has a reverse gear, operated by the same institutions through which the gains were won, and the historical literature of reversal is now large enough to have a mechanics of its own. Settlements are undone by legislation when majorities change, by judicial reinterpretation when courts turn, by administrative starvation when budgets are squeezed without any vote that headlines can report, by economic reorganisation that hollows a right’s material basis while leaving its text intact, as the decline of union density undid the practical content of labour law across several countries, and by the slow conversion of universal institutions into residual ones, whose narrowed constituencies can then be safely attacked. The British reversal after 1979 used every one of those channels in sequence, which is why it serves this article as the type case of the reverse gear in motion.

The ratchet’s defenders answer with the asymmetry of the evidence. Some creations of the reforming era proved effectively irreversible by the same test: no government of the advanced democracies has abolished universal health coverage where it was built as a universal right, social insurance’s core survives every change of administration, and the franchise, once extended, has never been durably withdrawn in a functioning democracy, suggesting that reforms which create universal constituencies and are experienced as rights of citizenship acquire a defensive strength that reforms experienced as favours or as sectoral property never do. That answer, examined closely, is Gorz’s distinction returning through the evidence: the durability of a reform tracks the power relations it creates, which is what the non-reformist reform criterion predicted. The ratchet debate therefore ends where several of this article’s threads end, in a conditional proposition both camps can use. Gains endure in proportion to the organised power and universal stake that stand behind them, and they dissolve in proportion as their beneficiaries become clients of a benefit rather than citizens of an institution. Accumulation, if the word is to mean anything testable, should be measured by that standard rather than by the statute book’s length.

The Popular Front test: reforming governments between the wars

The decades between the founders’ deaths and the postwar settlement ran the electoral experiment under the harshest available conditions, and the French Popular Front of 1936 is its most instructive single run. A coalition of socialists, radicals, and communists won the election of that year against the background of fascism’s advance and a general strike that swept the factories as the government formed, and Léon Blum’s administration, governing with communist support it did not include in the cabinet, delivered in its first weeks the Matignon agreements: collective bargaining rights, substantial wage increases, the forty hour week, and paid holidays for workers who had never taken a holiday in their lives. The achievements were real, popular, and woven permanently into French life; paid holidays in particular became one of those reforms that no subsequent regime dared repeal. Within a year the government had fallen, broken on the mechanisms the grid’s third row enumerates: capital fled the franc, the senate blocked its financial powers, employers organised against the forty hours, and the coalition’s own breadth, the price of its majority, forbade the measures its left wing demanded to answer the flight. Blum’s formula for his government, that it would exercise power without holding the fullness of power, deserves its place in the debate’s permanent record as the third row stated by a practitioner at the moment of its operation.

Spain’s parallel experience ran the same sequence at higher temperature and with a catastrophic ending in civil war and dictatorship, and the two cases together formed the strategic imagination of the generation that rebuilt the European left after 1945. The lessons drawn were contradictory in the usual pattern. One lesson held that the experiment had been sabotaged by the narrowness of its mandate and the disloyalty of the institutions around it, and that the answer was broader alliances and deeper preparation, the conclusion Berlinguer later formalised. The opposed lesson held that the experiment demonstrated the ceiling of the form itself: a coalition that must reassure the middle strata and the banks to exist will be destroyed by the same reassurance when it acts, and the interval between victory and destruction is best used otherwise. The interwar record also supplied a control case of a darker kind in Germany, where the party of the reforming road, governing Prussia and sharing national office through the republic’s crisis years, administered austerity under emergency decree as the depression deepened, believing legality and order its best defence, until the republic it was defending was handed to its enemies by the men who commanded the state’s heights. No strategist of any camp has since been able to treat the defence of democratic institutions as a simple good separable from the question of who holds power inside them, and the interwar chapters are the reason the third row of the grid is written the way it is.

