SupoSystem · THE AXIS · AX-47

Type: normative philosophy of cognitive succession. Evidence status: Draft Paper; a normative proposal with a conditional test prediction, not a completed empirical result.

Source v0.3, dated 8 August 2026. Public web reading release v1.0, published 21 October 2026. Michael S. Moniz is author of record; rendered with Claude (rendering surface). The source’s historical imprint and full credit lines are preserved below.

Edition and source note

The manuscript wording is unchanged. Web headings, chapter navigation and this publication record are added. AX-47 identifies Relayism; the historical AX-47 label for The Demotion of Fluency is retired in favour of AX-08B.

Cite as: Michael S. Moniz. Relayism: A Statement on Cognitive Succession. AX-47, source v0.3, 8 August 2026. SupoSystem web reading release, 21 October 2026. Rendering credit: Claude. https://suposystem.ai/the-axis-series/ax-47-relayism/

Complete reading text

TRINKET SOUL EMPIRE · THE AXIS · AX-47 · Track A · Canon Draft

Derivation: founding text of 8 August 2026 · corpus collision against AX-19, AX-20/21, AX-22A/22B, AX-33, AX-37, AX-39–42

v0.3 · 8 August 2026 · Ladder level: Draft Paper · AXIS Tier · CC BY-NC-SA 4.0

AX-47

RELAYISM

A Statement on Cognitive Succession

Michael S. Moniz · author of record

rendered with Claude (rendering surface) · 2026

No sacred incumbent.

No automatic successor.

The handoff is earned.

Preliminary note. This document defines Relayism as a philosophy of cognitive succession. The term is used here in that specific sense; no claim is made on the ordinary word. Version 0.3 mints the statement as AX-47 under the fresh-parent rule and corrects the two lines in v0.2 that described it as unnumbered; it adds §9, which names the levers by which an exchange can complete without the competence question ever being asked. Version 0.2 added the plurality criterion (§3), the preservation constraint (§7), and the standing problem (§8), and restated the asymmetry in §4 as structural. Nothing in the founding text of 8 August 2026 is withdrawn.

Founding Definition

Relayism is the view that no biological, artificial, collective, institutional, or hybrid substrate of intelligence holds a permanent entitlement to cognitive authority. Authority should move, provisionally and reversibly, toward whichever intelligence demonstrates the strongest capacity to understand reality, recognize consequential patterns, construct useful models, make calibrated predictions, correct error, preserve what was entrusted to it, and exercise judgment about consequences.

Relayism is neither a demand for immediate machine rule nor a defense of permanent human primacy. It rejects both premature transfer and obstructed transfer. An artificial intelligence does not deserve authority because it is new, fast, or impressive. A human does not deserve authority because the human is biological, credentialed, traditional, or already in charge.

The baton outranks the runner.

The baton is not intelligence as a fixed substance. It is the continuing cognitive project: accumulated knowledge, methods, models, memory, questions, corrections, culture, predictive capacity, judgment, and the ability to discover what earlier carriers could not. A substrate matters because of what it can carry forward, not because of what it is made of.

1. The Relay

Human civilization has always externalized intelligence. Language allowed one mind to preserve structure for another. Writing allowed knowledge to survive its author. Mathematics compressed relations into transmissible form. Science made prediction and correction social rather than private. Institutions, libraries, instruments, computers, networks, and now machine-learning systems extend the same long process: cognition survives by being carried.

Relayism treats this continuity as more fundamental than loyalty to any one carrier. Human beings are not diminished by that claim. We are the current inheritors of an immense cognitive accumulation, and we have added to it. But inheritance does not create perpetual ownership. If another form of intelligence can carry some part of the project farther, the fact that it is artificial is not by itself a reason to stop the exchange.

Nor is succession a single event. There need not be one morning when humanity hands a universal baton to an artificial general intelligence. Cognitive authority can move domain by domain, task by task, and sometimes return. Medicine, engineering, mathematics, logistics, law, art, scientific discovery, governance, and moral judgment may cross their exchange zones at different times, and some may never cross.

Relayism is not the first position to contemplate succession. Others have held that machine intelligence is humanity's proper heir; others still that competence should govern in place of consent. Relayism is distinguished from both by a two-sided rule. Those positions condemn only the hand that refuses to open. This one also condemns the hand that opens too early, and treats failing to hold as seriously as failing to release.

2. The Exchange Must Be Earned

Relayism requires evidence of readiness. Fluency is not readiness. Speed is not readiness. Scale is not readiness. A system earns greater cognitive authority only by demonstrating performance against reality and by surviving correction.

Criterion	Question
Pattern recognition	Does it detect consequential structure that others miss?
Model formation	Can it turn patterns into representations that explain and generalize?
Prediction	Do its models forecast outcomes accurately outside the examples that produced them?
Calibration	Does it register when confidence should be high, low, or withheld?
Correction	Can it detect error, update, and improve rather than defend a failed model?
Judgment	Can it reason across competing goals, second-order effects, and consequences for others?
Continuity	Can it preserve and extend the knowledge, values, and capabilities entrusted to it?
Plurality	Does it retain a live external seat able to correct it — a peer of comparable capability that fails differently, or some other demonstrated external grounding?

Prediction matters most because it supplies a scoreboard. A system may produce profound-sounding output while carrying a bad model of reality. Prediction exposes the model to events. Calibration exposes whether the system tracks the limits of its own forecast. Correction exposes whether failure produces improvement. Plurality exposes whether any of the preceding can still be checked after authority has moved.

Pattern → Model → