The withering away of the state is the most quoted promise in the Marxist tradition and the one its critics reach for first, because it seems to compress the whole doctrine into a single testable sentence: build the new society, and government itself will fade into a museum piece. Almost everything in that familiar summary is slightly wrong, and the errors compound. The phrase is not Marx’s but Engels’s. It was coined not as a free standing prediction but as one half of an argument with the anarchists, and its point was to refuse the demand that the state be abolished at a stroke. It presupposes a transitional state that does the opposite of fading for as long as classes survive. And the empirical record, which this article states plainly at the outset rather than apologising for later, is that no state has withered away anywhere, while several states constructed in the tradition’s name grew into the largest coercive apparatuses of their century. A glossary entry that hid any of those four facts would be a disservice dressed as a definition. The purpose here is precision: to recover what the claim actually said, to list the conditions it itself specified, and then to measure the record and the rival predictions against that specification.

Precision matters for this term more than for most, because the phrase does heavy rhetorical work for people who have never read the passage it comes from. Opponents quote it as proof that Marxism promises an impossible stateless Eden and then congratulate themselves when Eden fails to arrive. Sympathisers quote it as proof that the coercive states built under Marxist banners were betrayals of the doctrine rather than expressions of it. Both uses treat the phrase as a detachable slogan. In the texts it is nothing of the kind. It is a conclusion drawn from a theory of what the state is and where it came from, and it arrives attached to conditions that can be written down and checked. Restoring the term to its argument, which is the series habit this article follows, means setting out the theory of the state’s origin that makes disappearance thinkable, the sequence of stages the theory requires, the conditions named in the texts themselves, and the three great objections that were lodged against the idea, one of them before the ink of the founding quarrel was dry. Only after that work is done can the interesting question be asked with a straight face: was this a prediction that failed, a tendency claim that was never triggered, or a hope wearing the clothes of a law?

There is also a disciplinary reason for treating this term at length in a glossary series, and it is worth stating before the detail begins. Most contested terms in political theory are contested because their meanings are vague, and the glossary’s work is disambiguation. This term is contested for the opposite reason: its meaning, in the texts, is unusually determinate, while its fate is unusually well documented, and the gap between a precise claim and a closed record is where the interest lies. A reader who finishes this article should be able to do three things that general commentary on the phrase rarely equips anyone to do: state the claim in a form its author would recognise, list the conditions the claim itself specifies in a form a checker could use, and say which of the claim’s several statuses, prediction, theorem, tendency or hope, the evidence supports. Everything in the sections that follow serves one of those three competences, and the asides, the history and the quarrels are all, in the end, scaffolding for them.

What withering away means, and whose phrase it is

Begin with the definition the texts support. In the Marxist account, the state is not government in general, not administration, not coordination, and not every form of collective decision. It is a specific institution: a body of armed men, prisons, courts and officials standing apart from the population, whose function is to hold down the antagonism between classes that the economic order generates and cannot itself resolve. On that definition the state is a product of class division, historically late and functionally specific. Withering away names the process by which such an institution, once the classes whose antagonism called it into being have disappeared, loses its function and lapses, the way an organ lapses when the conditions that sustained it are removed. The image is biological on purpose. Nothing is smashed in the final stage and nothing is repealed. An institution dies because the work it existed to do no longer exists.

Attribution is the first precision, and it is the one print journalism most often fumbles. The canonical formulation stands in the third part of Engels’s Anti-Duhring, the polemic of 1878 against the Berlin lecturer Eugen Duhring, in the chapter that lays out the theoretical groundwork of socialism. Marx wrote no sentence of that book, though he read the manuscript, contributed a chapter to its economic part, and allowed his name to be associated with the defence of their shared position. When later editions and textbooks speak of Marx teaching that the state will wither away, they are crediting the wrong partner. Marx’s own vocabulary for the end of the state is different and, characteristically, less pictorial, and the difference between the two vocabularies is itself instructive, because it maps onto a difference in what each man chose to emphasise: Marx the transitional political form, Engels the long run destination.

Did Marx say the state would wither away?

The phrase is Engels’s, from Anti-Duhring. He argued that the state is not abolished, as the anarchists demanded, but dies away once classes disappear. Marx wrote of a transitional period of proletarian rule between capitalism and communism. Neither man described immediate abolition, and no state has withered.

The misattribution is understandable, which is why it survives. Marx and Engels presented a united front in their lifetimes, Engels spent his widowhood of the partnership editing Marx’s manuscripts and explaining their joint doctrine to a movement that treated the two names as one, and the phrase entered party education through Engels’s popularisations, above all through the short work Socialism: Utopian and Scientific, which extracted three chapters of Anti-Duhring and carried the formulation into dozens of languages. A reader who met the idea there met it as the doctrine of Marxism, and the movement’s own pedagogy rarely paused over which partner had drafted which sentence. Scholarship has no such excuse. The division of labour between the two men is documented in their correspondence and in the publication history of every major text, and a reference work that prints the phrase under Marx’s name is making an error of fact, not adopting an interpretation.

There is a second reason to insist on the attribution, beyond pedantry. Engels’s formulation carries commitments that Marx’s scattered remarks do not obviously share, and conflating the two men hides a genuine question inside the tradition: how far did Marx endorse the destination his partner described in such detail? Marx’s published references to the disappearance of political power are few and guarded. In The Poverty of Philosophy, his early polemic of 1847, he wrote that the working class would substitute an association in which the public power would lose its political character, a formulation that points the same way as Engels while stopping short of a timetable or a mechanism. In the Critique of the Gotha Programme, written in 1875 and published by Engels after Marx’s death, the question of what state functions remain in communist society is raised and deliberately left as a question, to be answered, Marx wrote, only scientifically and by the future itself. The asymmetry is real. Engels systematised; Marx, on this point, gestured and withheld. Keeping the phrase under its author’s name keeps that asymmetry visible, and the asymmetry matters when the claim’s status is assessed at the end of this article.

A third precision belongs in this opening section, because dictionaries routinely blur it. The state that withers is the state in the specific sense defined above: the separate coercive apparatus standing over society. Neither Engels nor Lenin, who took up the phrase and gave it its fullest elaboration, claimed that coordination, record keeping, planning or the management of common affairs would vanish. Those functions are expected to persist and even to expand; what disappears is their political character, their organisation as the monopoly of a separate body armed against part of the population. Whether that separation of administration from coercion is coherent is a serious question, and it gets its own section below, because the Weberian objection turns exactly on it. For the definition it is enough to hold the distinction steady: withering is predicated of the state as class instrument, not of organised collective life as such.

The broader framework into which this definition fits, the account of the state as a relatively autonomous apparatus arising from class antagonism and presenting itself as standing above it, is set out in the series treatment of the marxist theory of the state guide, and this article presupposes that framework without re-arguing it. What belongs here is the narrower glossary work: the phrase, its author, its contrast term, its conditions and its record.

The abolition contrast: the argument the phrase was built for

The single most valuable thing a reader can carry away from this article is the abolition contrast, and it deserves to be stated as a named claim because it reorganises everything else. The abolition contrast is this: withering away was formulated as a rejection of the anarchist demand for the immediate abolition of the state, and the phrase only makes full sense as one half of that nineteenth century argument. Engels did not coin a prophecy and then, separately, quarrel with anarchists. The coinage is the quarrel. In the sentence that made the phrase famous, the two verbs are set against each other like opposing counsel: the state is not abolished, it withers away. The first verb is denied; the second is asserted; and the denial is aimed at named contemporaries who had made abolition the centrepiece of their programme.

Those contemporaries were the anarchists of the First International, above all Mikhail Bakunin and his followers, for whom the state was not a symptom of class division but a primary evil in its own right, the source and guarantor of every other domination. On the anarchist analysis, revolution that left any state standing, even a revolutionary one, had already failed, because the apparatus would reproduce domination from its own resources regardless of the intentions of those staffing it. The anarchist demand therefore ran: abolish the state on the first day of the revolution, all states, everywhere, as the precondition of a free federation of communes and workplaces. Engels’s reply, sharpened across the polemics of the early 1870s and codified in Anti-Duhring, was that the demand confused a wish with a mechanism. A society emerging from capitalism still contained classes, scarcity and the habits of the old order. It would need coercion against the dispossessed possessing classes and against disorder generally. The instrument of that coercion would be a state, whether or not anyone chose the word. The honest course was to admit the transitional state, define it, and specify the conditions under which it could be expected to lapse, rather than to proclaim abolition and then improvise an apparatus in embarrassed silence.

Seen through the abolition contrast, the phrase changes character. Read alone, it sounds like a utopian guarantee, and it is attacked and defended as one. Read as the second half of the contrast, it is a claim about sequence and causation: coercion is a function of class antagonism, so coercion lapses when, and only when, its cause lapses; to demand that it lapse earlier is to demand effects before causes, which Engels regarded as the defining error of utopian politics. The phrase is thus less a promise than a refusal, a refusal to let the end of the state be scheduled by proclamation. This reading does not rescue the claim from its record, and nothing in this article will pretend otherwise. It does, however, fix what would have to be true for the claim to count as tested: the conditions named in the contrast, classlessness above all, would have to have obtained somewhere, and the apparatus would then have had to persist anyway, or to have dissolved on cue. The sections on the record and on the conditions defence return to that test and to the uncomfortable fact that it cuts both ways.

The argument with anarchism did not end in the 1870s, and its full history, from the Hague Congress expulsions through the rival strategies of the twentieth century, belongs to the companion treatment of marxism vs anarchism on the state, where the debate is given the room it needs. What this glossary article must hold is the minimum that makes the term intelligible: that the phrase was minted inside that debate, that its negative pole is abolition by decree, and that any reading which detaches the phrase from the contrast, whether to celebrate it as prophecy or to mock it as fantasy, has lost the thing the words were doing. Bakunin, notably, was not answered by the contrast so much as postponed by it, and his counter-prediction, that the transitional state itself was the danger, is treated at full strength in its own section below, because the brief of this article is that the standing test case for the whole doctrine is the man who lost the original argument.

One more feature of the contrast deserves attention, since it is rarely noticed and it disciplines loose talk on both sides. The contrast is asymmetric in its burdens. The anarchist demand could be satisfied by an act: a revolution could at least attempt abolition and be judged on the attempt. The Marxist reply could be satisfied only by a process measured in generations, and its authors knew it. Engels, in a letter to August Bebel in 1875, complained about the phrase free people’s state in the draft Gotha programme and remarked that the state, as the organisation of force, should be spoken of honestly during the transition and not prettified. The candour is worth noting because it complicates the lazy portrait of the withering doctrine as naive. Whatever its fate, it was not offered by men who thought the state a trifle that a proclamation could dissolve. It was offered by men who thought the state so stubborn a growth that only the disappearance of its root system would kill it.

The founding passage read closely

The Anti-Duhring paragraph rewards slow reading, because its argument is compressed into a chain of consequences that later quotation broke into slogans. Engels begins not with the state but with production. The proletariat, he writes, seizes state power and turns the means of production into state property, and he calls this act, in a phrase his readers have puzzled over for a century and a half, both the proletariat’s first act as ruling class and its last act as a state. The paradox is deliberate and it is the doctrine in miniature. The act of socialisation is the one exercise of state power that abolishes the basis of all state power, since by ending private ownership of the means of production it ends the class division that the state existed to manage. Every earlier state act had reproduced the conditions of the state’s necessity; this one, uniquely, consumes them. The elegance of the construction should be acknowledged before its difficulties are counted, because it explains the phrase’s hold on intelligent readers: the withering is not bolted onto the seizure of power as a consolation, it is derived from the seizure’s content, and a critic who wants the destination gone must break the derivation, not merely doubt the destination.

The derivation’s joints are, however, exactly where a close reading must press. The first joint is the identification of state ownership with social ownership. The act that sets the withering in motion vests the means of production in the state, on the assumption that a proletarian state is society organised, so that its property is everyone’s. The record’s central embarrassment lives in that assumption, since the vesting created, in every case, a distinct body of men administering everyone’s property and answerable, in practice, to themselves, and the distance between state property and social property turned out to be the distance in which Bakunin’s stratum grew. The second joint is the speed at which the chain is allowed to run. Engels’s prose moves from socialisation to the disappearance of class antagonism within a paragraph, treating as a consequence what history would treat as the work of centuries, if it happened at all; the compression is rhetorical, but rhetoric in a founding text becomes doctrine in the hands of successors, and Lenin’s chapters and the textbooks’ stages are essentially attempts to re-expand, honestly or evasively, the distance Engels’s paragraph had collapsed. The third joint is the treatment of the state’s interference in social relations, which Engels says becomes superfluous in one domain after another until the whole government of persons is replaced; the sequence within the sentence assumes that functions can be retired serially, each as its occasion lapses, whereas the record suggests functions retire, when they retire at all, into other functions of the same apparatus. None of this diminishes the passage as the tradition’s most consequential paragraph on the state. It locates, with unusual exactness, the assumptions on which the entire doctrine of the withering rests, and it shows that the assumptions were visible, to a careful reader, from the first edition. The failure, when it came, was not a surprise the text had hidden. It was a surprise the text had declined to price.

The German verb and the slow death it describes

Translation has softened the phrase in English, and recovering the German sharpens it. Engels’s verb is absterben, rendered in the standard English editions as to wither away or to die away. The verb is used of plants losing vitality, of limbs and organs ceasing to function, of customs and institutions lapsing as the life goes out of them. Its grammar matters as much as its imagery. Absterben is intransitive and agentless: things die away; nobody dies them away. There is no transitive counterpart in Engels’s sentence, no committee of liquidators for the state, no date of dissolution, no final session at which the apparatus votes itself out of existence. The linguistic choice performs the theory. If the state is an outgrowth of class relations, its end cannot be an administrative act performed upon it from outside, because any body capable of performing that act would itself be a state like organ, armed and separate, and the performance would reproduce the thing it ended. The end must instead be a loss of function registered across a whole society, the way a guild or a manorial court ended: not by decree but by the migration of real business elsewhere, until the shell stood empty and was swept into a museum.

Engels himself supplied the museum. In The Origin of the Family, Private Property and the State, published in 1884, he closed the historical arc by predicting that a society which reorganised production on the basis of free and equal association would consign the whole state machinery to the place it would then belong, a museum of antiquities, next to the spinning wheel and the bronze axe. The image is worth dwelling on because it specifies the manner of the ending better than any abstraction. Exhibits in a museum are not defeated enemies; they are tools whose world has gone. Nobody abolished the bronze axe by resolution. Iron arrived, and bronze became a curiosity. Engels is claiming the same modality for the state: not a conquest but an obsolescence, driven by the prior transformation of the conditions that made the tool necessary. The strength of the image is its clarity about mechanism. Its weakness, which the critics below exploit, is that obsolescence in the history of technology is driven by a superior replacement appearing, and the theory is obliged to say what replaces the state’s functions and why the replacement will not itself harden into a state.

What does the German word Absterben add to the phrase?

It adds the manner of the ending. Absterben describes plants, organs and customs dying out as the conditions sustaining them fail, and it has no agent: nothing is dismantled and nobody signs a dissolution. Engels chose a verb of obsolescence, not of demolition, which is why abolition language always mistranslates his claim.

Between Anti-Duhring and the museum image stands the short text that carried the phrase around the world. Socialism: Utopian and Scientific, assembled in 1880 from three chapters of the larger polemic at the request of the French socialist Paul Lafargue, reached readers who would never have opened a refutation of Duhring. In its pages the sequence is set out in the compressed form generations of party students memorised: the proletariat seizes state power, turns the means of production into state property as its first act, and by that very act abolishes itself as proletariat, abolishes class distinctions, and sets the state on the path of having nothing left to repress. The compression matters for the reception history. Read in Anti-Duhring, the withering sentence sits inside several hundred pages of argument about political economy and the history of socialism. Read in the extract, it sits a page from the end of a pamphlet, where it acquires the cadence of a peroration. Much of the phrase’s later career as a slogan, and much of the disappointment organised around it, is a consequence of that extraction: a conditional conclusion travelling without its conditions, at pamphlet length, in translation.

The German also disciplines a common English paraphrase. Renderings that speak of the state being dismantled, phased out or wound up smuggle an administrator back into the sentence, someone doing the dismantling, and with the administrator returns the whole problem Bakunin posed. Engels’s verb excludes that figure on grammatical grounds. If a future society contains an office charged with winding up the state, the office is the state under a temporary title, and the theory’s own logic requires asking when that office withers. The regress is not a trick; it is the point at which the doctrine’s internal consistency can be examined, and the examination is pursued in the section on the administration of things, where the question of who decides, and who compels, is put to the successor institutions directly.

The sequence: revolution, transitional state, then withering

The withering claim is the last term of a sequence, and most misunderstandings of it are sequence errors. The order the texts prescribe runs as follows, in prose rather than as a chart, because the transitions matter more than the labels. First comes revolution: the working class, organised as the ruling class for the purpose, seizes political power and breaks the political monopoly of the possessing classes. Second comes the transitional state, the period Marx named the dictatorship of the proletariat, in which the means of production pass into social ownership, distribution still follows contribution rather than need, and a coercive apparatus remains in being, now directed at suppressing the resistance of the old possessing classes and at defending the new order. Third, across a timescale the founders refused to fix, the class distinctions themselves dissolve: when everyone works and no one lives from ownership, there is no longer a class positioned to require repression, and production, socially organised, yields the abundance under which distribution by need becomes practicable. Fourth and finally, the state, deprived of its function, withers: the government of persons gives way to the administration of things, and political power properly so called disappears. Each stage is the condition of the next, and the withering is unintelligible, and was never offered, as a shortcut past the earlier stages.

The transitional stage is the hinge, and because it is a large subject with its own literature and its own controversies, it is treated in full in the series article on the dictatorship of the proletariat, to which readers wanting the institutional detail should go. Here the stage matters only for the discipline it imposes on the withering claim. The doctrine does not say that the state weakens steadily from the first day of the revolution, like a battery running down. On the orthodox account the transitional state is, in its coercive reach, potentially more extensive than the state it replaces, because it supervises the transformation of the whole economy and confronts the resistance that transformation provokes. Withering belongs to the far side of that exertion. Any presentation that has the state gently fading from the morning after the revolution has inverted the doctrine, and any criticism that scores points against that inversion has hit a target the texts do not defend. The texts defend something harder and, in the light of the record, more exposed: a state that first grows in function and only later, when classes are gone, lapses for want of employment.

Where does the sequence place the transitional state?

Between the revolution and the withering, doing the opposite of fading. The transitional state socialises production, suppresses the old possessing classes and administers distribution by contribution, and on the texts’ own account its functions expand before they can lapse. Withering is the last stage of the sequence, not the first.

Timescale deserves its own honest paragraph, since the founders’ refusal to name one has served every party to the dispute. Marx, asked in effect how long the transition would take, declined the question as utopian bookkeeping; Engels spoke of generations; Lenin, writing in 1917 with power still to be taken, insisted that no socialist had ever promised the higher phase by a date, and that the withering of the state even in its democratic form was a matter of a long and unspecified development. The vagueness is not incidental. A process defined by the disappearance of classes and the arrival of material abundance inherits the timescale of those enormous changes, and any founder who had named fifty years or five hundred would have been inventing. Yet the refusal has a cost that the assessment section must reckon with: a claim whose deadline recedes indefinitely is a claim that experience can never quite catch, and the difference between patience and evasion is precisely what the critics of the doctrine learned to press.

What the sequence excludes is as definite as what it includes. It excludes the anarchist shortcut, abolition at a stroke, already discussed. It also excludes the gradualist mirror image, in which the existing state, captured piecemeal by reform, quietly converts itself into the administration of things without a revolutionary break; Engels’s polemics against state socialist schemes of his day, and the whole logic of the state as a class instrument rather than a neutral tool, stand against that reading. And it excludes the identification of the transitional state with the destination. That identification, in which the apparatus built for the transition announces itself as the finished society and treats its own permanence as doctrine, is the specific deformation that the record section examines, because it is the one that actual ruling parties, confronted with the gap between the sequence and their situation, found most convenient.

The withering conditions table

Glossary precision pays its way when it produces something checkable, and the checkable form of this doctrine is a table. The withering conditions table lists, one row per condition, what the texts themselves require before the state can be expected to lapse, whether that condition has ever obtained, what happened instead in the societies that attempted the transition, and which critic predicted that outcome in advance. The rows are drawn from the texts rather than from later apologetics: the disappearance of classes, the social organisation of production, the replacement of the government of persons by the administration of things, the removal of any need for a special repressive force, and the material abundance under which the founders placed the higher phase of communist society. A reader who keeps the table in view can test any claimed instance of withering, and any defence of a non withering state, against the doctrine’s own stated requirements instead of against impressions.

Condition	What the texts require	Has it ever obtained?	What happened instead	Which critic predicted it
Classes and class antagonisms gone	No class left whose resistance requires organised repression	No	New strata formed around control of the state itself	Bakunin, from 1873
Production socially organised	Associated producers direct production in common, replacing private command of the economy	Only partially, under state direction	Direction concentrated in ministries and planning organs	Weber, on rational administration
Government of persons replaced by administration of things	Decisions become technical management of processes rather than rule over people	No	Political rule persisted inside every technical decision of scale	Weber; also Bakunin
No special repressive force required	Order maintained without a separate armed body standing over society	No	Police, armies and security services expanded in every attempted case	Bakunin
Material abundance	Scarcity sufficiently overcome that distribution by need replaces distribution by contribution	No	Shortage economies administered scarcity through the apparatus	The anarchist critics generally

Read the table from left to right and the doctrine looks demanding but coherent: if the first column’s conditions were jointly satisfied, the conclusion would at least be arguable, since an institution defined by its function could be expected to lapse with the function. Read it from the middle column outward and the difficulty stands bare. Not one condition has obtained in full anywhere, including in the societies that held power in the doctrine’s name for most of a century, and several rows show movement in the opposite direction from the one required. The rightmost column carries the sharpest embarrassment. The failures were not surprises that no one could have foreseen. The central one, the formation of a new ruling stratum around the transitional apparatus, was predicted in print, in detail, by the doctrine’s chief rival inside the workers’ movement, before the doctrine’s own canonical statement of withering had even been published in full.

The table also exposes the two honest ways of defending the doctrine, and forecloses a third, dishonest way. The first honest defence points at the middle column and argues that the experiment was never run: since the conditions never obtained, the conclusion was never tested, and a conditional claim is not refuted by the non arrival of its condition. The second honest defence concedes the record and revises the doctrine, treating the transitional state as a standing danger requiring institutional counterweights the founders did not design. The dishonest way, which the conditions table makes visible at a glance, is to redefine the conditions silently after the fact, so that whatever a ruling apparatus happens to be doing counts as a stage on the road, and the destination is redefined as the road itself. Twentieth century doctrine produced a specimen of exactly that manoeuvre in the theory of the state of the whole people, examined in the record section below, and the table is the instrument that catches it: a condition quietly edited is a row whose first column no longer matches the texts.

Marx’s transitional period: what Marx actually wrote

Because the phrase is Engels’s, recovering Marx’s own contribution means reading three short bodies of writing with care and resisting the temptation to make them say more than they do. The first is the letter to Joseph Weydemeyer of March 1852, in which Marx summarised his originality as showing that the existence of classes is bound up with particular historical phases of production, that the class struggle necessarily leads to the dictatorship of the proletariat, and that this dictatorship itself constitutes only the transition to the abolition of all classes and to a classless society. The sentence is the closest Marx came to a timetable of the state, and its logic should be noticed: the destination named is the abolition of classes, and the stateless society enters only as the presumed consequence of that destination, not as a separately argued terminus. Marx’s emphasis falls on the middle term, the transition, and on classes as the variable that decides everything else.

The second body of writing is the pair of texts on the Paris Commune of 1871, The Civil War in France and the preparatory drafts, where Marx praised the Commune for attempting not to transfer the state machine from one hand to another but to break it, replacing the standing army with the armed people, making officials elective and recallable at workmen’s wages, and fusing executive and legislative functions in a working body. The Commune writings are the nearest Marx came to institutional design, and their significance for the withering question is double. They show that his transitional state was meant to be built with anti bureaucratic devices from the first day, devices aimed precisely at preventing officials from consolidating into a caste. And they were written about a government that lasted seventy two days, which means the devices were never tested at the scale or duration where Bakunin’s prediction operates. The Commune is evidence of intention. It cannot be evidence of outcome, and presentations that treat it as a working model of a withering state are trading on its brevity.

The third and most important text is the Critique of the Gotha Programme, Marx’s marginal notes of 1875 on the draft programme of the united German workers’ party, circulated privately and published by Engels in 1891 against the party leadership’s wishes. Here the architecture of the whole doctrine stands in a few pages. Communist society is divided into a first phase, emerging from capitalism and stamped with its birthmarks, in which distribution follows labour contribution and a bourgeois right of unequal exchange persists, and a higher phase, in which the division of labour has been overcome, labour has become life’s prime want, and the springs of cooperative wealth flow abundantly enough for distribution by need. Between capitalist society and communist society, Marx wrote, lies the period of the revolutionary transformation of the one into the other, and of that period he said that the state can be nothing but the revolutionary dictatorship of the proletariat. Then comes the passage that matters most for precision. Marx asked what social functions would remain in communist society analogous to present state functions, and he refused to answer it in the programme, remarking that the question could be answered only scientifically. The founder of the doctrine, at the exact point where the withering should have been described, left a deliberate blank.

That blank is the most underreported fact in the entire literature of the phrase. Engels filled it with a confident mechanism; Lenin filled it with chapters; Marx left it as an open question addressed to a future science. Readers deciding how much weight the doctrine can bear should know which of the tradition’s authorities wrote the confident version and which declined to. The distinction also disciplines the attribution quarrel with which this article opened. To say the phrase is Engels’s is not to score a point off a careless textbook. It is to locate the doctrine’s specificity, its mechanism, its confidence, in the partner who specialised in system, and to note that Marx’s own texts supply the sequence and the conditions while withholding the destination’s description. Scholarship that honours that difference can argue about the doctrine’s worth with clean hands. Scholarship that merges the two men into a single oracle argues about a doctrine nobody quite held.

How Engels’s account of the state’s origin sets up its end

The withering claim is sometimes presented as an ornament bolted onto the theory late in Engels’s career, a consoling flourish for a movement facing long years of opposition. The publication record shows the opposite. The claim is the symmetrical completion of Engels’s account of the state’s origin, and it stands or falls with that account. In The Origin of the Family, Private Property and the State, Engels argued that the state is neither eternal nor natural nor a gift of reason. It arose at a definite stage of economic development, when society split into classes with irreconcilable interests and required a power, apparently standing above society, to moderate the conflict and keep it within the bounds of order. The state is, in his summary, a product of society at a certain stage, the admission that society has become entangled in insoluble contradiction with itself. An institution with a birthday can have a deathday. The whole force of the withering doctrine depends on that historical framing: what class division made, the end of class division can unmake.

The origin account does quiet work throughout the doctrine, and making it explicit shows which parts of the withering claim are load bearing. If the state arose because class antagonisms needed holding down, then its persistence is evidence about the persistence of antagonism, and its disappearance requires the antagonism’s disappearance and not merely a change of rulers, a new constitution or a proclamation of virtue. If, on the other hand, the state arose for other reasons, to coordinate irrigation, to organise defence, to administer complexity as such, then removing classes would leave those reasons standing, and the apparatus answering them would remain. This is why the Weberian objection, treated in its own section, is not a pedantic footnote but a rival origin story with a rival prediction: administration of a complex society requires a permanent professional apparatus whatever the class structure, so the apparatus will outlive every attempt to retire it. The debate between the withering doctrine and its strongest critics is, at bottom, a debate about why the state exists, conducted through predictions about whether it can end.

Engels’s origin story also explains a feature of the doctrine that puzzles readers who meet the phrase first: its patience. A state that took millennia to grow out of gentile society, acquiring its organs one by one as property and class hardened, is not the kind of growth that a single revolutionary season reverses. The Origin closes the arc begun in the kinship structures of early society, passes through the Athenian, Roman and Germanic formations, and arrives at the modern representative state as the most recent form of a very old institution. Against that backdrop, the museum sentence with which the book ends reads less like impatience than like a geologist’s confidence: the formation is impressive, and it is contingent, and contingent formations end. Whether the confidence was warranted is the business of the record section. The literary fact is that Engels built the ending into the foundations, and the doctrine cannot be trimmed of its destination without the origin story losing its point.

