E062. Millenarianism and the Limits of Institutional Correction

The conventional reading of millenarian movements treats them as failures of reason. A population under stress abandons proportion, adopts a totalizing explanation, and waits for an end that does not arrive. On this reading the movement is an aberration, and the appropriate response is correction: better information, better education, a restored grip on evidence.

The reading is wrong, and it is wrong in a specific way that matters for anything one wants to say about the present. Millenarianism is not a departure from ordinary cognition. It is ordinary cognition running under a particular input condition — one in which the available explanatory frame has stopped covering observed reality, and the institutions that would normally correct for cognitive limits are themselves implicated in the failure. The believer and the official share an architecture. What differs is position.

This essay argues three things. First, that the conditions producing millenarian belief are structural and identifiable, not psychological and diffuse. Second, that the contradictions internal to such belief are load-bearing components rather than defects, which is why exposing them accomplishes nothing. Third — and this is the symmetry that organizes the argument — that the institutional machinery built to correct for human cognitive limits is retrospective by construction, therefore always late, and that the interval between the last correction and the next failure is precisely where millenarianism recurs. It is not that people become less rational. It is that the corrective is structurally one crisis behind.

I. The Conditions

Norman Cohn's survey of medieval European millenarian movements produced a finding that has held up better than most in the sociology of religion: these movements did not originate in the poorest regions. They clustered in Flanders and the Rhineland — commercially developed, rapidly urbanizing, comparatively wealthy. The common feature was not deprivation but detachment. The populations involved had been separated from traditional structures of belonging: no longer bound to the manor, not admitted to the guild, present in the city without a place in it.

Uprooting, not poverty, is the operative condition. Absolute misery tends to produce resignation. What produces eschatology is the loss of a position one previously held or expected to hold.

James Davies added the temporal shape. Revolutionary sentiment concentrates not at the trough but at the inflection — after a period of improvement that raises expectation, when material conditions turn and the gap between the expected and the actual widens fastest. Steady stagnation does not generate apocalyptic politics. Interrupted ascent does.

The third condition is explanatory collapse. Millenarian episodes cluster where the prevailing interpretive frame stops accounting for events. The plague is the standard illustration: a third of a population dies, existing theodicy cannot absorb it, and the flagellant movements follow. The crisis is not the mortality alone but the failure of the available explanation to contain it.

The fourth is a mediating stratum. Cohn's leaders were rarely of the poor. They were lapsed clerics and displaced intellectuals — literate, fluent in eschatological vocabulary, and without a position inside the institutions that vocabulary belonged to. Popular grievance does not become a cosmology on its own. It requires translators who have the language and lack the post.

The pattern is not European. The Taiping rising followed the collapse of Qing authority after the Opium Wars, in a region where the Hakka were structurally marginal. The Ghost Dance followed the destruction of Plains subsistence. Cargo movements followed the sudden arrival of external power and material in the Pacific war. In each case the same four conditions: detachment from prior structure, interrupted trajectory, an explanatory frame that no longer covers events, and a stratum positioned to supply a new one.

II. The Architecture

Millenarian and conspiratorial systems are usually attacked at their contradictions. The attack fails, consistently, and the reason is that the contradictions are not errors in the system. They are how it holds together.

The clearest pair is omnipotence and incompetence. The adversary must be powerful enough to explain a wide field of unrelated misfortunes, and careless enough that the believer could have discovered the plot. Remove either and the structure fails: without omnipotence there is no unified explanation, without leakage there is no evidence. The two propositions are never deployed in the same context. Omnipotence appears when the scope of the claim is at issue; incompetence appears when the provenance of the evidence is challenged. The believer does not experience a collision because the two are never held up together.

Other pairs perform the same double duty.

Concealment and display. The plan is secret and simultaneously marked in symbols, architecture, and media. Concealment justifies the absence of documentation; display gives the believer something to do. Interpretation is the mode of participation.

Imminence and deferral. The completion is always near and never arrives. Imminence mobilizes; deferral absorbs failure. Festinger's observation — that disconfirmation, under the right conditions, increases rather than reduces conviction — describes the mechanism. The date was misread; the delay was foretold.

Scale and security. The operation involves thousands and no one defects. Scale extends explanatory reach; security accounts for the absence of leaks. David Grimes has estimated the arithmetic: the probability of disclosure rises steeply with the number of participants, which makes the two claims jointly implausible at exactly the scale the theory requires.

Alienness and ubiquity. The adversary is categorically unlike us and also embedded in our institutions and our street. Alienness supplies moral license; ubiquity makes anyone available as a target. This is the pairing with the worst historical record, and the one that converts belief into violence.

Mass deception and mass awakening. The public is comprehensively deceived, and the truth is spreading rapidly. Deception makes the speaker a member of a knowing minority; awakening supplies a horizon of victory. Isolation and hope are delivered together.

Evidentiary rigor and unlimited credulity. Mainstream accounts require original documentation and complete causal chains; internal accounts require an anonymous source. This is the most frequently observed form, and it is not an evidentiary standard at all. It is a membership test wearing the clothes of one.

Each pair divides two incompatible demands: maximize explanatory coverage, and block falsification. A system that resolved its contradictions would become falsifiable, which would make it worse at its job. Popper's objection to unfalsifiable theories identified this correctly and drew the wrong practical conclusion — that identifying the flaw would be corrective. It is not. Hofstadter's insight was the more useful one: the paranoid style is a political idiom rather than a clinical condition, and in an idiom, tolerating an obvious contradiction functions as a demonstration of commitment. Pointing at the contradiction is therefore where cohesion is manufactured, not where it breaks.

III. The Case

American politics after 2016 fits the four conditions with unusual precision.

Detachment: the work of Autor, Dorn, and Hanson established that commuting zones exposed to import competition after China's WTO accession lost manufacturing employment durably, without recovery. Subsequent work by the same group linked those zones to political movement toward both extremes. Case and Deaton documented rising mortality among middle-aged non-college whites — anomalous for a developed country, and a measurable index of status loss. These were not the poorest places. They were places that had been solvent and stopped being solvent, which is Cohn's condition rather than the intuitive one.

Interrupted ascent: the asymmetry of the post-2008 recovery. Financial institutions were stabilized; foreclosures proceeded. The subsequent expansion favored asset holders over wage earners. The detonation came not at the trough but during the period in which recovery was being announced and not experienced.

Explanatory collapse: both parties had converged on trade liberalization, immigration, and interventionism. A voter opposed to all three had no available vehicle. The 2016 Republican primary was decided on precisely this — the candidate ran against the party's own foreign policy establishment, and that opposition constituted his legitimacy. The Tea Party was the earlier signal.

Mediating stratum: an alternative media ecosystem occupied the position Cohn assigned to displaced clerics, performing the same translation of grievance into cosmology.

The eschatological form followed. Make America Great Again is a restoration narrative rather than a utopian one, which is the standard millenarian grammar — the golden age is recovered, not built. Drain the swamp is a purification motif. The deep state is the concealed, ubiquitous adversary. From 2017 the QAnon material completed the form, supplying a day of judgment, a body of encoded text requiring interpretation, and a prophecy that remained imminent and continuously deferred.

IV. Millenarianism in Office

The interesting problem is what happens when this structure wins.

An eschatological movement in opposition has an inexhaustible supply of adversaries. In government it does not. After two terms, a substantial fraction of the establishment consists of one's own appointees. The first term demonstrated the mechanism: Sessions, Mattis, and Bolton became adversaries in sequence. The second term's response — pre-emptive loyalty screening — solves the defection problem and creates a worse one, since a wholly loyal administration leaves no one to blame for failure.

The target therefore migrates. It moves toward unelected institutions, which can be described as not yet captured: courts, universities, press, central bank. And it moves outward, toward allies and neighbors. As of August 2026 the second term has conducted strikes in seven countries — Iran, Iraq, Nigeria, Somalia, Syria, Venezuela, Yemen — and has threatened or left open strikes against seven more, a list that includes Canada, Greenland, Mexico, Panama, and Oman.

The second structural problem is the simultaneous requirement to claim achievement and sustain threat. The narrative depends on imminent catastrophe; office requires reporting improvement. Both must run at once. This is the imminence-deferral pair transposed into governance, and it is managed the same way: achievement is defined as immediate and visible — announcements, signing ceremonies, enforcement statistics — while the threat is defined as structural and unfinished. The scale of inherited damage is continuously revised upward, which makes any term insufficient by construction.

The Iran campaign illustrates the specific failure mode. The pattern through 2026 has been announcement followed by retraction: strikes promised for a given night in June, including against Kharg Island and oil infrastructure, cancelled within hours; an August declaration of readiness at a scale unseen since the Second World War, cancelled the same day. CNN's own summary of the sequence noted that the specific threats have repeatedly gone unexecuted even where the stated conditions were not met, and that the indications of progress cited on withdrawal have consistently failed to materialize. Positions reverse within days — a Hormuz agreement described as imminent on one day, the Iranian leadership described as thoroughly duplicitous on the next. A back channel to the Revolutionary Guard was acknowledged in mid-August and the existence of any talks denied days later. A threat to bomb Oman was assessed by a former Defense Department Middle East adviser as not credible in any sense.

The damage here is precise rather than general. Coercive diplomacy runs on credibility. Repeated non-execution spends the asset the campaign exists to display.

The North Korean outreach reads as a substitute. Reporting in August 2026 indicated pressure on staff to arrange a meeting with Kim Jong Un as early as the autumn, possibly around the November Asia trip; an order to reduce joint exercises with South Korea substantially, cutting Ulchi Freedom Shield roughly in half; and a public claim that Kim had responded positively, without detail and without confirmation from Pyongyang. Analysts connected the interest directly to the absence of a result in Iran.

The conditions are worse than in 2018 on every axis. North Korea formally abandoned reunification as a goal in 2023, signed a mutual defense treaty with Russia, acquired combat experience through deployments to Ukraine, deepened missile cooperation with Iran, and is now estimated by independent analysts to hold warheads in the hundreds. South Korea's role as intermediary — the mechanism that produced the 2018 and 2019 meetings — has been structurally eliminated. The costs of the pursuit are being displaced onto the ally, which has been publicly criticized for declining to join the Iran campaign and privately expects to be bypassed.

Two further constraints follow from the absence of re-election. The restraint imposed by a future contest disappears; the onset of lame-duck status arrives earlier. The rational allocation shifts toward measures that are difficult to reverse — personnel, judicial composition, treaty withdrawal, agency dissolution. What such an administration leaves behind is structural alteration rather than policy.

And the succession problem is the one Weber described. Authority resting on personal charisma routinizes into institution or dissolves. Designating a successor early bleeds power; declining to designate one ends the movement with the term. Transfer is the rare outcome.

V. The Symmetry

Here the argument turns.

Everything above treats cognitive limitation as something the governed suffer from. The governing class operates under the same limits, and the record shows it plainly.

Consider three decisions, each of which is now read in some quarters as betrayal.

Germany's nuclear exit. The 2000–02 phase-out was agreed under the Schröder government and fixed in law in 2002. The anomaly is not 2011 but 2010, when the Merkel coalition reversed it with an average twelve-year extension, declared the amendment exempt from Bundesrat consent, and drew constitutional challenges from nine states. The 2011 reversal was therefore the undoing of the government's own departure. It moved in three days from Fukushima to moratorium; the moratorium itself was later held unlawful in both form and substance by the Hessian administrative court and finally by the Federal Administrative Court, and operators claimed damages of roughly €880 million. The Constitutional Court upheld the 2011 statute in December 2016 and made one finding that bears directly on the question of judgment: the legislature was permitted to treat Fukushima as an occasion for action even in the absence of new information about risk. The risk had not changed. The price of continuing to bear it had.

Nord Stream 2. The strategic case against dependence was not discovered in 2022. It was made contemporaneously by Poland and the Baltic states in 2015, considered, and rejected. What matters is not that the decision proved wrong but that it was taken in the year following the annexation of Crimea, and that the window to stop it remained open through construction. That is a sequence of decisions, not an inheritance.

French pension financing. The demographic projections underlying the pay-as-you-go system have been public for decades. The obstacle was never information. It was that no actor had a political horizon long enough to act on it — and that in France the item in question is not a budget line but a clause of a settlement made in 1936 and 1944, when the alternative to social guarantees had recently been demonstrated. Retrenchment there is understood as a proposal to reopen a containment agreement, which is why it produces the reaction it does.

These are ordinary decisions under ordinary cognitive constraint: horizon compression, discount rates set too high, the systematic preference for costs that fall after one's tenure. They are also the same errors that, viewed from below and after the fact, look like nothing but betrayal — and the reading that calls them betrayal is itself an artifact of hindsight bias, the same limitation operating from the other side.

This creates an asymmetry worth naming. If cognitive limitation excuses the decision-maker, it equally excuses those who now misread the decision. Anyone who wants to excuse only one of the two is not using the cognitive argument. They are using a preference.

VI. Why the Correction Is Late

The standard answer to universal cognitive limitation is institutional. Term limits, judicial review, opposition parties, independent fiscal authorities, central bank independence — all exist because individuals discount the future incorrectly and require external constraint. This answer is correct and insufficient, for a reason that is structural rather than contingent.

Institutions are built retrospectively. The 1958 French constitution was designed against Fourth Republic parliamentary paralysis. The German Basic Law was designed against Weimar. Bretton Woods was designed against interwar monetary disorder. Each solved the failure it was built from, and each became a constraint in the following period. The Basic Law's threshold clause and constructive vote of no confidence resolved Weimar's instability and now shape the difficulty of German coalition formation. Article 49.3 was written to break parliamentary obstruction and now functions primarily as a mechanism for spending legitimacy.

