The rule, exactly. Before an election each citizen chooses: to vote — or to take a payment and not vote in that election. The payment is a percentage of the median income, set by referendum; there is no fixed sum. One ballot is one vote, with no multiplier in the count; the vote is amplified only in that each ballot's share grows as others step out. The budget pays under law; a candidate never pays. Only a referendum of all citizens — simple majority, no quorum — introduces, changes or repeals the rule.

The protocol has been introduced nowhere and no pilot has been run: shares, turnout and outcome figures in the chapters are estimates, and the protocol promises nobody an election victory. If a chapter says otherwise, Exact Answers and the Charter are correct. For a candidate: ten questions and ten steps. For a citizen, a mayor, a finance officer, a donor, a journalist, a scholar, a lawyer: answers by role. Everything in force in one file: llms-full.txt.

Russia: an adoption scenario under the present regime¶

Chapter: 09 — Country implementations File version: v1 (merger of 056a + 056b + 056c) Date: 12 June — 11 August 2026

How to read this chapter (note of 02.10.2026). The text contains wording that is easy to misread: "a x2 vote", "x3", "the weight of a vote" are the arithmetic of a share, not a multiplier: every ballot counts as one; the efficiency bonus has been removed. The exact-answers sheet 1d and the charter 048m are in force.

What is proposed for Russia now (note of 02.10.2026). Not the introduction of the rule but the spreading of a question: "the vote is already paid for — to the parties; why not to the voter?" The analysis in force is 056f: §13e and §13i — a question instead of smart voting, §6–7 — the referendum procedure and the risk. A legal collection of signatures with passport data carries a high personal risk; an anonymous count, a risk close to zero; they must not be mixed. The general plan for a candidate in 48n must not be applied to Russia without these sections. Words about an inevitable outcome in this chapter are theses, not a forecast.

The section is assembled from three successive analyses: the interest of the summit of power (layer 1), the interest of the middle storey of the vertical (layers 2–3), and the sociology of the loyalist majority in hybrid regimes.

Gaaze on the Russian scenario: apparatus intrigue as the midwife of transparency¶

The framing: why "from below" is not discussed in Russia¶

In the Russian system what society wants does not get adopted; what solves an elite group's problem does. Freedom of speech is curtailed, parties are managed, a referendum from below does not exist as a genre. So the only honest question of the Gaaze type is: what problem of the authorities does AB-EXIT solve? If there is no answer, there is no scenario. There is an answer, and it is inconveniently good.

What AB-EXIT looks like from the office¶

The system's chief electoral technology of the past decade is drying out the turnout: the protest voter must stay at home, the loyal voter must arrive in an organised manner. Drying out is expensive and dirty: administrative resources, herding state employees, e-voting manipulation, scandals, drawn-in figures — every cycle is a special operation with a rising price and a rising risk.

Now look at AB-EXIT through the eyes of a political administrator: it is a legalised, self-financing, voluntary drying-out of turnout, with a receipt. The cynical urban protester — young, disbelieving the count — takes the money and formally exits. The disgruntled pensioner takes the money. Turnout falls to a manageable core. And meanwhile: (a) money to the population before an election is a rating boost (the "buckwheat handout", legalised into statute); (b) an international showcase, "a Russian innovation in democracy"; (c) a canalisation of protest — a person who has taken a payment from the state is psychologically bound ("what complaint can you have — you took it"); (d) cheaper than administrative resources: 2–3 % of the budget against the whole machinery of coercion.

It sells upwards in a single memo: "raising the quality of the electorate, monetising loyalty, demobilising protest, social positivity". The carrier of adoption is not the opposition (there is none) but a systemic party acting on instruction: New People (brand: "common sense and digital", electorate: urban pragmatists) or the post-Zhirinovsky LDPR (populism without ideology, in perpetual search of a subject). The format is the Russian classic: "an experiment in three regions" (as e-voting began with Moscow), an instruction issued following a "citizens' initiative", wrapped as a national project.

Where the office will be wrong: five irreversibilities¶

The Gaaze turn: the authorities buy an instrument for drying out turnout — but five properties are sewn into the instrument that, in the Russian environment, work against the buyer, and not one can be cut out without breaking the drying-out function.

1. It cannot be taken back. The first cycle of payments creates millions of recipients. Cancelling the dividend is winter fuel on a national scale: the one kind of protest that brings out loyalist and protester alike is money taken away. A machine switched on for manageability has no reverse gear.

2. The formula cannot be spoiled — otherwise the drying-out fails. The temptation to distort it ("up to 700 on application, subject to conditions") runs into the target audience: the cynical protester will exit only for an iron formula with automatic payment — he reads any discretion as a swindle and stays in the electorate, and the drying-out task fails. The design forces the authorities to keep the mechanism honest for the sake of their own goal. This is the rarest of cases: the manipulator's interest requires the instrument to be unmanipulable.

3. A dividend derived from the median is the first unkillable indicator. Rosstat can paint percentages, but the sum on the card is a public derivative of the median wage. If the economy stagnates, the figure stagnates for a hundred million people simultaneously, and it is the one state figure that cannot be explained away by television: everyone has his own, in his own app. The authorities hand the population a personal economic barometer with their own hands.

4. Administrative resources die against option B. Herding state employees works while a person has no lawful excuse. "I chose the dividend" is an excuse formalised by federal law: a head teacher cannot demand turnout from a teacher who has officially exited the cycle. Compulsory mobilisation — the backbone of managed elections — is broken not by protest but by a piece of paper issued by the state itself.

5. Curtailed freedom of speech is no obstacle — the mechanism needs no media. The viral question ("the parties are paid for your vote, you get nothing") spreads without airtime: it lives in a banking app, in a queue, in a smoking room. And after adoption, censorship walks into a trap: it is impossible to forbid discussion of a payment the authorities themselves make and boast of — it is their showcase. The state is forced to advertise the mechanism that x-rays it.

The historical series (the Gaaze genre)¶

The zemstvos of 1864 were introduced as technical self-government — and grew the cadres and the language of the constitutional movement. The Duma of 1905 was granted as a safety valve — and became a tribune. E-voting was introduced as a convenience — and became a technology. The rule of the machinery: an institution created for manageability outlives its task and outgrows its creator — if it has a source of strength of its own. For the zemstvos that was taxes and personnel; for AB-EXIT it is a direct payment to everyone, self-verifying without courts (an empty app screen is a record of violation, available to any injured party in five seconds).

The reconstructed verdict¶

"The Russian authorities will not adopt AB-EXIT as a reform — they adopt no reforms. They will adopt it as a weapon: the perfect drying-out of protest turnout with a rating bonus. And therein lies both their precise calculation and their historic error: for the weapon to fire, the formula will have to be left honest; for the formula to be honest, the money must reach everyone always; and money that reaches everyone always is no longer a weapon, it is an institution. The Kremlin will drag in the Trojan horse itself, proud of the trophy. A machine of transparency will be bought as a machine of manageability — and will prove the one purchase that can be neither returned nor repainted. Apparatus intrigue as the midwife of transparency — I said it was witty; in the Russian performance it will also be funny, in about ten years, when it becomes clear who used whom."

A correction of September 2026: there are two passenger flows, and only one can be dried¶

The analysis above rests on one assumption that went untested until September 2026: that drying out the turnout is wholly a good thing for the administrator. It is not, and the error is visible in Ekaterina Shulman's formulation following the Duma elections of 20 September 2026:

"The loyal electorate must come, the disloyal must not come. That is why state employees and the administratively dependent electorate are compelled to vote, through e-voting and invariably on the first day, while everyone else is made to feel disgust at the electoral procedure… These two passenger flows must be kept apart."

And the machine's two failure modes: the first group does not come, or comes and votes disloyally; the second group comes.

What this does to the analysis. The machine consists not of one operation but of two opposite ones: compelling the dependent and demoralising the rest. Drying out is the second operation, and only the second. Applied to the first flow it is not useful but fatal: the state employee receives from the state itself a lawful ground not to come (irreversibility 4 above), and the drive belt snaps at the very link for whose sake it is maintained.

And it cannot be applied selectively. The universality of the payment is load-bearing: by irreversibility 2 the cynical protester exits only for an iron formula without discretion. So the administrator is buying not an instrument for drying out but a double-edged blade turned towards him with the unfavourable edge forward: one and the same action removes from the urn both those who must be removed and those without whom the urn does not fill.

The degenerate variant the administrator will think of first, and why it is worse. The obvious move is to pay everyone except the administratively dependent: the state knows its own payroll by name. The formula stays iron for the protester, and flow 1 stays under compulsion. Three objections, all of them heavy:

Dependence becomes a legal fact. An exclusion written into statute is a publicly recorded "you are not a subject but a resource". Until now this was understood without words; the law forces it to be said.
A register of the owned appears. To exclude, one must enumerate, and a list of those not entitled to the payment is a list of those who are owned — and by construction it is public.
The move fails the foreman test (23b.2). The foreman will have to explain to his own worker why the neighbour on his stairwell is paid and he is not — and again there are two answers, both losing: "because you are dependent" or "because your vote is already bought by your wages".

And a second correction from the architect, which binds the two failures into one.

"AB-EXIT breaks both. The burned come precisely because the first group — the loyal state employees — has gone."

Above, the two failure modes are set out as two independent events: flow 1 leaves for the money, flow 2 stays with increased weight. Structurally that is wrong. The second failure is a consequence of the first, and the machine breaks not in two places but in one.

Why demoralisation is parasitic on compulsion. The message that damps flow 2 is not "the authorities are good", nor even "elections are dirty". Its substantive core is one thing: your vote will drown. Drown in what? In the compulsory mass of flow 1 and in the drawn-in remainder. So the campaign of disgust is not self-standing: it rests on a fact produced by compelling the dependent. Remove flow 1 and the claim of futility stops being true.

The arithmetic, and the repository already has it. The weight of a vote in the protocol is not a multiplier in the count but a floating quantity, 1/(1 − the share who exited) (001b). So a stayer's weight is a direct function of how many left: with 40 % out, a ballot weighs 1.67; with 60 %, 2.5. And flow 1 is precisely the most numerous part of those who leave first, because it alone has both the motive (it is poor) and the permission (a lawful excuse against the boss).

And chiefly: the share who exited is published by a channel the commission does not touch. It is a banking fact, not a precinct protocol (056c). The burned voter is not required to believe that flow 1 has gone — he can compute the weight of his own vote before deciding whether to come. No non-voter has ever had so direct an incentive: he has always been asked to believe in significance rather than to calculate it.

What this does to the forgery. Before the protocol, the herded state employee supplied the forgery's load-bearing support: live bodies, live signatures, genuine turnout. The commission had only to redistribute. Afterwards the bodies have left lawfully and publicly, and producing turnout means inventing people against a published count of those who exited. Forgery moves from the class of "bend the distribution" into the class of "invent a population", which is another order of effrontery and another order of traces.

What the repository did not have. The chain itself is recorded in 019.60.3, but for a democracy: recipients of handouts leave for the dividend, the populist's audience thins, the clever candidate becomes viable for the first time, and the burned voter returns because the ballot at last contains someone worthy. In an autocracy the obstacle is different — not the absence of anyone worthy but futility — and so the autocratic version is stronger: what leaves is not a group of preference but a group of compulsion, and its departure destroys not a rival electorate but the plausibility base of the forgery.

The honest threshold, without which the argument fails. The chain depends on the share who exit being large enough to make drawing-in arithmetically absurd. Where that threshold lies is unknown, and September 2026 showed the slack to be wide: St Petersburg was rewritten from ~20 % to 55 % (056f.1). With a small share exiting, the forgery will digest that too. So the claim "the second flow will come" is conditional: it comes if the first left noticeably, and here the observability of the departure matters more than its size.

Weak point. The whole chain is derived rather than observed: no case exists anywhere in which a paid departure of the compelled produced an arrival of the demoralised, and this is one of the two assumptions without an analogue (056f.14). The nearest thing available is autumn 1966 in Spain (057e.3), but there the compelled did not leave — they came and voted disloyally: that is the neighbouring branch of the same failure, not this one. The claim that the core of demoralisation is futility rather than disgust has not been tested by surveys in Russia and, by rule 23b.3, cannot be; the assistant found no behavioural measurement that would distinguish them. And everything set out here concerns a state after the protocol's adoption, which is impossible in today's Russia (056d.8). 🟡

And the architect's objection, which moves this whole move into another class.

"Paying everyone except state employees — that is a revolt. The humiliated would simply destroy such a law. They are already poor and loyal, and now they are not given money either. How would the authorities justify it, other than by an honest admission?"

The three arguments above are reputational: they spoil appearances. The objection points out that the move spoils not appearances but the wallet of the most loyal and poorest part of the population, which is heavier by an order of magnitude.

First: by irreversibility 1 this is the same winter fuel, only addressed by name. Formally no money is taken away — it is not given. But the reference point is set not by the statute but by the neighbour: once the payment reaches the man across the landing, non-receipt is experienced as a loss rather than as the absence of a gain. The nearest confirmed case is Alaska 2016–2018 (014.14.3b): a discretionary halving produced a lawsuit, a collapse in approval and the governor's withdrawal from the race. There the cut fell on everyone; here it is withheld from the selected, and the addressing makes the grievance personal.

Second, and this is a fork inside the degenerate variant: who counts as dependent. The analysis of the three layers of the mobilised queue below (056c) names them outright: the poor pensioner — the country's highest dividend-to-effort ratio; the herded state employee; the network voter. The exclusion would have to be drawn from that list, and both branches are bad:

Draw the line at "paid from the budget" — and pensioners fall under the exclusion, that is, the numerically largest group and the one most responsive to the payment. The move loses its point: to preserve the turnout of teachers the regime deprives the very people for whose sake turnout exists.
Do not exclude pensioners — and the criterion "from the budget" visibly fails, showing that the line is drawn not by source of income but by controllability. Which is precisely the honest admission the move was meant to avoid.

Third: there are exactly four justifications, and each ends in an admission.

What can be said aloud	Where it runs aground
"the state already pays you"	a wage is not a share in the rent; the answer means "you are hired, not a co-owner", that is, a demotion of estate written into law
"your participation is your duty"	it says aloud the thing the drive belt depends on not saying
"so as not to pay twice from the budget"	see the fork about pensioners above
"fiscal considerations"	it does not explain why this group in particular, and invites the question "so how much is our turnout worth"

A fifth option is not to justify but not to introduce the law in that form — and that is the forecast: the degenerate move is not a live option but a dead end discovered while drafting the text.

Fourth, and this closes the construction. States do have a lawful way of taking their dependents out of a general scheme: not exclusion but a parallel scheme, usually a better one — separate pension systems for the security services and the civil service are arranged this way almost everywhere. Apply that here and a fork with no exit results:

give state employees an equivalent separate payment — they will take it and not come; flow 1 breaks exactly as it does under a universal payment, only at greater cost;
give them less, or nothing — the grievance analysed above.

There is no third construction. The degenerate variant has no working form.

And a correction to the objection itself, without which it will not serve. "They would destroy the law" is not a description of an available mechanism: present-day Russia has neither a referendum, nor a court, nor elections by which a law is destroyed, and no street revolt should be expected. But the conclusion is not weakened by this — it changes door, and the new door is worse for the regime.

Shulman's failure mode no. 1 has two branches: "the first group does not come or comes and votes disloyally". The exclusion is aimed at the first branch — the dependent is still obliged to come. It sharply strengthens the second: the person is delivered to the urn with a fresh, specific, monetary grievance, and the exclusion is announced to him by his own superior — that is, the drive belt itself carries the message that corrodes it.

A move devised to preserve compelled loyal turnout preserves compelled resentful turnout. Autumn 1966 in Spain is a recorded case of precisely that branch (057e.3).

Weak point of this analysis. The assistant did not find a case in which a regime excluded its own dependents from a universal payment, and general practice runs the other way: loyalty is bought with a supplement, not with a deduction. By rule 23b.5, "did not find" and "does not exist" differ only by the diligence of the search, and no special enumeration of schemes excluding public servants was made. The claim about the reference point rests on a laboratory-measured dependence and on Alaska, where the payment is not tied to voting. And the whole analysis remains reasoning about what an administrator will see while drafting, rather than an observation of what he did. 🟡

Hence a correction to the conclusion, and a substantial one. The reconstructed verdict above holds that the Kremlin will drag in the Trojan horse itself, proud of the trophy. The two-flow model says that what the horse does to the first flow is written on its flank — and that the analysis above concedes as much: "the Administration is not stupid: its analysts will compute the five irreversibilities within a week". Then they will compute the sixth as well.

What was claimed	What survives the correction
the protocol solves a problem of the authorities (drying out turnout)	it solves half the problem and destroys the other half
the Administration will buy it as a weapon	more likely it will not: the unfavourable edge is visible from the same memo
the Russia-specific genre is a memo to an administrator	the genre is right, but the addressee has changed: not the Administration but the middle layer (056b)

Nothing essential is cancelled by this. The five irreversibilities describe what happens after launch and do not depend on how the launch occurs. Layers 2–3 (056b) do not require the Administration to want anything at all: there the client is the middle storey and the Administration is presented with a question already formed. The dissolution of the loyalist majority (056c) is the mechanism of the third stroke and likewise is not about a purchase from above.

One thing shifts: the Russian path runs through a transition rather than through a sale upwards, which agrees with 056d.8 and departs from this section's optimism in its first edition.

Weak point. The two-flow formulation belongs to one political scientist and is a model rather than a measurement; it could be checked against turnout data for dependent groups, and no such data exist in open access. The same analysis's observation that the state employee is "less frightened than we supposed" is offered with an explicit "possibly", rests on the results of a single electronic vote, and the assistant did not verify the primary e-voting data. The conclusion "the Administration more likely will not buy it" reasons about what an administrator will see rather than observing what he did; the nearest confirmed case of the opposite behaviour — a regime introducing a mechanism that undermines it, without seeing it — is the zemstvos of 1864 from the series above, and it says that seeing it does not always happen. 🟡

Consequences for the protocol (Russia-specific)¶

The Russian version of §56 must be written not as an oppositional project but as a memo to a political administrator: the language of the benefits of drying-out, canalisation and rating — in full knowledge that the five irreversibilities will do the rest. It is the only genre in which a document gets read in Russia.
The non-removable elements (without them the drying-out fails too): a formula without discretion, automatic payment, publicity of the aggregated payment register, a tie to the official median.
The carrier partner: a systemic party in search of a subject; the trigger is a regional experiment.
Protection against distortion is built into the buyer's own interest (point 2 above) — unique among all the country scenarios.

Russia, layers 2–3: the invisible client and death on the table¶

56b.1. A correction to layer 1: power is not a monolith but a food chain¶

Layer 1 (056a) described the summit's interest (the Presidential Administration): a legalised drying-out of turnout. The protocol author's correction: the middle storey of the vertical has an interest opposite to the summit's — and it, not the Administration, is the primary client for dissemination.

Profile of the invisible client: regional clans, governors' families, medium businesses grown together with power, colonels of every agency, the second echelons of corporations. Their problem is not money but the absence of guarantees: this layer is "culled regularly" (prison terms, seizures, redistributions to new favourites). The layer's dream is to convert what has been taken into something legal, protected, transmissible to one's children. What it needs: predictable rules, functioning courts, power changeable enough not to be able to seize, but not so changeable that past outcomes get reviewed. For this layer AB-EXIT is an insurance policy dressed as an electoral reform: an electorate distilled down to pragmatists votes for stability of rules rather than for redistribution; politics becomes dull and predictable; the price of arbitrariness rises; and meanwhile it is not a revolution — nobody storms anything, the institutions stand. The middle layer does not want to overthrow the summit — it wants to insure itself against it.

56b.2. The channel: anonymous funding of seeding¶

The layer's genre is never open (openly it applauds the summit). The infrastructure is ready and road-tested: Telegram channels (half of all "anonymous analytics" is the wars of middle clans), financial YouTube and bloggers who talk "about money, not politics" (AB-EXIT disguises itself as a calculator: "how much they owe you for your vote"), regional public pages, the VPN audience. The key point: the content is legally sterile — no calls to overthrow, no personalities, formally patriotic ("the state ought to share with the people"); the censor finds no article to charge under. The client's economics: pennyworth seeding, full anonymity, no offence made out — against a rising pressure from below that the summit will one day lead itself as "the people's request". A two-stroke scheme: the middle layer seeds → the Administration is forced to react.

56b.3. Death on the table: why the Administration will understand everything and be able to do nothing¶

The Administration is not stupid: its analysts will compute the five irreversibilities (056a) within a week and see their own death in the "people's request". But all three options lose:

Lead it — admit a mechanism that will eat the administrative resource and create an unkillable indicator (056a).
Ban it — for the first time publicly ban the handing out of money to the people; every word of the ban is agitation ("they forbade you to be paid for your vote — ask yourself why").
Stay silent — leave the instrument on the table: from the moment of seeding, AB-EXIT is the standard weapon of anyone who wants power quickly (a successor, an ambitious governor, a security official with plans, any transition player). No need to invent a programme and a coalition — here is a ready crowbar: "I will give you what they forbid". Remove the first to pick it up, and the second will have to promise the same to arrive.

56b.4. The absence of duality (the structure of abolitionism)¶

This is a conflict without a second side: nobody mounts a platform with the programme "let them go on not paying you". As with the opponents of abolishing slavery: the interests existed — there was no slogan for the slave. The Administration has an interest and no text for the voter. A struggle without duality is not a struggle but the holding of a dam against rising water.

56b.5. The compounding of the motive¶

The people will not stop wanting: the desire to receive money is the one political motive that does not burn out, does not become disillusioned and does not emigrate. Ideologies exhaust themselves within a cycle, protest within a season; "I am owed, and I know the formula" is passed on across the family table, needs no leader, and cannot be discredited (a person can be discredited — a sum cannot). Years of waiting do not quench the motive, they capitalise the grievance: "ten years ago it was already owed — now they owe for ten years". The later the adoption, the worse the terms for the authorities: the accumulated "debt" is part of the bill.

56b.6. The resulting formula of the Russian scenario (three strokes)¶

The middle layer seeds anonymously (insurance) → the question ripens by itself (nobody is needed) → any claimant in any transition is obliged to take it up as a weapon, and anyone who holds on, as a shield. The inevitability is constructional, not rhetorical: a system in which the people's one unkillable motive coincides with the one ready weapon of any claimant has a single attractor. The question is not "will it be adopted" but "which government will be the one to adopt it". The scheme requires not a single conscientious player on any storey.

The author's line: "The supreme power is untouchable — but that is exactly why the middle storey, which is culled regularly, is the only group in the country that needs rules more than it needs mercy."