There is a companion relationship here with Engels’s other great posthumous reputation, the dialectics of nature project, and readers following the textual history will find the two enterprises share a method: take a domain that common sense treats as eternal, exhibit its historical emergence, and let the emergence argue the possibility of its transformation. The series treatment of the engels dialectics of nature guide examines that method in the natural sciences, including the places where it overreached. The state writings apply the same method to politics with, if anything, greater textual discipline, because the historical materials were richer and the author’s hand steadier. Noticing the shared method matters for the final assessment of the claim’s status: a thinker who everywhere converts eternal seeming things into historical processes will naturally state their endings as tendencies of those processes, and the withering doctrine may be best understood in that register, as the origin story read backwards, rather than as either a dated prophecy or a pious wish.

Engels’s last word: the valedictory introduction

Engels returned to the state question at the end of his life, in the introduction he wrote in March 1891 for the twentieth anniversary edition of Marx’s Civil War in France, and the text deserves a section of its own because it is the author’s own final gloss on the phrase, delivered with the authority of the man who had coined it and the melancholy of a strategist watching the German party grow respectable. Two passages carry the weight. In the first, Engels identified the disease he thought fatal to every workers’ movement that came near power: a superstitious reverence for the state, absorbed from childhood, which made even revolutionaries regard the apparatus as a sacred cow to be milked rather than a machine to be judged. The state, he insisted in that introduction, is at best an evil inherited by the proletariat after its victorious struggle for class supremacy, an evil whose worst sides the victorious proletariat, like the Commune, would be compelled to lop off as soon as possible, until a generation reared in new and free social conditions could throw the whole lumber of the state on the scrap heap. The image repays attention. The lumber of the state is not dissolved by chemistry or by time. A generation reared differently throws it away, and the throwing is an act, performed by people formed to be capable of it. Even Engels, at his most valedictory, quietly reintroduced the agent his famous verb had excluded.

What was Engels’s last word on the state?

The introduction of 1891 to Marx’s Civil War in France. Engels called the state an inherited evil, warned against superstitious reverence for it, praised the Commune’s devices for lopping off its worst sides, and pictured a generation reared in free conditions throwing the whole apparatus on the scrap heap.

The second passage is the one the anarchists of the next century quoted back at the Marxists with interest, because Engels used the introduction to restate the difference between the two camps in its sharpest form. The anarchists, he wrote, had taken to throwing the phrase dictatorship of the proletariat at the Marxists in horror; let them look at the Paris Commune, he answered, that was the dictatorship of the proletariat. The sentence binds the destination to the transition with a knot neither side has since untied. The Commune, seventy two days of besieged improvisation, is offered simultaneously as the model of the transitional state and as the down payment on the state’s disappearance, and its brevity is made to serve as proof of both characters. Everything problematic in the doctrine lives inside that double use. A form that lasted ten weeks under artillery fire is asked to demonstrate both that a workers’ state can wield power and that such a state inclines by nature toward its own dissolution, and the demonstration works only because the experiment ended before either proposition could be tested by duration. The 1891 introduction is thus the perfect emblem of the phrase’s whole career: the author’s own last statement of it is confident, quotable, and resting on the one case too short to prove anything, delivered to a party that was already settling into the reverence he had just diagnosed as the disease.

Lenin’s elaboration: the mechanism in State and Revolution

The phrase would have remained a respected sentence in a long polemic had Lenin not rebuilt an entire theory of the state around it in the late summer of 1917. State and Revolution, written in hiding in Finland between the July Days and the October seizure of power and left unfinished when events overtook its author, is the canonical elaboration, and its fifth chapter, on the economic basis of the withering away of the state, is where the mechanism receives its fullest statement. The readers guide to that text, state and revolution readers guide, walks the argument chapter by chapter; this section isolates what Lenin added to Engels, because the additions, as much as the original, are what the record then tested.

Lenin’s first addition was emphasis, and it was polemical. Writing against the social democratic orthodoxy of Karl Kautsky, which he accused of keeping the revolutionary vocabulary while quietly dropping the destination, Lenin insisted that the withering of the state was not a decorative clause but the dividing line between Marxism and opportunism. A Marxist, on his account, was someone who extended the recognition of the class struggle to the recognition of the dictatorship of the proletariat and beyond it to the disappearance of the state altogether. The test was deliberately set at the doctrine’s most exposed point. Where Engels had stated the destination as the conclusion of a historical argument, Lenin made it a badge of fidelity, and that shift in register, from conclusion to badge, shaped how the claim functioned inside the movement he led: not as a hypothesis to be checked against conditions but as a loyalty line, deviation from which marked the heretic. The subsequent history of the doctrine, in which questioning the apparatus’s growth could be treated as disloyalty to the destination, is partly written in that shift.

His second addition was the mechanism, drawn from the Gotha Critique’s two phases and stated with an economist’s care. In the first phase of communist society, the state that persists is, in Lenin’s notorious formulation, a bourgeois state without the bourgeoisie: distribution still follows contribution, equal right remains the right of unequal men to unequal shares, and the enforcement of that accounting requires an apparatus, though one no longer serving a possessing class. Only when the higher phase arrives, when labour has become habitual and its productivity has made scarcity recede, does the need for any enforcement of distribution lapse, and with it the state. Lenin was explicit that people would become accustomed to observing the elementary rules of social life without force or subordination, and that the habit, not a decision, is the solvent of the apparatus. The mechanism is sociological rather than constitutional: no article abolishes the state; a population that no longer needs compelling stops generating the demand for compellers.

His third addition is the one least quoted by the doctrine’s defenders and most useful to its critics. Lenin extended the withering to democracy itself. Democracy, as a form of state, is the organised systematic use of force by one part of the population against another, and it too belongs to the transitional inventory: with the disappearance of classes, democracy in this strict sense also withers away, and humanity passes to administration without subordination. The passage matters because it removes a common evasion. It is sometimes suggested, by way of rescue, that the doctrine promised only the end of some narrow repressive ministry while free institutions flourished on. Lenin’s text forecloses the comfort. The destination is the end of the state as such, the state his movement was busy constructing included, and the promise was printed, in millions of copies, by the very party that then built the largest state of the century. The gap between the text and the record is not a subtle scholarly construction. It is the plainest fact in the whole dossier, and the next sections keep it in view.

The timetable quarrel inside the Second International

Between Engels’s death in 1895 and Lenin’s pamphlet of 1917, the withering doctrine lived the quiet life of an inherited clause, recited in party programmes and rarely interrogated, and the manner of its keeping explains much about its condition when the record began. The German Social Democratic Party, the movement’s flagship, carried the destination in its Erfurt Programme of 1891, a document drafted under Engels’s watchful correspondence, and its theoreticians treated the end of the state the way established churches treat the millennium: as a truth whose remoteness excused the institution from organising its affairs around it. Karl Kautsky, the programme’s commentator and the era’s pope of Marxism, wrote voluminously on the state the workers’ movement would build and very little on the mechanics by which that state would oblige the doctrine by disappearing. The clause survived by not being examined. Nobody attacked it, because nobody in the mainstream intended to rely on it, and nobody defended it in detail, because its defence would have required specifying conditions whose absence in the present was the party’s standing excuse for gradualism.

The one serious internal challenge came from Eduard Bernstein, whose revisionist writings of the late 1890s proposed to strike the destination altogether and keep only the movement: the final goal, he announced, was nothing, the movement everything. Bernstein’s target was the catastrophic side of the doctrine, the expectation of collapse and seizure, more than the withering clause itself, but his logic touched it directly. If socialism arrives as the cumulative product of reforms inside the existing state, then the state is not an instrument to be wielded and retired but a terrain to be occupied and improved indefinitely, and the withering destination becomes a fairy tale the movement tells itself about a journey it has no intention of completing. The party condemned revisionism formally and absorbed it practically, which is to say it kept the clause in the programme and the gradualism in its conduct, and the clause’s function settled into what it remained for most of the century: a horizon that legitimised the organisation’s patience. Horizons do not wither, and organisations do not dissolve into them. They point, and pointing is enough.

Lenin’s violence against this settlement in 1917 is unintelligible unless the settlement is seen clearly. When he accused Kautsky of having distorted Marxism into opportunism, the specific charge was that the orthodoxy had retained the transitional state, which legitimised building party and union apparatuses, and dropped the withering, which alone made those apparatuses provisional. A doctrine whose destination clause is treated as decoration produces, Lenin saw, a movement whose every institution plans for permanence while speaking the language of transition. The accusation boomeranged within a decade, when the state Lenin founded began performing the same operation on his own text, keeping the chapters on smashing the old machine in ceremonial circulation and letting the fifth chapter’s destination fade from the catechism. The pattern generalises beyond any party. The withering clause is the one element of the doctrine that indicts the doctrine’s own custodians, and custodians of every generation, orthodox, revisionist and revolutionary alike, have responded with the same editorial instinct: revere the phrase, and change the subject. That instinct, repeated across a century by men who agreed on nothing else, is itself evidence in the assessment, because it suggests the clause was experienced from inside, by its friends, as the doctrine’s most dangerous sentence.

There is one more figure in the timetable quarrel who deserves rescue from footnote status, because she posed the question neither the orthodox nor the revisionists would pose. Rosa Luxemburg, criticising the Bolsheviks from within the revolutionary camp in 1918, warned that freedom restricted to the supporters of the government and members of one party, however numerous, is no freedom at all, and that without unrestricted democracy the life of public institutions dies out and the bureaucracy alone remains the active element. The warning is Bakunin’s mechanism restated by a Marxist of unimpeachable revolutionary credentials, with the causal arrow pointed at political monopoly rather than at state power as such. Luxemburg’s version matters for this article because it identifies the variable the founders’ texts left unspecified: not the existence of a transitional state, but the presence or absence inside it of independent public life capable of checking the apparatus. Her prediction, like Bakunin’s, was borne out by the case she was addressing, and the fact that the tradition produced the warning twice, from opposite wings, half a century apart, and filed it both times as an embarrassment, belongs in any honest dossier on the phrase.

The administration of things and what it still requires

Every account of the withering sooner or later leans on the formula Engels took over from the French socialist tradition and transmitted to Lenin and the textbooks: the government of persons is replaced by the administration of things and the direction of the processes of production. The formula descends from Saint-Simon and his school, for whom the passage from governing men to administering things expressed the coming rule of producers and engineers over the politics of priests and warriors, and Engels adopted it with acknowledgment, observing that Saint-Simon had already in 1816 declared politics to be the science of production. Few borrowed phrases have carried so much weight. It is the doctrine’s only sketch of the destination’s daily texture, the single sentence asked to describe how a complex society runs its railways, hospitals, power grids and food supply without anyone ruling anyone.

Where does the Saint-Simon formula enter the argument?

At the destination. Engels adopted the older French formula to describe what replaces political rule once classes are gone: the administration of things and the direction of production processes, in place of the government of persons. It is a sketch of the end state’s texture, offered by the tradition as the answer to what remains when rule lapses.

Subject the formula to the pressure it invites and three questions detach themselves, each simple to state and none answered in the founding texts. The first is the question of decision. Administration settles how a chosen end is achieved; it does not choose ends. A society must still decide how much to invest against how much to consume, which risks to run, which regions to favour, which research to fund, and these decisions distribute advantages unevenly among real people with real disagreements. Deciding among contested ends, under scarcity of time and knowledge even amid material plenty, is the activity the rest of political theory calls government. Renaming the room in which it happens does not change the activity. The second question is enforcement. Rules of distribution and safety bind only if defectors can be checked, and the higher phase’s own description concedes the point by relying on habit: people, Lenin wrote, would become accustomed to the rules without compulsion. Habit is a real social force, and societies do run largely on it, but every known society maintains a coercive margin for the minority whom habit does not govern, and the doctrine must either extend that margin to its destination, readmitting a repressive apparatus under a politer name, or claim that the defecting minority will disappear with its causes, which is a psychological prediction the texts assert rather than argue.

The third question is knowledge, and it is the one the formula most obviously evades. Administering things at the scale of a modern economy means aggregating information about capacities, needs and consequences that is dispersed among millions of people and much of it tacit. The classical liberal objection, stated most sharply in the socialist calculation debate of the 1920s and 1930s, held that without market prices the administrators of things would lack the signals by which to direct production rationally. Whatever verdict a reader reaches on that debate, and this article takes none, its bearing on the formula is direct: the administration of things turns out to be not a quieter, simpler residue left when politics departs, but the hardest problem in the whole system, and the people solving it would wield discretion on a scale no ministry of the old type possessed. Discretion at scale, exercised by a identifiable body over others who must comply, is what the doctrine’s own definition calls a state function. The formula thus risks describing the destination in words that its own theory of the origin refuses to accept: if an apparatus directs everyone’s production and settles everyone’s disputes, the origin theory asks what class antagonism sustains it, and if the answer is none, the theory owes an account of why this apparatus, alone among historical apparatuses, will administer without ruling.

Against these pressures, the strongest available reply within the tradition deserves its hearing. It runs that the questions smuggle the premises of class society into the destination. Decision among ends is politics as known, the reply concedes, but politics as known is deformed by the fact that the ends themselves are class ends, the investment question being at bottom a question of who appropriates the surplus. Remove the appropriation and the residue of genuine disagreement is technical and temperamental, settleable by rotation, lot, expertise and revocable delegation, the Commune devices generalised. Enforcement shrinks with the incentives to defect, which are overwhelmingly incentives created by unequal access to goods. Knowledge problems belong to the first phase and its accounting, not to a society of abundance where planning can proceed in physical terms against known needs. The reply is coherent as far as it goes. Its cost should be registered: it defends the formula by leaning, at every joint, on the abundance condition and on a transformed human motivation, the two least secured planks in the entire structure, and it thereby confirms the assessment reached later in this article, that the destination stands or falls with premises about human beings under plenty that the nineteenth century asserted and the twentieth did not supply.

The record: states that grew instead

The empirical record can be stated without rhetoric, and it is stronger stated so. No state has withered away. No state constructed under Marxist auspices has withered away. In each of the major cases, the apparatus of coercion and administration grew across the whole life of the regime, and where those regimes ended, they ended by collapse, negotiated surrender or transformation from above, never by the gradual lapse of functions into a self managing society. The Soviet Union, the founding case, built upon the Tsarist inheritance a party state whose security organs, from the Cheka through its successor agencies, employed coercion on a scale the old regime never approached, whose ministries directed an entire industrial economy, and whose final decades were marked not by functions draining away but by their proliferation into every corner of social life. The regime’s own doctrinal custodians registered the difficulty in the only way open to them: at the Twenty Second Congress of the Communist Party in 1961, the programme adopted under Nikita Khrushchev announced that the dictatorship of the proletariat had been succeeded by the state of the whole people, a formula that preserved the vocabulary of stages while quietly detaching the stage theory from any observable withering, since the new stage was defined to last until the full construction of communism at an ever receding horizon.

The internal dissent tells the same story from inside the doctrine’s own categories. Leon Trotsky, exiled and writing The Revolution Betrayed in 1936, argued that the Soviet bureaucracy had consolidated into a privileged caste resting on the state it administered, and he treated that caste’s grip as the central fact of the regime. His explanation differed from Bakunin’s, locating the cause in scarcity and isolation rather than in the nature of state power as such, and the difference matters to historians of Marxism. For the withering question, though, the two hostile witnesses agree on the datum: the transitional apparatus, far from lapsing as its tasks were completed, acquired interests of its own and defended them. Trotsky’s scarcity explanation is in fact the conditions defence in embryo, offered by a revolutionary with every motive to believe the destination reachable, and even on his account the apparatus grew for as long as he could observe it. The full Soviet dossier, its economic debates, its cultural machinery and its long stagnation, belongs to the soviet marxism complete guide; this article needs only the trajectory, and the trajectory is one directional.

China supplies the second great case, and it repeats the pattern with variations that strengthen rather than soften the conclusion. The Chinese party state after 1949 constructed administrative penetration of village and workplace beyond anything the imperial or republican orders had achieved, and the decades of market reform that followed 1978 changed the economy’s operating system without loosening the apparatus: the party’s committees, the security services and the administrative hierarchy adapted to governing a partly market economy and grew in technical capacity as they did so. If withering were a tendency generated by socialised production, the Chinese case, with its vast state sector and its decades of proclaimed socialist construction, offered the conditions under which some function somewhere might have begun to lapse into social self management. The observable tendency ran the other way, toward a more capable and more comprehensively informed state. Nor does the smaller print of the century help the doctrine. Yugoslavia’s self management reforms after 1950 genuinely devolved decisions to workers’ councils and are the nearest historical approach to an official policy of de etatisation from within a ruling communist party; yet the League of Communists retained its monopoly, the federation’s coercive organs remained intact, and when the state finally disappeared it did so by violent disintegration, the opposite modality from Engels’s verb. Cuba, Vietnam and the states of Eastern Europe each add their own variant of the same trajectory: apparatus first, permanence after, and an end, where an end came, by rupture rather than by lapse.

Set against this record, the scattered instances sometimes offered in the doctrine’s support dissolve on inspection. Functions do migrate out of states in ordinary politics: industries are denationalised, services contracted, currencies handed to independent banks. But those migrations occur inside states that remain armed, taxing and supreme over the territory, and they transfer functions to markets and agencies, not to a classless society’s self administration. The withering claim is not the claim that states rearrange their portfolios. It is the claim that the state as such, the separate apparatus of rule, lapses when classes lapse. Nothing in the record shows a step down that road, and a great deal shows traffic in the other direction. The failure, moreover, cannot be quarantined as a failure of this or that leadership’s sincerity, because it is uniform across leaders of opposite temperaments, continents and decades, from revolutionaries of undoubted commitment to apparatchiks of none. Uniformity of outcome across varied agents is the signature social scientists look for when they suspect a structural cause, and a structural cause is exactly what the doctrine’s two great critics claimed to identify: one in the sociology of organisation, one in the logic of power itself, and it is time to give those critics the floor at the length they have earned.

The social democratic detour: the state that grew by consent

The revolutionary road was not the only one down which the doctrine’s heirs travelled, and the other road supplies a control case of a kind historians rarely get. Across the twentieth century, social democratic parties in Scandinavia and elsewhere took office by election, held it for generations, socialised the risks of sickness, age and unemployment, compressed wage differentials, and built the most extensive public provisions in the capitalist world, all without a dictatorship of the proletariat, without abolishing classes by decree, and in declared allegiance, for much of the period, to the same destination the revolutionaries claimed. If the withering of the state were a tendency generated by the socialisation of provision as such, by the removal of life’s necessities from the market’s discipline, then the Scandinavian states should show it earliest and most clearly, since they decommodified more of ordinary life, for longer, than any revolutionary regime managed to do. What they show instead is the most instructive split in the whole dossier. Their coercive cores are, by the standards of the century, remarkably slender: small prison populations, lightly armed police, high trust between citizen and official. Their administrative wholes are enormous: the state touches income, housing, health, education, childcare and old age at a density no nineteenth century liberal would have believed compatible with freedom. Provision grew; rule, in the founders’ narrow sense of a repressive apparatus standing over a subject class, arguably thinned; and the sum of the two movements is not the withering of the state but its transformation into a service bureaucracy of exactly the kind Weber described, legitimated by election rather than by doctrine.

The detour matters for the assessment because it isolates a variable the revolutionary cases confound. In Russia and China the apparatus’s growth can be blamed, and by defenders always is blamed, on backwardness, invasion, civil war and the siege mentality they bred. In Scandinavia none of those excuses is available. Here were advanced, peaceful, literate societies, governed for decades by the doctrine’s own parliamentary wing, moving steadily in the direction of socialised provision under conditions of abundance the founders would have recognised as their premise arriving. The apparatus grew anyway, not as a conspiracy and not as a betrayal, but as the arithmetic of the promises: every service universalised is a service administered, every entitlement a file, every file a staff. The Scandinavian finance minister does not need Bakunin to explain why his ministry never shrinks. The lesson the detour teaches is the Weberian one with the menace removed, and it is in a way the hardest lesson for the doctrine, because it cannot be answered by pointing at terror or encirclement. Administration grows because complex societies ask it to, through ballot boxes as readily as through party congresses, and a doctrine that predicts the opposite movement owes an account of the asking, of why a population freed from class rule would stop wanting the coordination whose supply is the state’s growth. No text in the canon supplies that account. The founders assumed that the demand for the apparatus came from above, from the needs of a ruling class. The detour suggests the demand also comes from below, from the rational preferences of ordinary people for insured lives, and that a withering which must override those preferences would require precisely the coercion it exists to abolish.

Kropotkin’s alternative: withering by federation

Bakunin supplied the anarchist camp’s prediction; Peter Kropotkin supplied its positive theory, and the theory deserves examination here because it is, in structure, a rival doctrine of the state’s end rather than a mere demand for its abolition. In The Conquest of Bread and in the long argument of Mutual Aid, Kropotkin proposed that the state is a recent and parasitic growth upon an older and hardier stock of institutions, the commune, the guild, the federation of free associations, which had handled coordination, welfare and defence for centuries before the modern state absorbed their functions. On this account the end of the state is not an event but a regrowth: as free associations resume, one by one, the functions the state monopolised, the apparatus is hollowed from the outside, and what remains is a federation of communes and workplaces bound by agreement rather than by subordination. Readers of Engels will hear the echo before it is pointed out. Kropotkin’s mechanism is absterben with the agency restored and the direction reversed: the state dies away, not because its class function lapses after a transitional dictatorship, but because rival institutions outcompete it in the present, function by function, until the shell stands empty. The two doctrines agree that the state is historical, that it can end, and that its end will look like obsolescence rather than like a bonfire. They disagree about everything that decides the matter: the motor, the sequence and the role of the transitional power that one side builds and the other refuses.

The disagreement is worth anatomising because it shows the withering problem from the one angle the Marxist texts never occupy, the angle of construction before power. Kropotkin’s route avoids Bakunin’s dilemma about the transitional state by declining to build one, and avoids the Weberian dilemma by proposing that the successor institutions grow up inside the old society, trained and tested, rather than being improvised after a seizure. Its difficulties are of the opposite kind, and they are severe. The federal institutions of the medieval commune and the guild, on which the theory’s history rests, operated at small scale, inside economies of limited interdependence, and were themselves, as Kropotkin’s own chapters concede, forever hardening into oligarchies that the state then absorbed partly by invitation. The scaling problem that defeated the Commune devices returns in federal dress: a federation of communes that must coordinate a railway network, a power grid or a war acquires delegates, committees and standing staffs, and the argument about whether those organs are a state under federation’s name is the argument of 1872 restated with better examples. The Spanish experience, treated in the next section, is the one full trial the federal route received, and its trajectory, from free militias and collectives to ministers in a republican cabinet within a single year, suggests that the pressures of scale and war act on federations as they act on parties. What survives the trial, on the anarchist side as on the Marxist, is the diagnosis rather than the destination: the state is not eternal, its functions can migrate, and every attempt to retire it by either road has instead refurbished it. The two traditions, expelled from each other’s internationals and from each other’s histories, converge in the end on the same unintended monument, an apparatus that outlasted both of their theories of its death.

The Spanish mirror: the anarchists’ own experiment

Fairness to Bakunin requires one more exhibit, and it is the exhibit his own tradition least likes to discuss. In July 1936, in response to a military rising, the anarchist movement of Spain, the strongest in the world and the heir of Bakunin’s federalism through the CNT and the FAI, found itself holding effective power across much of Catalonia and Aragon. Here, if anywhere, was the abolitionist experiment the nineteenth century had argued about: a movement doctrinally committed to dissolving the state on the first day, armed, organised and in possession of factories, collectives and militias. What the movement did with its power is the mirror image of the Marxist record, and the mirror is worth describing without pleasure, because it completes the dossier’s logic. The collectives were real and, in places, remarkable achievements of self organisation. But the war against a professional army demanded coordination, discipline and central direction; the militias were militarised; a central committee of antifascist militias functioned as a government in all but name; and within four months the CNT itself entered first the Catalan government and then, in November 1936, the government of the Spanish Republic, supplying four ministers to a state it existed to abolish. The anarchists in power discovered, in compressed and tragic form, the functional pressures the Marxists had predicted for them in the 1870s: coordination at scale secretes directing organs, and directing organs accrete the permanence of states.

Did the anarchists abolish the state where they held power?

No. In Spain in 1936 the anarchist movement, holding real power in Catalonia, built coordinating bodies that functioned as governments, militarised its militias under the pressure of war, and within months entered first the regional and then the national government it existed to abolish.

The Spanish mirror does not rescue the withering doctrine, and it is important to be exact about why, since both camps have tried to spend it as rescue. What Spain shows is that the pressures toward apparatus are general: they acted on the abolitionists as reliably as on the gradualists and the revolutionaries, which confirms the Weberian and Bakuninist diagnosis that the mechanism lives in scale, war and coordination rather than in any party’s doctrine. But that confirmation damages the withering doctrine rather than defending it, because the doctrine’s claim was never merely that other people’s states would persist. It was that its own road led through the apparatus to the apparatus’s end. A world in which every road, including the road marked abolition, leads to a state that stays, is a world in which the withering destination has been detached from every available means of reaching it. Spain also refines the assessment of Bakunin, and a glossary article should record the refinement. His prediction about the Marxists was vindicated. His prescription was tested once, under the most favourable conditions his movement ever enjoyed, and it collapsed into statehood faster than the Marxists’ prescription did. The nineteenth century’s great argument thus ends with both sides’ mechanisms in ruins and only the diagnosis standing: Bakunin’s of the transitional state, Engels’s of the abolitionist shortcut. The two diagnoses are compatible, and together they describe the whole of the subsequent record with an accuracy neither man achieved alone. That is perhaps the deepest result this article has to offer, and it is not a comfortable one for any tradition.

Bakunin’s warning: the counter-prediction from inside the movement

Mikhail Bakunin made his prediction before the doctrine he was predicting against had finished being written, and the timing is part of its sting. In Statism and Anarchy, published in 1873 from within the wreckage of the First International, he asked what it could mean for the proletariat as a whole to be raised to the position of ruling class. If the proletariat rules, he pressed, over whom does it rule? There would remain another stratum, the peasantry above all, standing under the workers’ state as its subject population, and rule over subjects requires rulers: a governing minority drawn from the workers themselves, who would cease in the act of governing to be workers and would come to regard the labouring world from the altitude of the state. The former workers, he wrote, once they govern, look down on the world of labour from the heights of power, represent no longer the people but themselves and their pretensions, and whoever doubts it knows nothing of human nature. The prediction is specific enough to be checked: not merely that revolutions disappoint, but that the transitional state itself manufactures a new ruling stratum out of its own personnel, whose interest is the state’s permanence.

Bakunin added a second string that is often forgotten beside the famous one. He warned that a government of the most learned, the scientific socialists as he called the Marxian school, would be the most oppressive of governments precisely because of its confidence: rulers who believe their rule is the application of a science of history will treat dissent not as disagreement but as error, and error as a symptom to be corrected in the erring. The prediction generalises beyond class. It says that the danger of the transitional state lies not only in the material interests of its officials but in the doctrine the officials carry, which certifies their rule as a stage of a necessary process and thereby disarms the question of when the stage should end. Readers of the record section will recognise the mechanism in the state of the whole people formula of 1961, where doctrine obligingly reclassified the apparatus’s permanence as progress. Bakunin did not foresee the formula. He foresaw the function it would perform, because he had identified, before the fact, the incentive structure that produces such formulas.

Justice to the record also requires stating where Bakunin’s own position was weakest, since this article assesses arguments rather than canonising critics. His alternative, the immediate abolition of the state and its replacement by the free federation of communes, was never given a mechanism for the questions the Marxists pressed on him: how a federation defends itself against a surviving hostile state, how it coordinates production across regions without a directing centre, how it prevents the federation’s own delegates and militias from consolidating into the thing abolished. The Marxist counter charge, that Bakunin’s abolition would reproduce the state informally and therefore worse, unaccountable because unacknowledged, was a serious argument and remains one wherever informal power flourishes under anti authoritarian banners. Bakunin also wrote as a conspirator who believed in secret directing organisations of the convinced, a practice in obvious tension with his theory of spontaneous freedom, and his polemics against Marx were disfigured by personal and ethnic venom that no modern reader defends. None of that dulls the prediction’s edge. A forecast is not refuted by its author’s other failures. The question the doctrine must answer is why the forecast, made in 1873 by a rival who lost every organisational battle of his lifetime, describes the trajectory of the victorious tradition’s states more accurately than the tradition’s own destination describes them.

The fairest framing, and the one this article adopts for the final assessment, is that Bakunin stands to the withering doctrine as a control experiment designed in advance. The founders asserted that a transitional state, holding power on behalf of the immense majority and charged with abolishing classes, would differ in kind from all previous states and would lapse when its work was done. Bakunin asserted that it would differ only in personnel, that the work would never be declared done by those whose position depended on its continuance, and that the doctrine of stages would supply the vocabulary in which permanence was explained as progress. A prediction contest does not get cleaner than that: rival mechanisms, specified beforehand, incompatible outcomes. The outcome is in, and it is not ambiguous. Whatever else the twentieth century tested, it tested this, and the point of maximum embarrassment for the withering doctrine is not that its destination failed to arrive but that the rival’s mechanism, offered in the doctrine’s own century by the man the doctrine’s founders expelled, names the process by which it failed.

The Weberian objection: bureaucracy as a permanent fixture

Bakunin attacked the transitional state from inside the revolutionary camp; Max Weber attacked the destination from inside the sociology the founders’ century was building, and his objection is in one respect the colder of the two. Bakunin’s prediction allows, in principle, that a transitional state staffed by saints and watched by vigilant communes might escape his mechanism. Weber’s argument, developed across Economy and Society and the political writings of the war years, denies the destination to any complex society whatever its class structure or the virtue of its rulers. Bureaucracy, on his account, is not a excrescence of capitalism or of class rule. It is the technically superior form of administration as such: precise, continuous, calculable, staffed by trained officials working from written files under fixed rules, and it advances wherever tasks grow complex, in the state, in the army, in the church, in the capitalist firm itself. A socialism that socialised the means of production would not dissolve this apparatus. It would complete it, by uniting the bureaucracy of the state and the bureaucracy of industry into a single hierarchy, and Weber said in terms that the dictatorship of the official, not of the worker, was on the march.

The objection strikes the withering doctrine at its foundation rather than at its timetable. Recall that the doctrine’s confidence rests on an origin theory: the state exists because classes exist, so ending classes ends the state. Weber’s sociology substitutes a different engine. The apparatus exists because coordination at scale requires specialised, continuous, rule bound administration, and that requirement is generated by the complexity of the tasks, not by the antagonism of classes. Abolish classes tomorrow and the railways still need dispatching, the hospitals rotas, the power grid balancing, the water system maintaining, and each of those activities, Weber’s analysis implies, will secrete its trained staff, its files, its hierarchy of competence, its corporate interest in continuity. What would wither, on this account, is only the overtly repressive superstructure, the political police and the class courts, and even those would be under pressure to return in administrative dress, as inspectorates and licensing bodies, the moment coordination failures demanded enforcement. The administration of things, in short, is not the residue left when the state departs. It is the state’s most durable form, the part that every modern order, capitalist or socialist, has in fact expanded without interruption.

Weber also supplied the micro mechanism that Bakunin’s rhetoric lacked, and it deserves stating because it converts the objection from a mood into a theory. Officials, he argued, possess two structural advantages over any nominal master: expertise and secrecy. The trained official knows the files, the precedents and the practical possibilities in a way no rotating amateur, no workers’ council meeting after shifts, no minister without a staff can match, and the official can defend that advantage by treating knowledge of the apparatus as official secret. Political control of bureaucracy is therefore always partly fictitious, and every attempt to control it requires building a second apparatus to watch the first, which inherits the same advantages. Applied to the transitional state, the mechanism predicts exactly the Soviet trajectory without needing any hypothesis about the bad faith of revolutionaries: the soviets and councils, staffed by amateurs and exhausted volunteers, would within a few years be taking their papers from the permanent officials, and the permanent officials would become, in substance, the government. Historians of the early Soviet state have documented that migration of real initiative from elected bodies to appointed staffs in granular detail, and it began while Lenin still lived and while the destination was still being preached.

The tradition’s replies to Weber exist and should be weighed rather than waved at. One reply accepts the description of bureaucracy and disputes only its eternity: Weber, it is said, mistook the administrative forms of a scarcity society, where officials guard access to insufficient goods and derive their power from that gatekeeping, for the administrative forms of any society whatever. Under abundance the gatekeeping function dies, and what remains is technical coordination of a kind engineers already perform inside firms without generating a ruling caste. A second reply distinguishes hierarchy of competence from hierarchy of command: that some people know more about the grid than others does not entail that they rule the others, any more than the surgeon’s knowledge makes the hospital a dictatorship. A third reply turns the origin question back on the critic: Weber’s own examples of bureaucracy’s advance are drawn from states and firms inside class societies, so his induction to all possible societies outruns his evidence in exactly the way he accused the Marxists of outrunning theirs. Each reply has force, and the third in particular is a fair methodological hit. But the replies share a burden they cannot discharge: they must explain why, in the one set of cases where the experiment was actually run, the apparatus behaved as Weber’s induction said it would, growing through every phase, under leaders who had read their Engels and intended the opposite. An induction confirmed by the only available trials, against the intentions of the triallists, is not a conjecture to be balanced politely against a hope. It is the standing state of the evidence.

Scarcity: the material premise under the microscope

Every condition in the table leans, at the last joint, on the abundance premise, and the premise deserves the microscope before the defence built on it is priced. The founders’ version is more careful than its caricature. Engels and Marx did not predict infinite wealth; they predicted that the productive forces unleashed by capitalism, once rationally organised, would suffice to reduce necessary labour drastically and to make the struggle for necessities, the arena in which classes form, a marginal rather than central fact of life. Distribution by need in the Gotha Critique is tied to that reduction: the higher phase begins when the springs of cooperative wealth flow more abundantly, a comparative and historical threshold, not a cornucopia. Even in that careful version, the premise carries three difficulties that a century of economic reflection has sharpened rather than blunted. The first is that needs expand with the means of satisfying them, a process the founders themselves documented for the working class of their own century, whose necessary consumption rose with civilisation; a threshold defined against expanding needs recedes as it is approached, and the higher phase risks the same infinite postponement as the geopolitical condition examined above. The second difficulty is positional goods: a class of goods, status, location, precedence, the best view and the rarest skill, whose value consists in others not having them, and which no growth of production can make abundant because their scarcity is their substance. A society may abolish scarcity of bread and find its distributive conflicts relocated, intact, to the goods that cannot be multiplied, and conflicts over distribution are, on the doctrine’s own theory, the material from which apparatuses grow. The third difficulty is time. Even under material plenty, human time remains finite, and the allocation of attention, care and precedence among persons generates queues that someone must administer and disputes that someone must settle.

Can abundance abolish scarcity altogether?

Not on the evidence of the concept itself. Abundance in the founders’ careful version means necessities secured and necessary labour shrunk, but needs expand with means, positional goods are scarce by definition, and time stays finite, so distributive conflict survives the threshold the doctrine sets for the higher phase.

None of this makes the founders foolish; it makes their premise a nineteenth century extrapolation, drawn when the margin between production and subsistence was the overwhelming fact of social life and when it was reasonable to treat scarcity as one problem with one root. The assessment that matters for this article is narrower. The withering doctrine requires not merely prosperity but the specific social consequence of prosperity, the drying up of the conflicts that employ an apparatus. A century in which the richest societies in history operate the most extensive administrative and coercive apparatuses ever built, apparatuses substantially occupied with the distributive and positional conflicts of affluent populations, is a century that has run the prosperity variable upward and watched the apparatus follow it. That observation is not a refutation of the conditional, whose antecedent includes classlessness and not mere wealth. It is, however, a direct hit on the doctrine’s background picture of how the antecedent was supposed to be approached, through the development of the productive forces under socialised direction, since the variable that was supposed to starve the state has instead, in every observable case, fed it. A glossary article that failed to register that reversal would be leaving its readers with a map whose contours all slope the wrong way.

The Commune devices under the microscope

Because the safeguards are the doctrine’s only engineering, they deserve the same itemised scrutiny the conditions received, and the scrutiny is more respectful than either dismissal or incantation. Four devices make up the kit, all drawn from the Paris Commune’s practice as Marx reported it: universal election of officials, instant recall of the elected by their electors, rotation in office, and the workman’s wage as the ceiling of official pay. Each device targets a specific gear in Bakunin’s mechanism. Election targets appointment from above, the channel by which an apparatus reproduces itself in its own image. Recall targets the interval problem, the years during which a representative, once mandated, is free to become a ruler. Rotation targets the formation of a caste, the slow conversion of office into identity and livelihood. The wage ceiling targets the material root, the official’s separation from the condition of the governed, by making office financially indistinguishable from the bench and the lathe. As a design brief, written in 1871 against a prediction published in 1873, the kit is remarkable: it is the prediction’s own checklist, answered point for point before the prediction’s most famous statement had appeared.

Subject each device to scale and duration, the two variables the Commune never supplied, and the difficulties accumulate in an instructive order. Election scales well; modern states elect millions of officeholders without strain, and no one supposes the result has dissolved their apparatuses, because election decides who staffs the offices and leaves the offices, their powers, their files and their permanence untouched. Recall scales badly into administration: a planning official whose every consequential decision can be reversed by an assembly that meets after working hours will either make no consequential decisions or make them where the assembly cannot see, and the second outcome, decision migrating to the unrecallable, is the Weberian mechanism entering through the safeguard itself. Rotation collides with expertise in exactly the posts where the apparatus is most consequential, the grid, the epidemic service, the monetary authority, where the knowledge is tacit, accumulated and dangerous to interrupt; a society may rotate its magistrates and prosper, but it rotates its epidemiologists at a price counted in lives, and the posts exempted from rotation, exempted for excellent reasons, become the nucleus of the permanent state the rotation existed to prevent. The wage ceiling collides with the labour market for scarce skill: cap the surgeon’s administrative salary at the bench wage and the ablest administrators exit to wherever their skill is priced, leaving the apparatus to the less able and, worse, to the corruptible, for whom the ceiling is no deterrent since their income was never going to be the salary. None of these difficulties is a refutation in isolation; institutions can be patched, exemptions sunsetted, salaries supplemented in kind. But the patches all lean the same way, toward permanence, expertise and exception, which is to say toward the apparatus, and the kit’s history is the history of the patches winning. The devices were not defeated in argument. They were outcompeted, post by post and decade by decade, by the requirements of the work they were meant to supervise, and an assessment that blames only the wickedness of successors has not understood the problem the founders set themselves. They designed for a world in which administration was simple enough to be rotated through a citizenry, and the world industrialised out from under the design.

This reading also locates the exact point where the doctrine might have developed and did not. A tradition that took its own safeguards seriously as engineering would have spent the twentieth century iterating them: designing recall that bites without paralysing, rotation that respects tacit knowledge, transparency that defeats the official’s monopoly of the files, the very monopoly Weber named as the apparatus’s second advantage. Scattered attempts exist, in cooperative governance, in sortition experiments, in the open data practices by which some administrations now publish their own files, and they are the living descendants of the Commune kit, though rarely under its flag. The doctrinal mainstream, by contrast, treated the devices as relics of the heroic age, quoted them on anniversaries, and governed through the apparatus they were built to dissolve. The phrase and its safeguards share a fate: both survive as citation. The difference is that the safeguards, unlike the destination, could still be tested by anyone willing to build them at scale, and the tradition that owns them has shown, for a century, a settled preference for quoting over building.

Where the critics converge: power and knowledge

Bakunin wrote as a revolutionary romantic with a conspirator’s biography; Weber wrote as a liberal sociologist with a civil servant’s eye for files. They agreed about almost nothing else, and their convergence on the transitional state is therefore the most underused fact in the whole debate. Both men located the fatal mechanism inside the apparatus itself rather than in the intentions of its founders. Both denied that the class character of the personnel would decide the outcome: Bakunin’s ex workers governing from the altitude of the state, Weber’s officials governing from the depth of the files, are the same figure seen from different windows, the new man of power produced by the position he occupies. And both treated doctrine as an accelerator rather than a brake, Bakunin through the certainties of scientific government, Weber through the official’s conviction that the file’s way is the rational way, so that the rulers experience their permanence not as usurpation but as competence. When two hostile witnesses, working from incompatible theories of history, independently specify the same mechanism fifty years apart and the mechanism then describes the central cases of the following century, the specification has earned the standing this article gives it: not a refutation of everything the founders wrote, but a defeater for the one claim that required the mechanism not to exist.

A third convergence arrived from economics, and it closes the triangle. The public choice tradition of the later twentieth century, building on the older insight that bureaus seek budget and jurisdiction the way firms seek market share, modelled officials as maximisers within their constraints and derived the growth of apparatuses as an equilibrium rather than a betrayal. The vocabulary is alien to both founders and critics of the earlier rounds, but the conclusion is theirs: an organisation charged with a temporary task and staffed by people whose livelihood, status and expertise are the task will not, absent an external force with both the power and the interest to dissolve it, schedule its own redundancy. In the transitional state the external force is supposed to be the associated producers themselves, organised beneath and around the apparatus. The whole question, then as since, is whether that force can exist in the required form: informed enough to judge the apparatus, organised enough to act, and independent enough not to be staffed by it. The Commune devices, election, recall, rotation, workmen’s wages for officials, were the founders’ engineering answer to exactly that specification. The record shows the devices either never installed or installed as ceremony. What the record does not show, and what fairness requires noting, is a case in which the devices operated at scale, across decades, under a population determined to enforce them, and failed anyway. The engineering was never stress tested. It was bypassed. A doctrine may still be blamed for designing safeguards that every real revolution found inconvenient, since a safeguard that cannot survive contact with its intended environment is a design fault and not an excuse, but the blame should be assigned to the right ledger: the founders saw the danger, named it, and built against it, and the building was skipped by men who quoted them.

The conditions defence and the price it charges

Every serious defence of the withering doctrine mounted since the record accumulated takes the same form, and the form must be understood precisely, because the final assessment turns on it. The defence runs: the doctrine is conditional. It asserts that if classes disappear, if production is socially organised, if abundance replaces scarcity, then the state loses its function and lapses. None of those conditions has ever been satisfied. The revolutions occurred in backward, isolated, war ruined countries, the opposite of the advanced, international, prosperous settings the founders envisaged. Therefore the consequent was never triggered, the doctrine was never tested, and a conditional is not refuted by the failure of its antecedent. Put formally the defence is valid. A claim of the shape if P then Q is untouched by any number of cases in which P is false. The defenders are right about the logic, and critics who announce that the Soviet Union disproved the withering of the state, full stop, have committed an elementary error, the same error as announcing that a barometer is broken because the storm it was never exposed to failed to register.

But the defence pays for its validity, and the price list has three items. The first is that it relocates the doctrine’s real content. If the withering consequent is safe behind its conditions, then all the doctrine’s testable weight shifts onto the implied promise that the transitional process tends toward those conditions, that a state built to abolish classes actually abolishes them rather than entrenching a substitute. That tendency claim was tested, repeatedly, and it failed: the apparatus did not wither, and the conditions did not approach; in case after case the distance to classlessness grew as the apparatus grew, because the apparatus itself became the salient stratification. The second item is the mechanism gap, which is Bakunin’s point restated in the vocabulary of political science: the doctrine specifies no process by which the holders of state power are compelled to preside over the conditions of their own redundancy. Elections, recall and rotation were the founders’ devices, and they were overrun in every case by the Weberian advantages of expertise and continuity. A conditional whose antecedent can only be brought about by the voluntary self dissolution of an interested party, with no enforcement mechanism specified, is a conditional whose truth may be impeccable and whose relevance is nil. The third price is methodological, and it is the deepest. In the philosophy of science, a theory protected from every failure by an auxiliary clause that is never itself independently checked exhibits the pattern Popper diagnosed in the Marxism of his day: the followers reinterpreting theory and evidence until the two fit. The conditions clause is not automatically such a device; conditions clauses are the ordinary furniture of honest science. It becomes such a device at the moment its defenders can no longer say what would count, in advance, as the conditions having been approximated closely enough for the tendency to show. That moment, in the public literature of the doctrine’s defence, arrived long ago.

The objection can now be stated at its full strength, as this article’s brief requires, without softening. A theory whose central political destination has failed to materialise in every society that attempted the road to it, whose failure was predicted in advance, in print, by a contemporary rival through a mechanism that describes the failures case by case, and whose defence consists in conditions that recede exactly as fast as they are approached, such a theory, the objection runs, should be treated as refuted in the only court available to political doctrines, the court of its own specified tests, rather than as an unblemished conditional awaiting a world that will never supply its antecedent. The answer available to the doctrine is the distinction drawn above between the conditional claim and the unconditional tendency, and it is a real distinction: the conditional may stand while the tendency falls. But an honest assessment must then say what the concession costs. What falls with the tendency is the doctrine as a guide to politics, since politics lives entirely in the tendency, in the question of what building a transitional state actually does. What survives is a theorem about an imaginary society: in a world without classes, coercion would have nothing to do. That theorem may even be true. It is also, standing alone, very much less than the phrase has been used to promise for a century and a half.

The world scale condition: encirclement, armies and the last function

Buried inside the doctrine is a condition so large that it is easy to read past, and the founders stated it plainly whenever they were pressed: the transformation was expected to be international, or it was not expected at all. Marx and Engels held that communist society could arise only as the act of the dominant peoples all at once, on the stage of a world market that capital itself had built, because a single country attempting the transition in isolation would face scarcity at home and hostile states abroad, and both pressures would feed the apparatus. The withering, in particular, has a world scale prerequisite that its other conditions lack. A classless society might conceivably organise its internal affairs without a repressive apparatus, but as long as neighbouring states maintain armies, the society at the destination must maintain one too, and an army is not an administrative convenience. It is the historical nucleus of the state itself, the armed body apart from the population from which, on Engels’s own genealogy, the rest of the apparatus grew. Defence is the last function, and it is the one function whose necessity does not depend on class antagonism inside the society at all.

The tradition’s custodians understood the difficulty and answered it with an amendment that deserves to be quoted in every glossary entry on the phrase, because it quietly converted the doctrine. In his report to the Eighteenth Congress of the Soviet party in March 1939, Stalin announced that the state would remain even under communism if capitalist encirclement had not been liquidated and the danger of military attack from abroad persisted. The amendment was presented as a development of Engels rather than a departure from him, on the grounds that Engels had imagined the victory of socialism in all countries simultaneously. As exegesis the claim is not absurd; the world scale condition is genuinely in the founders. As doctrine the amendment was transformative. It supplied a permanent, externally defined reason why the apparatus could never begin to lapse, a reason whose satisfaction lay entirely outside the society’s own development and whose judge was the apparatus itself. After 1939 the withering in Soviet doctrine was not denied; it was postponed behind a geopolitical condition that no Soviet leader would ever certify as met, and the certification question answers itself. The amendment is the conditions defence hardened into state doctrine, and it marks the precise point at which the clause ceased, in the doctrine’s largest homeland, to be even a tendency claim and became a formula for explaining why the tendency would never be observed.

The encirclement problem also clarifies what a non cynical version of the doctrine would have to say. If defence is the last function, then withering is not a national process at all but a property of a world system, and its first observable symptom would not be the shrinking of any ministry but the demilitarisation of relations between post capitalist societies, armies lapsing together as the occasion for them lapses. Nothing resembling that symptom has appeared. The states that claimed the doctrine treated each other, within a generation, as strategic threats requiring the full apparatus of borders, armies and intelligence services, and the Sino Soviet split supplied the spectacle of two governments, each officially committed to the withering of the state, aiming the largest armies in their histories at a shared frontier. A reader assembling the dossier is entitled to find that spectacle decisive, not because it refutes the conditional theorem, which it does not touch, but because it displays the mechanism problem in its purest form: even where the doctrine’s holders faced each other, with class enemies in common and the texts memorised on both sides, the apparatuses generated antagonism between themselves. The state, it turns out, does not need classes in the founders’ sense to find its work. It needs only another state.

Stateless orders in the anthropological record

One body of evidence is usually missing from arguments about the withering, because both sides prefer to fight over the twentieth century, and a glossary article is the right place to admit it. Stateless societies are not a fiction. Anthropology and archaeology document many orders, some of them large and durable, that coordinated production, settled disputes, organised ritual and even war without a separate coercive apparatus standing over the population: segmentary lineage systems in which order emerged from balanced opposition between kin groups, age set systems that distributed authority across the life course, councils and assemblies whose decisions bound because custom bound, and sanctions that ran through reputation, compensation and feud rather than through police. The Nuer of the Sudan, in Evans-Pritchard’s classic account, numbered in the hundreds of thousands with no government in the European sense at all, and their order was not chaos but a different technology of order. The founders, who read the ethnography of their day with appetite, drew on exactly this material for the early chapters of the origin story: the gentile constitution of Morgan’s Ancient Society is the empirical base on which Engels built the claim that the state has a birthday.

The same evidence, followed honestly, cuts against the destination as sharply as it supports the origin. The documented stateless orders are, without exception, small scale, kinship saturated and economically simple beside the societies the doctrine addresses. Their coordination technology, the balanced segment and the customary sanction, works because everyone is related to everyone, because exit is possible into open land, and because the collective tasks are few and recurrent. None of those enabling conditions is available to a society of hundreds of millions running power grids and pharmaceutical supply chains, and the anthropological record contains no case of a complex, populous, technically advanced society shedding a state it once possessed. States in the record arise, persist, collapse under conquest or exhaustion, and fragment; the one transition that never appears is the orderly lapse of a mature apparatus into self administration as its functions become unnecessary. Where functions did become unnecessary, as with the guild and manorial jurisdictions already mentioned, the apparatus above them absorbed the jurisdiction rather than dissolving with it. The anthropological verdict, stated fairly, is that the founders were right about the state’s contingency and wrong about the direction of travel at scale: statelessness is humanity’s past at low complexity, and nothing in the record shows it as humanity’s future at high complexity. A doctrine of withering would need that future case, or a mechanism for producing it, and it has neither.

There is a subtler lesson in the same material, and it concerns the habit mechanism on which Lenin rested the destination. Stateless orders maintained themselves, insofar as the ethnography shows, by sanctions that were anything but gentle: the feud, the blood price, ostracism that amounted to a death sentence in a world without anonymous refuge. The absence of a state did not mean the absence of coercion; it meant coercion distributed through the kin network and applied without appeal, record or limit. The founders’ imagination of a post state order quietly assumed that the enforcement residue in their destination would be milder than the state’s, habit doing gently what police had done harshly. The anthropological record suggests the opposite presumption is at least as plausible: that enforcement, deprived of a specialised and therefore visible and criticisable body, does not evaporate but diffuses into hands no one elected and no one can recall. This is, once more, the Marxist charge against Bakunin’s federation, now recovered from the ethnography the founders themselves prized, and it belongs in the dossier as the point where the origin story’s own evidence testifies against the origin story’s ending.

Prediction, tendency, or hope: settling the claim’s status

The assessment just reached suggests the classification with which a glossary article should close its analytical work, because the three candidate statuses are not interchangeable and the claim behaves differently under each. As an unconditional prediction, that states built on the doctrine’s road will wither, the claim is dead: the sample is closed, the direction uniform, the rival forecast vindicated. As a conditional theorem, if classes and scarcity were abolished the coercive state would lose its function, the claim is alive but nearly empty, since its antecedent describes a society no one knows how to reach and its truth follows largely from the definitions it begins with; define the state as the organ of class repression and its lapse in a classless world is an unpacking of the definition, not a discovery about the world. The interesting status, and the one that best fits the founders’ own practice, is the middle one: a tendency claim about a historical process, of the same family as the tendency of the rate of profit to fall in Marx’s economics, where a law like pressure operates through counteracting influences and explains the shape of events without dictating their outcomes. Read as a tendency, the withering doctrine asserts that classless production exerts a standing pressure against political rule, and it invites the empirical question whether that pressure can be detected anywhere in the record, bending events even where it does not decide them.

That invitation should be accepted honestly, and the honest answer is that the pressure is hard to find. The places where state functions genuinely shrank in the last century shrank under pressures the doctrine does not name: fiscal limits, electoral choice, technological substitution, the collapse of legitimacy, and they shrank inside states whose coercive cores were untouched. The places where the doctrine’s own preconditions were pushed furthest, socialised production under a party committed to the destination, show the pressure running in reverse, the apparatus thickening as socialisation deepened. A tendency that manifests only as its opposite, in every observation, over a century, is a tendency in name only, and the classification exercise ends where the record section began. What remains is the third status, and it should be named without contempt: a hope. The withering of the state expresses a hope that rule over persons is a historical episode rather than a permanent feature of the human condition, that the museum of antiquities has a room reserved for the instruments of domination. Hopes are not refuted the way predictions are, and this one has done real work in the world, much of it honourable: it kept a suspicion of permanent officialdom alive inside movements that built officialdoms, it supplied generations of dissidents within those movements with a text to quote against their own apparatuses, and it still performs that office wherever the phrase is read against the grain of power. A glossary article’s duty is not to extinguish the hope but to label the shelf it belongs on. It does not belong on the shelf of established tendencies, and it was shelved there by its custodians for a hundred years at the cost of the doctrine’s credibility on everything else.

Two distinctions protect the assessment from overshooting, and precision requires them. The first is between the doctrine and the theory of the state that carries it. The origin theory, the state as a historically emergent apparatus bound to class division, remains a serious research programme with genuine explanatory victories to its name in the study of early states and of state formation, quite apart from the fate of its terminal prediction. A reader who discards the origin theory because the destination failed would be repeating, in reverse, the error of the custodians who protected the destination by immunising the theory. The second distinction is between the claim’s truth and its historical function. As an account of what the nineteenth century’s most consequential socialist tradition expected, and of why its states behaved as they did when expectation met apparatus, the withering doctrine is indispensable evidence, and historians will go on needing it long after its last believer. Glossary precision serves both distinctions: it keeps the phrase attributed, conditioned and classified, so that neither its defenders nor its prosecutors can borrow authority it never had.

When was the withering supposed to begin?

Only after the transitional work was finished: classes dissolved, production socially organised, abundance in view. The founders refused dates, Engels spoke of generations, and Lenin treated the higher phase as a long development. The state was expected to expand its functions first and lapse last.

What would count as a state function withering

One useful discipline remains before the misreadings are cleared away, because it converts the assessment from a verdict on the past into an instrument for the future. Suppose a reader wishes to know what evidence, in any society, would count as a state function genuinely withering rather than being renamed, outsourced or disguised. The doctrine’s own definitions supply criteria strict enough to be interesting. A function withers, on the account the texts give, when the need it answers disappears from the population’s life, so that the function lapses without a successor body inheriting it, without the population experiencing the lapse as a loss of protection, and without coercive enforcement migrating to a private or informal holder. A police function that passes to a private security industry has not withered; it has been sold. A court system replaced by party committees has not withered; it has been renamed. A border force disbanded because the neighbouring state has merged into a common political community has not withered either, on the strict reading, because the apparatus persists at the larger scale. The strict reading is demanding, and deliberately so: the doctrine promised the lapse of rule, not its rearrangement, and any fair test must hold the doctrine to its own promise.

Apply the criteria generously to the most favourable material available and a small residue survives, worth naming because an assessment that concedes nothing is as untrustworthy as one that concedes everything. Conscription has lapsed across much of Europe without successor bodies, because the mass army’s function, defence against a neighbouring mass army, genuinely receded; yet the professional forces that remain are states’ cores, and the lapse belongs to a function of interstate rivalry rather than of class rule. Certain moral police functions, the censorship of print and the prosecution of blasphemy, have withered in liberal states to the point of antiquarian curiosity, and they withered, notably, as the social conflicts they policed lost their charge, a mechanism Engels would have recognised even though the classes involved were not the ones his theory featured. Dueling, sumptuary enforcement and the guild restrictions were once state business and are now museum pieces in the most literal sense. The residue shows that functions can die when their occasions die. It also shows the pattern that undermines the doctrine’s specific claim: the functions that died were peripheral to the coercive core, and the core, taxation, criminal law, armed force, the administration of identity and movement, has in the same period done nothing but refine itself. A tendency visible only at the margins of an institution, while its centre consolidates, is not the tendency the doctrine predicted. It is the ordinary metabolism of states, which shed obsolete offices the way they shed obsolete weapons, in order to persist.