The debate restated after the war: Crosland, Bad Godesberg, and the Swedish model

The postwar generation restated the debate inside the reforming tradition itself, and the restatement is the strongest evidence that Bernstein’s questions were permanent rather than dated. In Britain, Anthony Crosland’s The Future of Socialism, published in 1956, performed for the Labour Party the operation Bernstein had performed for the German: an audit of the inherited doctrine against the transformed facts. Capitalism, Crosland argued, had changed its nature under full employment, the welfare state, and the managerial reorganisation of the firm; the Marxist categories of ownership and crisis no longer described the society British socialism actually faced; and the tradition’s goal should be restated as equality, pursued through education, social services, and the distribution of income and opportunity rather than through further nationalisation for its own sake. The parallel with 1899 extends to the reception: Crosland’s critics inside his party accused him of dissolving socialism into liberal amelioration, the precise charge Luxemburg laid against Bernstein, and his defenders answered with Bernstein’s answer, that loyalty to the evidence is the tradition’s scientific honour. In Germany, the Social Democratic Party completed the journey institutionally at Bad Godesberg in 1959, adopting a programme that renounced the Marxist framework, embraced the market economy’s framework of competition and private ownership, and redefined the party as a people’s party open to all strata, the formal registration, sixty years on, of the practice Bernstein had described and the goal he had dissolved.

Sweden’s labour movement supplied the restatement’s most intellectually ambitious variant, and the contrast with its neighbours is instructive. The Swedish model of the postwar decades, built by the Rehn Meidner strategy, was not mere amelioration: its solidaristic wage policy deliberately used centralised bargaining to squeeze low productivity firms out of existence while active labour market policy moved their workers into expanding sectors, making the union movement an engine of economic restructuring rather than a brake on it, and its architects understood the system as a functional equivalent, inside capitalism, of some of planning’s claims. When the same movement later proposed the wage earner funds, it was attempting to add the missing element, ownership, to a model that had already pushed the other elements to their limits, and the funds’ defeat therefore stands in this article’s evidence as the point at which the most sophisticated reformism ever constructed met the second row of the grid. The postwar restatements, taken together, changed the debate’s centre of gravity in a way the grid records faithfully. After Crosland and Bad Godesberg, the argument was no longer between a revolutionary movement and its revisionist wing. It was between traditions that had formally abandoned transformation and a shrinking company of those, inside and outside the old parties, who insisted the accumulation question remained unanswered rather than merely unfashionable.

The sheltered settlement: why the postwar decades flattered both sides

The postwar decades occupy a peculiar place in the evidence because they supplied both camps with their favourite exhibits, and an analytical account must explain the peculiarity rather than merely inherit it. The reforming road’s greatest achievements cluster in the thirty years after 1945, and the revolutionary expectation of deepening crisis suffered its longest embarrassment in the same years. The clustering was not an accident of political virtue. The settlement rested on material conditions that amounted to a shelter: capital controls under the Bretton Woods arrangements limited the flight mechanism that had destroyed Blum; the destruction of capital values in war and the backlog of reconstruction demand sustained investment; the United States underwrote the system’s security and its markets; energy was cheap; and the memory of the 1930s, held by electorates and elites alike, made the price of social peace seem low to those who set it. Under shelter, the dilemma relaxed. Governments could tax, spend, and regulate while capital, unable to leave and facing profitable demand at home, invested anyway, and parties could broaden their coalitions without immediately mortgaging their programmes to the confidence of mobile wealth.

The shelter’s dismantling, from the collapse of Bretton Woods in the early 1970s through the liberalisation of capital movements in the following decades, is therefore among the most consequential entries in the whole record, because it changed the parameters within which every row of the grid operates. Mobile capital restored the third row’s veto in a stronger form than the founders knew, since the portfolio decisions that discipline governments became continuous and instantaneous rather than episodic. The second row’s prospects narrowed correspondingly: accumulation strategies that presuppose a national economy whose commanding firms can be gradually socialised confront firms whose ownership, production, and tax residence are distributed across jurisdictions by design. None of this settles the debate, and both camps have drawn their morals with the usual symmetry: reformists conclude that the frame must be rebuilt at the scale capital operates, revolutionaries conclude that a system whose stability requires capital’s permission has confessed the location of power. The analytical point stands beneath both morals. The postwar decades were an experiment conducted under shelter, and arguments that generalise from them without naming the shelter, in either direction, are arguing from the grid’s most misleading row of evidence.