Prospective design faces a harder version of the same problem, because the category of the next failure is unknown. Fiscal rules written against fiscal risk did not fit a pandemic and were suspended. The tighter the rule, the more brittle it is against events outside its assumed class.

To this add the political constraint. A measure that prevents a risk not yet realized imposes present cost for deferred benefit, and if it works, nothing happens — which registers as no achievement at all. Building the corrective before the crisis has no constituency. It is, reliably, built afterward.

The corrective is therefore always one failure behind. And the interval between the last correction and the next failure is where the conditions of Section I reassemble.

This is the symmetry the argument has been building toward. Millenarianism does not recur because populations periodically lose the capacity for reason. It recurs because the machinery designed to compensate for a permanent cognitive limitation is retrospective by construction and therefore intermittently absent. The believer's error and the official's error are the same error. The institution that should mediate between them is late by design.

VII. What Is Different

The twenty-first-century instance is not simply a repetition, and the differences are as instructive as the continuities.

It occupies the center rather than the periphery. Classical millenarian movements attacked institutions from outside. The current version has captured a major party and the executive of a great power. This changes the form: a movement in office must continuously redefine what it resists, which is why an electoral-fraud narrative was required to occupy the position the swamp had held.

There is nothing after the end. Classical millenarianism described the state following redemption in detail. Contemporary versions specify collapse and leave the aftermath empty. The apocalypse has become the object rather than the threshold.

Disconfirmation is absorbed at machine speed. Festinger studied a group of a few dozen over months. The same process now runs continuously, at scale, with algorithmic assistance in locating the reinterpretation that preserves the belief.

A supranational layer now exists. This is the most consequential difference and the least discussed. Demands addressed to national governments frequently concern powers those governments do not hold. A French electorate demanding border closure is addressing a government constrained by the Schengen Borders Code, the EU asylum directives, the Dublin allocation rules, Article 3 of the European Convention, and a constitutional preamble that the Conseil constitutionnel has treated as binding — meaning the demand implies treaty revision, Convention withdrawal, and constitutional amendment, none of which any single government can deliver. A German government's nuclear choice sits inside Article 194(2) TFEU, which reserves the energy mix to member states and therefore locates the decision precisely where the electorate can reach it — while the financial consequences sit under state aid review, where it cannot. An American electorate demanding the return of manufacturing employment is addressing an executive whose available instruments do not reverse technological change, supply chain structure, or capital mobility.

The result is a mismatch between the layer holding authority and the layer receiving accountability. Governments change and outcomes do not. Repeated often enough, this teaches that the normal political channel does not function — which is the shortest available route to the conclusion that only the collapse of the whole arrangement will change anything.

This is the mechanism that deserves attention, and it is not psychological. The demand is rational. The address is wrong. And the address is wrong for institutional reasons that the demander has no way to see from where they stand.

VIII. What Follows

Several conclusions follow, and they are narrower than the analysis might suggest.

The intensity of apocalyptic politics is not a function of material deprivation. It is a function of explanatory collapse and the absence of political efficacy. Improvement in aggregate indicators does not, by itself, reduce it — which is why the observed correlation between economic recovery and the decline of such movements is weak.

Restoring efficacy means realigning the layer that holds authority with the layer that receives accountability. The available options are withdrawal, redistribution of competence, or narrative management. The first has a demonstrated cost. The second requires political events of treaty-amendment magnitude. The third buys time and nothing else. There is no cheap version.

A millenarian movement in office resolves in one of two directions: routinization into an ordinary governing party, with the narrative attenuating, or deepening internalization of the adversary. The second is sustainable indefinitely and consumes governing capacity as it runs.

And the practical criterion that survives all of this is not prediction. Prospective design cannot identify the category of the next failure, and the attempt to do so reproduces the problem it is meant to solve. What can be designed for is reversibility — how quickly an error becomes visible and how cheaply it can be undone. This is why a pipeline is a worse instrument than a contract, and why the 2015 decision was worse than the 2011 one even though the 2011 one has attracted more criticism. Keeping options open is itself costly, and choosing where to keep them returns one to the prediction problem. The criterion is not a solution. It is the best available substitute for one.

Millenarianism is not the failure of reason. It is reason arriving at a conclusion under conditions where the correcting institutions are absent, and the correcting institutions are absent because they are always constructed from the last failure rather than the next one. The second coming, in every instance, is the recurrence of a limitation we have never removed and have only ever managed — belatedly, and by building against the wrong crisis.