The criteria also expose, one final time, why the conditions table is the right instrument for this term and why the article built it early. Every other way of discussing the withering invites drift: into sincerity, into intentions, into the alternative histories of what might have happened in Russia with different leaders or in China with different policies. The criteria and the table recall the discussion to the doctrine’s own text. The state, defined as the organ of class rule, was to lapse when classes lapsed. Classes did not lapse; in the attempted cases, stratification reformed around the apparatus itself. The answer to the glossary question is therefore double, and both halves are now earned. The phrase means a specific, conditioned, carefully argued expectation about the fate of political rule. And the expectation, in the only form in which history could test it, was disappointed by the very process built to fulfil it.

How the phrase should be cited: a note on the texts

Precision in citation is the practical residue of everything argued so far, and a glossary article earns its keep by leaving the reader with habits that survive the closing of the page. The canonical passage stands in the third part of Anti-Duhring, in the chapter setting out the theoretical foundations of socialism, where Engels moves from the seizure of the means of production to the consequence drawn throughout this article. The same material, in the author’s own abridgement, forms the closing movement of Socialism: Utopian and Scientific, which is the form in which most readers in most languages have met it, and citation should say which form is being quoted, since the pamphlet’s compression is responsible for part of the phrase’s slogan like career. The German verb should be given at least once in any scholarly treatment, Absterben des Staates for the process as a noun phrase, because the English renderings divide between withering away and dying away and the division conceals the botanical and physiological register on which the metaphor depends. The museum sentence belongs to a different book, The Origin of the Family, Private Property and the State, and conflating the two sources, a common shortcut in textbooks, obscures the fact that the destination is argued twice, from the political economy of the present in one work and from the anthropology of the past in the other, by the same author in the same decade.

How should the phrase be cited in careful writing?

Attribute it to Engels, name Anti-Duhring or its abridgement Socialism: Utopian and Scientific, give the German Absterben once, and state the condition the sentence carries, the disappearance of classes, in the same breath. A citation that detaches the phrase from its author, its text or its condition repeats the standard error.

Three further habits protect the writer who uses the phrase in examinations, journalism or scholarship. The first is to keep the founders’ voices separate in the sentence itself: Engels formulated, Marx presupposed, Lenin elaborated, and the verbs are not interchangeable, since the whole attribution argument of this article stands behind them. The second is to mark the modality whenever the phrase is summarised: the state, on the doctrine, does not get abolished, is not dismantled, and is not scheduled for closure, and a paraphrase that installs an agent in the sentence has already falsified the claim it reports. The third is to carry the record with the definition in any context where the definition will be used as evidence about the world: a reference work that defines the withering of the state without noting that no state has withered is performing the same service for the doctrine that the Soviet textbooks performed, and the fact that the performance now serves scholarship rather than a party does not improve it. The series standard, applied throughout the state cluster of which this article forms a part, is that a term and its fate travel together or not at all.

The phrase as a standing rhetorical target

No glossary of this term is complete without an account of the work the phrase does in public argument, because for most readers the withering of the state is met not in Engels but in speeches, leaders and lecture halls, where it serves as the standing exhibit in the case against the entire tradition. The rhetorical use has a fixed shape. The speaker quotes the destination, usually attributing it to Marx, contrasts it with the nearest available gulag, police ministry or party headquarters, and invites the audience to conclude that the tradition is refuted by its own promise. The move is effective and, within limits, legitimate: a doctrine that advertises a destination cannot complain when the destination is checked against the journey. But the move as usually performed contains the two errors this article has been at pains to correct, and correcting them improves the criticism rather than blunting it. The attribution error matters because it lets the speaker skip the sequence, in which a transitional state that expands is part of the doctrine and not a deviation from it; the honest contrast is not between the promise and the apparatus, but between the sequence’s later stages and the apparatus’s refusal to enter them. And the conditions matter because the speaker who ignores them attacks a prediction the texts never made in that form, while missing the attack the texts cannot parry, the attack on the tendency, on the mechanism, on the century long failure of the distance to the conditions to close. Precision, in this corner of political argument, is not the enemy of polemic. It is what polemic looks like when it has done the reading.

The phrase performs the opposite office inside the tradition’s surviving movements, where it survives less as doctrine than as conscience. Dissidents within the states that claimed Marx’s name reached for it across the century: the Yugoslav Praxis philosophers asking why self management stopped at the factory gate, the workers of Gdansk in 1980 demanding that the unions and councils the doctrine promised be allowed to exist outside the party that monopolised them, the reformers of the Prague Spring reaching for a socialism with a human face in the vocabulary the founders had left them. In each case the phrase functioned as an internal standard against which the existing apparatus could be measured and found, in the doctrine’s own terms, unambiguously wanting. That function explains the custodians’ editorial instinct described earlier, their wish to keep the phrase revered and unexamined, because a phrase that measures the apparatus is a phrase the apparatus cannot safely gloss. It also explains the term’s curious afterlife in an age when the states that monopolised it have gone: the phrase remains in circulation, quoted by people who expect nothing from it as prediction, because as a standard of what rule claims to be for, and of how rule actually behaves, it has lost none of its edge. Glossary entries exist for words whose meanings are contested. Few words in the political vocabulary carry a contest in which the definition itself is the weapon.

One further public use deserves correction before the catalogue of misreadings, because it appears in respectable reference works and not only in speeches. The phrase is often glossed as the Marxist expectation that government will eventually become unnecessary because human beings will have outgrown conflict, a gloss that makes the doctrine a species of moral utopianism and its failure a failure of human nature. The texts support neither half of the gloss. Conflict, for Engels, is not outgrown; the specific conflict between classes is abolished at its economic root, and the administration that remains is a real activity with real objects, not a choir. The gloss matters because it relocates the doctrine’s failure to the one place where nothing can be learned from it, the supposed permanence of original sin, and away from the places where the failure is instructive: the sociology of apparatuses, the economics of scarcity, the geopolitics of encirclement, the design of safeguards. Readers who take their definition from this article will at least be disappointed in the doctrine for the reasons the doctrine itself supplies, which is the most a glossary can do for either side of an argument.

The corollary nobody quotes: democracy withers too

Lenin’s extension of the withering to democracy itself, noted in the elaboration section, deserves a section of its own, because it is the point where the doctrine’s logic becomes fully visible and where its modern readers, sympathetic and hostile alike, most often flinch. The reasoning is strict. If the state is defined as the organised coercion of one part of society by another, and democracy is a form of state, the form in which the majority coerces the minority through institutions, then democracy shares the state’s definition and must share its fate. In the higher phase, on Lenin’s account, there is no minority positioned by class for the majority to coerce, decisions approach unanimity as their material stakes fall, and the machinery of voting, counting and enforcing outcomes lapses for the same reason as the police: its function has gone. Engels had supplied the premise years earlier in Anti-Duhring, where democracy figures as a form of the state like monarchy and aristocracy, differing in who holds the apparatus rather than in whether an apparatus is held. The corollary is not a Leninist eccentricity. It is the doctrine completing its own syllogism, and any presentation of the withering that stops at policemen and prisons, leaving parliaments and elections standing in the destination like furniture the theory forgot to remove, has not understood the destination’s design.

The corollary repays the same functional scrutiny as the administration of things, and it fares no better, for reasons that clarify the whole doctrine. Majority decision is not only an instrument of class coercion; it is also a technology for closing disagreement among equals who share no class antagonism at all, disagreeing instead about risk, taste, evidence and the proper price of the future. A classless society deciding whether to build the dam, fund the observatory or license the experiment faces the full menu of human dissent with no class key in which to resolve it, and it must either vote, delegate, draw lots or fight, the four historically attested closers of collective argument. Voting and delegation are the state in the doctrine’s strict sense, coercion of the dissenting minority by the prevailing number. Lot is a genuine alternative with an honourable ancient pedigree, and the founders’ silence about it is a missed resource: sortition closes argument without a permanent apparatus and without a coercing majority, though it decides without deliberating and scales awkwardly into technical questions. The doctrine, committed to the disappearance of the problem rather than to a machinery for managing it, never developed these alternatives, because its premise was that the disagreements requiring closure were class disagreements in disguise. The premise is the wager identified earlier, now visible in its institutional form. If even a modest residue of genuine, non class disagreement survives into the destination, then the destination needs a closer of arguments, and a closer of arguments that can enforce its closures against the recalcitrant is, by the founders’ own definition, a state, however small, however rotated, however politely named. The withering, followed rigorously, thus abolishes itself at the last step unless human disagreement itself withers, and the texts, which are materialist about everything else, become quietly idealist at exactly this door, assuming a harmony of interests they nowhere argue for.

This is also the point at which the doctrine’s democratic critics and its authoritarian practitioners reveal their shared evasion. The practitioners built states in which the corollary was inverted: democracy, in any sense the founders would have recognised, withered first and fastest, while the apparatus, which was supposed to wither last, did not wither at all, a sequence that reads like the doctrine run backwards as farce. The democratic critics, for their part, rarely press the corollary, preferring to defend elections against the destination in the name of freedom without noticing that the doctrine, read strictly, regards their elections exactly as it regards the police, as machinery for a stage of society, admirable in the stage and obsolete beyond it. A glossary that restored the corollary to the definition would at least force the argument onto its proper ground. The question the withering poses to democrats is not whether they favour freedom. It is whether they believe majority coercion is a permanent necessity of associated life or a historical instrument with conditions of its own, and that question, posed by a doctrine whose answer failed, remains genuinely open in a way the doctrine’s other questions do not. It is the one corner of the edifice where the ruin is still a building site.

After the attempts ended: the phrase in a colder climate

The states built in the doctrine’s name have now nearly all ended, and the manner of their ending supplies the dossier’s final exhibit, which an article written with the record complete is obliged to register. The Soviet Union and its allied regimes did not wind down their apparatuses as their official ideologies lost conviction. The apparatuses remained at full establishment to the last season, and then the states themselves ceased, by dissolution from above, by negotiated round tables, by the simple refusal of their own officials to defend them. The withering doctrine had imagined the state’s end as the slowest of processes, a lapse measured in generations as functions drained into society. The actual ends were among the fastest political events on record, states with armies, police and ministries vanishing as organisations within months, while their functions, taxation, borders, criminal law, the issue of passports, were inherited intact by successor states within the year. History, offered the choice between the doctrine’s modality and its opposite, chose the opposite with unusual unanimity: states turned out to be mortal not by obsolescence but by collapse, and everything beneath them turned out to be immortal, passing from apparatus to apparatus like equipment between tenants. A doctrine of how the state ends that has the modality exactly backwards has failed in a deeper place than its timetable. It has failed in its picture of what a state is, mistaking for an organism, which dies when its conditions fail, what is closer to a legal shell over a stock of functions, a shell that can crack in a season while the functions march on under new management.

The scholarly reassessment that followed the endings has been more discriminating than either the celebrations or the funerals of the time, and its discriminations belong in this article’s balance sheet. Historians of the Soviet experiment, working the opened archives, have largely confirmed the apparatus story told here while complicating its causes: the security organs’ growth owed as much to the rulers’ fear of their own population as to any functional necessity, and the bureaucracy’s weight crushed initiatives for which no Weberian logic can account. Theorists of the state, meanwhile, have quietly detached the founders’ origin question from their destination, and the study of how states arose, why they persist and what they are for proceeds in several disciplines with the nineteenth century’s questions still on the table and its answers under permanent revision. The phrase itself survives the reassessment in the way this article’s status section predicted: as a standard and a hope, quoted now most often by critics of actually existing administration of every political colour, a measuring rod for the distance between what rule promises and what it does. There is a fittingness in the destination that no one foresaw. Engels promised that the state would end in a museum of antiquities. It is the phrase that has entered the museum, carefully labelled, attributed at last to its correct author, and visited by readers who come to study not a prophecy but the most instructive failure in the modern theory of the state: the failure of the one doctrine that made the end of rule its test, and was tested.

Misreadings the phrase attracts

The corrections scattered through the preceding sections gather naturally into a short catalogue, stated here in prose because these are the errors a reader is most likely to meet in the wild and the ones the series exists to retire. The first misreading is the attribution: the phrase is Engels’s, from Anti-Duhring, and crediting it to Marx is an error of fact that also conceals the difference between Marx’s guarded transitional writings and Engels’s systematic destination. The second misreading is the collapse of the sequence: withering is the last stage of a process that begins with a state expanding its functions, and any summary that has the state dissolving at the moment of revolution has confused the doctrine with the anarchism it was coined against. The third misreading is the equation of withering with abolition, as though the two verbs were stylistic variants; the whole point of the passage is that they name opposite modalities, one an act of will performed on the state, the other a loss of function suffered by it, and the doctrine stands or falls on the distinction. The fourth misreading holds the claim to be untestable, a wisp of utopian vapour; in fact it specifies conditions with unusual clarity for a political doctrine, which is precisely why its record can be assessed and why the assessment is damaging. The fifth misreading, flattering to the doctrine’s memory, is that Lenin expected an early withering and was betrayed by successors; Lenin’s own text places the destination beyond a lengthy transition, extends the withering to democracy itself, and supplies the badge of fidelity by which later custodians policed the faithful, so the gap between text and record cannot be billed to his heirs alone.

A sixth misreading belongs to the doctrine’s opponents and is worth correcting in the same spirit of fairness. It is sometimes said that the withering doctrine proves its authors were innocents who imagined human beings without conflict, envy or ambition. The texts do not support the charge. Engels’s state arises from real antagonism and his destination depends on material transformation, not on moral improvement; Lenin’s mechanism is habit formed under new conditions, not a new species of man; and the founders’ polemics against utopian socialism were precisely attacks on politics founded on exhortation. The doctrine’s vulnerability lies elsewhere and is more interesting: in its confidence that the antagonisms requiring coercion are all, at root, class antagonisms, and that no residue of conflict over status, belief, risk or love of power would survive their abolition to keep an apparatus in business. That is an empirical wager about the sources of human conflict, stated by implication and never defended head on in the texts, and it is the wager the twentieth century, with its conflicts raging inside officially classless orders, most plainly called. A criticism that locates the failure there has found the doctrine’s true weak joint, and it is a better criticism than the cartoon of innocence it replaces.