The means question the slogan hides: force, legality, and the defence of gains

A further bundle hides inside the slogan, closely related to the four questions but distinct enough to name separately, and it concerns means rather than destinations: what part force plays in transformation and in the defence of whatever transformation achieves. The caricature allocates violence to one camp and legality to the other, and the founders’ texts dismantle the allocation from both sides. Luxemburg opposed putschism, terrorism, and the substitution of a determined minority’s action for the movement of a class with a consistency that cost her standing among the impatient of her own camp, and her theory of the mass strike assigned the initiative to the spontaneous action of millions, in which force appears as the response of a society in motion rather than as a technique applied to it. She also held, against the legalists, that a ruling order facing the loss of its property would not reliably respect the legality it had written for quieter times, and that a movement which renounced in advance every means beyond the ballot had announced the limit at which its opponents could safely stop it. Bernstein’s position was the mirror image with the same structure: he defended legality as the terrain on which his movement was strong and his enemies divided, while the movement he advised kept its distance from the converse promise, that it would acquiesce in the abolition of the democracy through which it worked. The German party’s catastrophe of 1933, in which a legalist leadership, commanding mass organisations and a paramilitary formation of its own, declined every extra legal defence of a republic being dismantled around it, stands in the record as the means question’s hardest case, and every later strategist of the democratic road has had to write in its shadow, usually by arguing that the case’s conditions, a lost election, a divided left, an army and judiciary already hostile, will not be theirs.

The analytical yield of separating the means question is the discovery that it does not align with the others. Positions on force cut across positions on accumulation: history offers legalist revolutionaries and insurrectionary defenders of gradual gains, reformists who built militias to protect printing presses and revolutionaries who forbade their followers to answer provocation. The means question also has its own evidence, distinct from the grid’s rows, in the record of how transformations and settlements actually ended: rarely by persuasion, often by exhaustion, and at the decisive points by the disposition of organised coercion, which is why the state theory tradition treats the armed apparatus as the final entry in the third row’s inventory. A debate that keeps the means question bundled inside the destination question will keep mistaking arguments about one for arguments about the other, and the founders, who argued about both with more care than their slogans suggest, are the first witnesses for the separation.

What would be built: the destination problem inside the accumulation question

The accumulation question conceals a further difficulty that both camps, for most of the debate’s history, preferred to leave unexamined, and naming it completes the analytical survey. Suppose the reformist mechanism worked exactly as its strongest version claims: reforms accumulate, the balance shifts, and a government commanding a durable majority begins the transfer of productive ownership to social hands. The question that then arrives, and that no quantity of prior reform answers, is what social ownership consists in as an operating system: which body decides what is produced and invested, by what signals it learns what is needed, how managers are chosen and removed, what disciplines failure, and how the liberties of minorities and dissenters are secured inside an economy whose commanding positions are publicly held. The revolutionary tradition’s historical answer, central planning by a party state, collapsed as an exhibit between 1989 and 1991, and its collapse damaged the reformist road as well, since a gradual transfer toward an unspecified destination is a journey whose later stages no one can describe, a fact electorates noticed. The market socialist literature, descending from the interwar calculation debate in which Lange proposed that public ownership could use prices as its signals, supplies designs of varying completeness, and the cooperative and municipal traditions supply working fragments, but the debate’s mainstream, on both sides, spent a century arguing about the road while treating the destination as a promissory note.

Bernstein’s gradualism had, in this respect, a hidden advantage he never quite claimed: a transformation defined as a direction rather than a blueprint does not need the destination specified in advance, since each reform can be judged by the criteria of the present, more security, more equality, more democracy, and the sum can be allowed to be whatever it becomes. Luxemburg’s tradition had the corresponding burden in its sharpest form, since a rupture justified by its destination must say what the destination’s institutions are before the rupture’s costs can be weighed against them, and her own writings on democracy after the revolution, with their insistence on unrestricted public life as the new order’s operating condition, are best read as her attempt to pay that burden honestly. The destination problem does not change the grid’s verdicts. It explains one reason the second row has resisted settlement even in argument: the accumulation question cannot be fully answered until the thing accumulated toward is specified, and the specification, attempted seriously, divides each camp internally as deeply as the road divides the camps from each other.

Other homelands of the argument: possibilism, Fabianism, and the roads not taken in Germany

The German debate became canonical because the German party was the movement’s flagship and its antagonists its finest writers, but the argument had other homelands with other chemistries, and two of them clarify what was contingent in the German form. In France, the split over gradualism predated Bernstein by a generation: the possibilists of the early 1880s, associated with Paul Brousse, broke with the Guesdist orthodoxy precisely over the attainable, arguing that a programme should demand what each moment makes possible and treat municipal office, where socialists could pave streets, run services, and demonstrate competence, as the school and showcase of the larger ambition. French gradualism was thus born as a factional programme inside Marxism before Bernstein supplied its theory, and its later career through Millerand to the Popular Front gave France the debate’s most complete national record, the same sequence of office, achievement, constraint, and reaction running at shorter intervals and higher speed. In Britain, the Fabian Society built gradualism without a Marxist host to revise: the Webbs and their associates proposed to permeate the existing parties and the civil service with collectivist measures, trusting expertise and administrative momentum rather than class mobilisation, and their gradualism differed from Bernstein’s in the decisive respect that it had no revolutionary past to shed and no mass party of its own to hold together. British reformism’s characteristic products, the municipal enterprise of the progressive cities and eventually the administrative architecture of the welfare state, bear the Fabian stamp of the blueprint drawn by experts, a stamp whose strengths and limits the British record displays in pure form.

The comparison sharpens the German case by contrast. Bernstein’s revision was explosive because it occurred inside a doctrine that promised rupture and a party that practised reform, so that the argument was always also about honesty, about whether words and deeds would be brought into line by changing the words or the deeds. Where no such doctrine existed, as in Britain, gradualism needed no revision and provoked no Luxemburg, and where the movement was weaker and more fragmented, as in France, the argument dissolved into factional chronology rather than concentrating into classic texts. The Scandinavian movements supplied a fourth chemistry, in which strong unions and a disciplined party achieved the reforming road’s best results while retaining, well into the postwar period, a transformative self understanding that their German sister party formally renounced, a combination that suggests the doctrine mattered less to outcomes than the organisation and the shelter, a suggestion both camps find uncongenial and the evidence keeps making. The debate’s geography thus teaches the same lesson as its grid: the variable that decides what a strategy becomes is rarely the slogan under which it marches. It is the composition of the coalition, the strength and independence of the organisations, the posture of the state, and the shelter or exposure of the national economy, the entries, in short, of the rows themselves.

Why the slogan survives its analysis: the uses of an unsettled argument

If the separation of the questions is as clarifying as this article claims, a fair reader may ask why the bundled slogan has survived a century of intelligent criticism, and the answer completes the analysis by turning it on the debaters themselves. The slogan persists because it does work that analysis does not. For a movement, a binary choice between roads is an instrument of identity: it sorts members from sympathisers, gives a founding story its villains and its faithful, and converts the slow ambiguities of strategy into a loyalty that can be tested in a single question. Parties preserve their founding quarrels the way nations preserve their founding wars, as charters, and a cadre formed on the story of Bernstein’s betrayal or of Luxemburg’s martyrdom possesses a usable past that a cadre formed on a four row grid conspicuously lacks. The slogan also performs recruitment’s oldest trick, borrowing the energy of the destination to animate the means: few organisers ever filled a hall by promising a marginally better administration of the existing order across three generations, and few revolutionaries ever filled one by dwelling on the apparatus problems of the morning after. Each camp’s rhetoric therefore inflates exactly the component its analysis knows to be doubtful, and the inflation is not hypocrisy but the occupational physics of mobilisation, which runs on certainty the way engines run on fuel.

Recognising the slogan’s functions explains the pattern of its flare ups. The argument returns to incandescence at moments of defeat and of unexpected victory, when movements must explain failure or decide what success is for, and it subsides into scholasticism during the long periods when organisations are busy surviving. It also explains why the debate’s analytical advances, the grid’s separations, Gorz’s criterion, Przeworski’s arithmetic, Michels’s law, have been absorbed by scholars and largely ignored by the movements whose ancestors produced them: an instrument for thinking is not an instrument for marching, and the two instruments obey different designs. None of this makes the analysis futile. It locates its audience and its use. The reader who has separated the questions cannot be recruited by the bundle, cannot be told that a welfare state refutes a theory of ownership or that a failed ministry refutes a century of reform, and can follow any future form of the argument, under whatever new names its parties travel, by asking the four questions in order and listening for which one the speaker is avoiding. That is the competence this article set out to build, and the slogan, survived by its analysis, remains available as the map’s legend rather than as its territory.