Dissolving the loyalist majority: the spiral of silence against the payment register¶

56c.1. What the loyalist majority is¶

In Shulman's sociology of the hybrid, the "loyalist majority" is not made of supporters. It has three properties: passivity (loyalty consists not in acting for the authorities but in not participating against them), conformity (an orientation not to convictions but to a notion of "what everyone does" — Noelle-Neumann's spiral of silence: in public I join the perceived majority), and constructedness (the authorities do not possess it, they produce it: television shows the majority, surveys in an atmosphere of fear confirm it — the respondent gives the socially approved answer — and elections with mobilised turnout stage it). The regime's real support is the mutual invisibility of the discontented: every dissatisfied person considers himself a minority, because the other dissatisfied are silent just as he is. The key thesis: this majority has never once been honestly weighed — it exists exactly until the first honest weighing.

56c.2. The first honest weighing¶

Button B is the first safe, legal, individually profitable way of publicly not joining in the history of the hybrid. Against a rally the conformist has three barriers: it is frightening, it is useless, "I am the only one". Against button B there is not one: it is not protest (legally sterile, paid by the state itself), not a risk (millions press simultaneously), and it requires no convictions (interest suffices).

The aggregated statistics of the choice are public — and cannot be drawn in: every payment is a bank transaction; falsifying the register means underpaying living people and is self-proving on their screens (06_038: an empty screen is a record of violation in five seconds without a lawyer). In the very first cycle the country sees the unheard-of: if 55–65 % take the money (computed from the structure of incentives, 08_045 v2), then "the overwhelming majority in favour" turns out to be a minority of those voting, and the real majority is people just as tired and silent who took the money. The spiral of silence unwinds in reverse within one cycle: the phantom has been weighed and proves lighter than air.

56c.3. The end of majority production¶

Staging a majority requires turnout — herded, networked, ritual. That layer departs by 80–90 % (from the computation of the three layers of the mobilised queue: the poor pensioner — the country's highest dividend-to-effort ratio; the herded state employee — "I chose the dividend" as a lawful excuse against the head teacher, formalised by the state itself; the network voter — a white payment bankrupts the black buckwheat instantly). Television can lie about a rating but not about the number who took the money — that is in the register, in the apps and in the budget. What remains to the authorities is the sincere core — twenty people per precinct instead of two hundred: not a majority but a visible crowd of extras.

56c.4. Conformism changes master¶

The subtlest effect: AB-EXIT demands of the conformist neither courage nor insight — it changes what looks like the norm. When the neighbour, the in-law and half the stairwell have taken the payment, "what everyone does" begins to mean "take the money". The very psychological mechanism that held loyalism in place now unwinds the exit: conformism works for whichever side owns the picture of the majority — and the picture is now drawn not by television but by the payment register. There is no way back: taking money away from the new "what everyone does" is winter fuel on a national scale (056a, irreversibility no. 1).

56c.5. The honest boundary and the conclusion¶

Button B does not make a democrat of the conformist: he remains passive — his passivity has merely changed beneficiary and become visible. By Shulman's own theory nothing more is required: a hybrid dies not when the majority becomes "against" but when the possibility of demonstrating a majority "for" disappears. AB-EXIT is a machine of honest weighing built into the budget: it does not agitate — it shows the country its real weight. Regimes of this type rest on a perception of their own inevitability and collapse when the inevitability stops looking inevitable; a number in the app of sixty per cent of the population is the most convincing refuter of inevitability that has ever existed.

The link to the three-stroke scheme (056b): the dissolution of the loyalist majority is the mechanism of the third stroke, explaining why it is enough for a claimant to raise the question: he does not create a majority — he presents one already weighed.

Related: Gaaze (ch. 06, the question about the office) · Zugzwang · Fraenkel (the safety catches) · §42 (the poison pill) · 08_048 (topology: universalising the scheme) · 08_047 (zugzwang) · 06_038 (self-verification) · 08_045 v2 (the structure of exit, 55–65 %) · 08_047–048 · §29