The phrase in the official textbooks: doctrine under management

Between the founders’ texts and the citizens of the states built in their name stood the textbook, and the textbook’s handling of the withering is the most revealing document in the whole dossier, because it shows what an institution does with a sentence that predicts the institution’s end. The Soviet textbooks of dialectical and historical materialism, through which generations of students, officials and officers passed, retained the phrase without exception; it was too canonical to drop and too dangerous to examine. Their solution was choreography. The withering appeared in the final chapter, after the stages, the phases and the laws, in the optative mood of a destination whose conditions were recited and whose arrival was assigned to a communism that the same textbooks placed beyond every planning horizon in use. The 1961 programme of the Soviet party attempted one act of scheduling, promising that the material and technical basis of communism would be built by about 1980 and that the rising generation would live under communism, and the promise’s quiet burial, unmentioned in the textbooks that succeeded it, taught the system’s own students the operational meaning of the destination more plainly than any commentary: the horizon is for pointing, not for reaching. A doctrine’s textbook afterlife is not the doctrine, and scholars err when they grade Engels by his catechists. But the afterlife is evidence about the doctrine’s usability, and a destination clause that every custodian, across three generations and several continents, found it necessary to embalm rather than to operate has failed a test of its own, the test of whether the tradition that owned it could afford to believe it in public.

The Chinese textbooks added an innovation that completes the typology of management strategies. Where Soviet doctrine postponed the destination behind encirclement and programme horizons, the Maoist doctrine of the continuation of the revolution under the dictatorship of the proletariat revised the sequence’s middle term: classes and class struggle, it held, persist throughout the socialist period and repeatedly regenerate, so that the transitional state must not merely endure but periodically convulse itself against its own bureaucracy. The theory had the merit of addressing the record honestly at one point, since it conceded that stratification reappeared inside the apparatus and named it a problem in the doctrine’s own vocabulary. Its practical career completed the pattern this article has followed from the start. The campaigns launched to prevent a new class from consolidating were administered by the apparatus, enlarged it, and left it, after each convulsion, more comprehensively in command than before, while the destination receded behind a theory that made its preconditions self renewing: if classes regenerate under socialism by doctrinal necessity, then the withering, whose first condition is their disappearance, is postponed not by failure but by design. Between the Soviet postponement, the Chinese regeneration clause and the social democratic quiet dropping of the clause altogether, the twentieth century exhausted the available strategies for keeping a destination one does not expect to reach. The phrase survived all of them, in print, in examinations, in the mouths of dissidents who turned it against its custodians. Survival as citation is the afterlife the status section assigned it, and the textbooks, meaning to preserve the doctrine, preserved instead the evidence for the assessment this article has given.

For students and exam candidates

The mark carrying detail on this topic is the abolition contrast, and an answer built around it will outscore an answer built around the slogan. Open by attributing the phrase correctly: it is Engels’s, from Anti-Duhring, stated against the anarchist demand that the state be abolished at a stroke. Then set the sequence in order, revolution, transitional state, disappearance of classes, withering, and resist any phrasing that lets the state fade from the first day. Examiners reward the candidate who can also state the conditions the texts name, classes gone, production socially organised, the government of persons replaced by the administration of things, because conditions convert a vague summary into a testable account. Keep Marx and Engels in separate sentences: Marx wrote the transition, Engels the destination, Lenin the mechanism. Close the answer with evaluation that names its witnesses, Bakunin’s prediction of a new ruling stratum from 1873 and the record of states that grew, rather than with a general remark about utopia. Candidates preparing notes on the state cluster can keep the attribution, the contrast and the conditions table together in the Marxism study notes companion, where the distinctions this article draws are set out for revision.

For teachers and seminar leaders

This term teaches best as an argument reconstructed before it is judged, and the seminar room is the one place where the reconstruction can be made an exercise rather than a lecture. Set the Anti-Duhring passage beside Bakunin’s counter-prediction from Statism and Anarchy and ask the group to specify, before any verdicts are allowed, what evidence would settle the contest between them; students discover quickly that the two texts were designed, half a century before the experiments, as rival mechanisms with incompatible outcomes, and that the twentieth century ran the experiment for them. The conditions table in this article works well as a handout for the second meeting, with each row assigned to a pair of students who must argue both that the condition was never met and that it was never going to be. Guard against the two classroom failure modes: the session that treats the phrase as a curiosity of Victorian optimism, and the session that treats the record as a debating trophy. The material is stronger than either. It shows a great tradition specifying its own test, a rival specifying the counter test, and history returning a verdict detailed enough to grade. Few topics in political theory let students watch a prediction contest with the entry forms preserved.

For researchers, historians, and journalists

Three professional habits cover most of the errors this phrase generates in print. The first is attribution: the formulation is Engels’s, and copy that gives it to Marx is wrong in a way a letters page can correct, since the passage stands in Anti-Duhring and in its abridgement Socialism: Utopian and Scientific, while Marx’s cognate texts, the Gotha Critique and the Commune writings, argue the transition and leave the destination’s description open. Journalists quoting the phrase against or for any government should therefore quote it as Engels and pair it with its condition, the disappearance of classes, rather than with a date or a promise no text contains. The second habit is modality: the state in this doctrine dies away and is not abolished, dismantled or wound up, and paraphrases that install an administrator falsify the claim in the act of reporting it. The third habit is separation of the claim’s layers, prediction, tendency and hope, which the assessment section of this article distinguishes; most public argument trades on sliding between them mid sentence. Researchers will find the primary dossier compact: Anti-Duhring and The Origin of the Family for the formulation, the Gotha Critique for the conditions, chapter five of State and Revolution for the mechanism, Bakunin’s Statism and Anarchy for the counter-prediction, and the 1939 encirclement amendment and the 1961 programme for the doctrine’s official fate in power. Staffers preparing a principal for questions on the phrase need above all the conditions and the record in one breath, since the phrase is a standing rhetorical target precisely where those two are quoted apart.

Frequently Asked Questions

Q: What does the withering away of the state mean?

It means the gradual lapse of the state, understood as a separate coercive apparatus standing over society, once the classes whose antagonism called it into being have disappeared. The phrase is Engels’s, from Anti-Duhring. On the theory, an institution defined by its function, holding class conflict within bounds, loses its reason to exist when the function goes, and dies away as organs and customs die away, without being abolished by decree or dismantled by any office. Administration, coordination and the management of common affairs are expected to continue; what ends is rule over persons by a body apart from the population.

Q: Did Marx say the state would wither away?

No. The phrase and its canonical formulation are Engels’s, published in Anti-Duhring in 1878 and popularised through Socialism: Utopian and Scientific. Marx’s own writings on the end of the state are fewer and more guarded: the letter to Weydemeyer of 1852 names the abolition of all classes as the destination, the Civil War in France praises the Commune for breaking the state machine rather than capturing it, and the Critique of the Gotha Programme describes the transitional period while expressly leaving the question of remaining state functions to a future science. Attribution matters because the confident mechanism belongs to Engels, and Lenin after him, rather than to Marx.

Q: What is the difference between abolishing and withering away?

Abolishing is an act: a revolution or assembly decrees the state out of existence at a stroke, which was the anarchist demand Engels was answering. Withering is a process: the state loses its function as classes disappear and lapses the way an organ or a custom lapses, with no agent performing the ending and no date attached. Engels set the verbs against each other deliberately, arguing that coercion is an effect of class antagonism and cannot be repealed before its cause. The abolition contrast is the key to the phrase: withering was coined as the rejection of immediate abolition, not as a synonym for it.

Q: Why would the state wither away?

On the theory’s own logic, because its cause would have been removed. The state, in Engels’s account, arose when society split into classes whose conflict required a power standing apparently above society to keep order. If production is socially organised, classes and their antagonisms disappear, and distribution in conditions of growing abundance no longer needs enforcing, then the apparatus has nothing left to do: no class to hold down, no scarcity to police, no resistance to suppress. Lenin added the mechanism of habit, arguing that people accustomed to observing the elementary rules of social life without compulsion would simply stop generating the demand for a coercive body.

Q: Has any state ever withered away?

No. The record contains no case of a state lapsing as its functions became unnecessary, and the states built in the Marxist tradition’s name moved uniformly in the opposite direction, expanding their coercive and administrative apparatuses across their whole existence. The Soviet Union revised the doctrine instead, first through the 1939 amendment that the state would remain while capitalist encirclement lasted, then through the 1961 formula of the state of the whole people. When those regimes ended, they ended by collapse or negotiated surrender, and successor states inherited their functions intact. Individual functions have lapsed at the margins of states, but the state as such has never withered anywhere.

Q: What is the administration of things?

It is the formula, descending from Saint-Simon and adopted by Engels, for what replaces political rule in the destination: the government of persons gives way to the administration of things and the direction of the processes of production. The image is of a society whose collective business has become technical management, scheduling, maintaining, allocating, rather than command over people. Critics press two questions the founding texts leave open. Administration does not choose ends, and choosing among contested ends is what politics is. Administration also requires enforcement against defectors, and an enforcing body begins to resemble the apparatus the formula was meant to replace.

Q: Does bureaucracy make withering impossible?

Weber’s argument says effectively yes, in principle rather than merely in practice. Bureaucracy, on his account, is the technically superior form of administration for any complex task, staffed by trained officials whose expertise and command of the files give them structural advantages over every nominal master. Socialising production would not dissolve that apparatus; it would fuse the state’s bureaucracy with industry’s into a single hierarchy. The record of the attempted transitions, in which elected bodies steadily yielded real initiative to permanent staffs, is consistent with the argument. Replies within the tradition hold that Weber generalised from scarcity societies, but no observed case shows an apparatus retiring as he predicted it never would.

Q: What did Bakunin predict about a workers state?

Writing in Statism and Anarchy in 1873, Bakunin predicted that a workers’ state would manufacture a new ruling stratum out of its own personnel. If the proletariat rules, he asked, over whom does it rule, and he answered that the former workers who govern would cease to be workers, would regard the labouring world from the altitude of the state, and would defend their position as every ruling group does. He added that a government of scientific socialists would be especially oppressive because its certainty would treat dissent as error. The prediction was made before the canonical withering texts had finished appearing, and it describes the trajectory of the later party states closely enough to stand as the doctrine’s standing test case.

Q: Is withering away a prediction or a hope?

The article’s assessment separates three statuses the phrase slides between. As an unconditional prediction about states on the doctrine’s road, it has failed: the direction of every case runs the other way. As a conditional theorem, if classes and scarcity were abolished the coercive state would lose its function, it stands largely by definition and has never been triggered. As a tendency claim it should show some pressure in the record and shows, if anything, the reverse. What survives intact is the hope that rule over persons is a historical episode, and that hope has done real work as an internal standard against which the tradition’s own apparatuses were judged by their dissidents.

Q: What conditions does withering away require?

The texts name a connected set. Classes and class antagonisms must have disappeared, so that no group requires holding down. Production must be socially organised under the associated producers rather than directed by private owners or by a ministry standing in for them. Material abundance must have advanced far enough for distribution by need to replace accounting by contribution, the higher phase of the Gotha Critique. And the government of persons must have given way to the administration of things, with no special repressive force needed to keep order. The withering conditions table in this article sets each condition against whether it has ever obtained and what happened instead.

Q: What would count as evidence that a state function is withering rather than growing?

The doctrine’s own definitions supply a strict test. A function withers when the need it answers disappears from social life, so the function lapses without a successor body inheriting it, without enforcement migrating to a private or informal holder, and without the population experiencing the lapse as lost protection. A police task sold to a private security firm has been outsourced, not withered; a court replaced by a party committee has been renamed; a ministry absorbed by a larger state has been relocated. Evidence of genuine withering would be a function whose occasion has demonstrably gone, conscription after the mass army’s disappearance is a partial example, leaving no apparatus in its place.

Q: How does Engels’s account of the state’s origin set up its disappearance?

In The Origin of the Family, Private Property and the State, Engels argued that the state is not eternal but arose at a definite stage, when society divided into classes and required a power, apparently standing above it, to moderate their conflict and preserve order. An institution with a birthday can have a deathday: if the state exists because class antagonism exists, then the end of class antagonism removes the state’s reason for being, and withering is the origin story read backwards. The whole claim stands or falls with that genealogy, which is why rival accounts of the state’s origin, above all Weber’s, function as rival predictions about whether it can end.

Q: Why does the administration of things still need decision and enforcement?

Because administering things settles how chosen ends are achieved, while societies must still choose the ends: how much to invest, which risks to run, which regions and projects to favour. Those choices distribute advantages among people who disagree, and reaching and enforcing them is the activity called government under any name. Enforcement remains necessary because rules bind only if defectors can be checked, and the founding texts meet the point with the mechanism of habit rather than with institutions. Critics from Weber onward conclude that a body empowered to decide and enforce at scale is a state function, whatever the destination calls it, and that the formula renames the problem instead of solving it.

Q: What should a student separate in Bakunin’s warning to Marx?

Three things. First, separate the prediction from the prescription: Bakunin’s forecast that a workers’ state would produce a new ruling stratum was vindicated by the record, while his remedy, immediate abolition and free federation, collapsed into statehood in its one great trial, Spain in 1936. Second, separate the two strands of the warning, the material strand about office holders acquiring interests of their own and the doctrinal strand about rulers armed with a science of history treating dissent as error; the strands can be assessed independently. Third, separate the argument from the polemic around it, including the personal venom of the original quarrel, which neither side’s modern readers defend and which has no bearing on whether the prediction was accurate.

Q: How long a timescale does the withering claim quietly assume?

A long and deliberately unspecified one, and the refusal to specify is part of the claim’s structure. Marx declined to answer questions about the destination’s arrangements in advance, Engels spoke in terms of generations reared under new conditions, and Lenin treated the higher phase in which the state could lapse as a development no one could date. The sequence requires classlessness and abundance as preconditions, so the withering inherits the timescale of those transformations rather than the timescale of a parliament or a plan. Critics note the cost: a destination without a deadline can absorb any amount of contrary evidence as prematurity, which is why the assessment of the claim turns on tendencies rather than on dates.