For students and exam candidates

The four questions grid is the essay structure this topic rewards, and candidates should build answers on it directly. Begin by separating the bundled slogan into its components: whether capitalism can stabilise itself, whether reforms can accumulate into a change of ownership, whether office delivers power, and whether reforms are worth winning regardless. Then show the convergence the separation reveals, that serious reformists and serious revolutionaries agree on the worth of reforms, largely agree on the constraints surrounding office, and divide at full strength mainly over accumulation. Support each row with one named case: the recurrence of crises against the stabilisation claim, the Swedish wage earner funds against the accumulation claim, the British government of 1945 for the difference between transforming conditions and transforming ownership. Bernstein should be presented through his evidence rather than his famous sentence, and Luxemburg through her distinction between means and end rather than through any supposed hostility to reforms. Keep a working file of the primary texts and the grid side by side; the study notes at VaultBook are built for exactly that kind of structured revision, with the four questions as headings and the cases filed beneath them. Examiners reward the candidate who can state what would count as evidence against a position, so end by naming the experiment that has never been run in clean form.

For teachers and seminar leaders

Teach this debate as an exercise in separating questions before staging it as a contest, since students arrive with the slogan and its caricatures pre installed. A productive sequence opens with the grid incomplete: give the seminar Bernstein’s four empirical claims and ask the room to audit each against the subsequent record before Luxemburg is introduced, so that revisionism is met first as an evidence based challenge rather than as a heresy awaiting its refutation. Introduce Luxemburg’s reply second, with explicit attention to her defence of the daily struggle, which dismantles the laziest assumption in the room. Przeworski belongs at the centre of the second half, because his dilemma converts the moral vocabulary of betrayal, which students find irresistible, into arithmetic they are obliged to argue with. Set the Meidner plan as the discussion case: it is the rare episode in which a movement designed a reform to test the accumulation thesis directly, and a seminar can divide over its dilution with genuine evidence on both sides. Close by asking each participant to state which row of the grid their own instinct answers, and to notice how much of the room’s disagreement survives the separation. Assessment should credit the location of claims on the grid at least as heavily as the verdicts reached.

For researchers, historians, and journalists

The research frontier of this debate is empirical, and it lies in the audit of specific claims rather than in further commentary on the canonical texts. Bernstein’s predictions repay disaggregation: the literature on the separation of ownership and control, on the recomposition of the middle strata, and on the socialisation of investment through pension funds each revises one component of his case while leaving the others standing, and work that treats revisionism as a single claim repeats the error the four questions grid exists to correct. The comparative history of reforming governments offers a second frontier, in particular the conditions under which settlements proved reversible, where the contrast between the durability of universal services and the fragility of public ownership in the British case awaits systematic treatment across countries. A third frontier is the formal literature descended from Przeworski, which has refined the electoral dilemma with better data on class composition and party strategy, and which historians have not yet absorbed as fully as political scientists. Journalists covering any movement that promises transformation through elections will find the debate’s vocabulary unavoidable and its grid directly usable: the questions that expose a programme are the grid’s second and third rows, asking by what mechanism reforms are said to accumulate into changed ownership, and what the programme expects from the apparatus, the currency, and the owners of capital in its first hundred days. The archival record, from the congress protocols of the German party to the Swedish union congresses that debated the funds, is rich, published, and still underworked relative to the commentary literature, and researchers who go to it will find the founders arguing with a precision their summarists rarely match.

Frequently Asked Questions

Q: Can socialism be achieved through elections?

The honest answer separates the question into parts. Elections can certainly be won by parties committed to socialism, and elected governments have used office to build welfare states, extend rights, and nationalise particular industries. What no elected government has done is complete a transfer of the ownership of the productive economy to society as a whole. Reformists argue that accumulated reforms, extended far enough, amount to that transfer by degrees. Revolutionaries argue that structural constraints, the dependence of the economy on private investment and of the government on an inherited apparatus, prevent the final steps. The question therefore turns less on elections themselves than on whether reforms can accumulate into a change of ownership.

Q: What was Bernstein’s revisionist argument?

Bernstein argued, at the end of the 1890s, that Marxist orthodoxy had misread the direction of capitalist development. Misery was not deepening as expected, share ownership and the middle strata were growing rather than vanishing, small firms persisted, and credit, cartels, and improved communications were making general crises less likely. From that evidence he drew a strategic conclusion: since capitalism was proving adaptable and democracy was proving real, socialism should be pursued as the cumulative result of reforms, union organisation, and democratic advance, rather than awaited as the sequel to a collapse. His formula that the movement is everything and the final goal nothing summarised the priority of the daily work over the prophesied destination.

Q: Was Bernstein right about how capitalism developed?

He was right in parts, and the parts must be scored separately. On living standards he read his period correctly, and the long rise of real incomes and social security in the advanced economies extended his point. On the middle strata he was substantially right: the space between capital and labour kept refilling with salaried, professional, and technical occupations instead of emptying. On ownership he confused the diffusion of shareholding with the dispersal of power, since control concentrated even as titles spread. On stability he was wrong: credit and combination did not prevent the crash of 1929 or the later crises, and at times amplified them. The mixed score is precisely why the debate he opened never closed.

Q: Why did Luxemburg reject gradualism?

Luxemburg rejected gradualism as a theory, not reforms as a practice. She held that reforms are the means of the movement and the conquest of political power its end, and that the two cannot be separated without changing the character of both. Her economic reason was that reforms, however many accumulate, operate inside the wage relation: they improve the terms on which labour power is sold without ending the sale, so quantity of reform never sums to a different quality of system. Her political reason was that detaching the movement from its goal converts a socialist party into an agency for administering capitalism more kindly, which she regarded as a different project wearing the same name. She defended unions, elections, and democratic rights throughout as indispensable means.

Q: What is the structural dilemma facing socialist parties?

The dilemma, given its classic form by Adam Przeworski, arises from the composition of the electorate. Manual workers have never formed a majority of voters in an advanced democracy, so a socialist party that appeals to workers alone on an undiluted programme assembles a minority and loses. To win, it must attract salaried employees, professionals, and other strata, and those allies will not support a programme that threatens their own positions, so the programme is diluted as the coalition widens. Electoral success and transformative ambition thus stand in structural tension, whoever leads the party. A second horn waits in office, where a government that exceeds its diluted mandate faces withheld investment and capital flight. The dilemma explains moderation by position rather than by betrayal.

Q: Did social democracy achieve anything socialist?

The answer depends on the measure, and both traditions supply one. By the measure of people’s security and freedom, social democratic governments achieved a great deal that their founders would have recognised as the substance of their hopes: universal welfare states, full employment for long periods, compressed inequality, public health and education, and strong rights at work, most completely in Scandinavia. By the measure of ownership, they achieved almost nothing durable: productive assets remained privately owned, the public sectors they built were largely reprivatised when political conditions changed, and wage labour survived untouched as the economy’s organising relation. Whether the first list counts as something socialist, or as capitalism humanised, is exactly the point the two traditions dispute.

Q: What is a non-reformist reform?

The term is André Gorz’s, from the 1960s, for a reform designed to shift the balance of power rather than merely to improve conditions within the existing balance. A reformist reform, in his contrast, is granted on the system’s own criteria, leaves the distribution of power unchanged, and turns its beneficiaries into clients. A non-reformist reform is won by organised pressure, creates rights and institutions its holders can defend and extend, raises demands the system cannot comfortably absorb, and is consciously one step in a sequence toward a different order. The Swedish wage earner funds are the classic attempted example. The practical difficulty, acknowledged by Gorz’s readers ever since, is that reforms often do both kinds of work at once, and which they do may only become clear in retrospect.

Q: Does winning office mean holding power?

No, and the distinction is one of the debate’s permanent acquisitions. Office means occupying the formal positions of government after an electoral victory. Power means the capacity to make decisions effective, and it depends on a civil service and judiciary the government did not appoint, on armed forces with their own corporate outlook, on treaties and central bank commitments, and above all on an economy whose investment decisions remain in private hands and can be withheld from a government their owners distrust. Governments repeatedly discover that they hold responsibility for outcomes before they hold command over causes. Reformists and revolutionaries dispute how far determination, preparation, and popular mobilisation can close that gap. Neither serious tradition denies the gap exists.

Q: Is the reform or revolution question still live?

It is live wherever a movement promises deep change through democratic means, because the questions composing it have never been answered by events in a form all sides accept. Its vocabulary changes with each era: the names of parties, the design of programmes, and the immediate policy stakes differ completely from those of 1899, while the underlying structure, a coalition broad enough to win being rarely radical enough to transform, and a government in office facing an economy it does not command, recurs in recognisable form. The debate also persists in scholarship, where the audit of Bernstein’s predictions and the formal study of electoral dilemmas remain active literatures. What has died is only the expectation, held by the founders on both sides, that history would shortly grade the argument for them.

Q: What would settle the reform versus revolution argument?

The accumulation question at its centre would be settled by the experiment history has never run in clean form: a government elected on an explicit programme of transferring ownership, holding a durable majority and a united movement behind it, free of war and overwhelming foreign intervention, persisting across several electoral cycles, and then either consolidating the transfer or being visibly stopped by identifiable mechanisms. Partial cases exist, in Sweden, Britain, and Chile, but each carries a confounding factor that partisans can cite, a narrow mandate, a divided movement, an external enemy, a programme that never aimed at ownership in the first place. Until a cleaner case exists, the argument will be conducted, as this article has conducted it, by separating its questions and auditing the evidence row by row.

Q: How does the four questions grid separate Bernstein from Luxemburg?

The grid shows that the two antagonists were not opposed at every point, which the slogan version of their quarrel assumes. Separated into the four questions, their disagreement concentrates overwhelmingly on one: whether reforms can accumulate into a transformation of ownership, where Bernstein’s whole strategy says yes and Luxemburg’s whole theory says no. On the worth of reforms they agree, since both organised for them relentlessly. On the power of office they are closer than the polemic admits, since Luxemburg analysed the limits of parliament as an instrument and Bernstein analysed the resistance a reforming government must expect. Even on stability, where they differ sharply, both are arguing from evidence about crises. The grid turns a clash of temperaments into a located dispute about accumulation.

Q: What should a historian check before calling a reform revolutionary?

A historian should check four things, matching the grid. First, what the reform did to ownership, not merely to conditions: did control over productive assets change hands, or only the distribution of income and security around an unchanged structure? Second, what it did to power: did it create institutions and rights that their holders could use to win further changes, or did it leave its beneficiaries dependent on the goodwill that granted it? Third, how it was won, since reforms conceded from above to forestall pressure have a different history from reforms extracted by organised movements. Fourth, what survived: whether the reform endured changes of government and economic weather, or was reversed at the first opportunity. A reform that changed ownership, shifted power, was won by organisation, and endured would have a strong claim to the adjective. Most celebrated reforms pass some of these tests and fail others, which is why the label needs the audit.

Q: Why did winning elections repeatedly fail to transfer economic power?

Because an election transfers office, and economic power sits largely outside the offices being filled. The owners of productive assets retain their property the morning after the count, and with it the decisions about investment, employment, and prices on which the government’s own success will be judged. A government that moves against ownership therefore faces an opponent who does not need to win a vote: capital can be withheld, moved abroad, or priced against the currency, and the resulting economic distress is charged to the government’s account by the electorate it must face again. Added to this stand the inherited apparatus, the courts, the treaty commitments, and the party’s own need to keep a broad coalition intact. Reformists reply that these are obstacles to be prepared for rather than laws of nature, but the mechanisms themselves are facts of the record, not inventions of either camp.

Q: How does Luxemburg’s mass strike argument answer gradualism?

The mass strike writings, composed after the Russian revolution of 1905, answer gradualism by challenging its picture of how change happens. Gradualism supposes a smooth ascent in which each reform is added to the last by deliberate decision. Luxemburg argued that real movements develop in waves: a strike begun over wages becomes a strike over rights, economic and political demands fuse, defensive actions turn offensive, and the boundary between a reform campaign and a revolutionary situation dissolves in practice, with the movement’s own activity educating it faster than any programme of preparation could. On that picture, the choice between reform and revolution is not a menu leaders select from in advance. Situations escalate by their own logic, and a party’s task is to be organised, present, and clear about its goal when they do. Gradualism errs, she held, about the rhythm of history as well as about the sum of reforms.

Q: What do reformists and revolutionaries unexpectedly agree on?

They agree on more than either camp’s polemic admits. Both hold that reforms are worth winning and that neglecting the daily struggle for them forfeits the people a movement claims to serve; Luxemburg insisted on this as firmly as Bernstein. Both hold that elections can be won and office used for real gains. Both, in their serious forms, hold that an elected government faces structural constraints from the apparatus it inherits and the economy it does not own, differing over whether persistence can overcome the constraints rather than over their existence. Both even agree that Bernstein’s empirical questions were the right ones to ask, since the revolutionary tradition spent a century answering them. The surviving disagreement, over whether reforms can accumulate into a transformation of ownership, is narrower than the famous quarrel and more precise, which is what makes it worth conducting carefully.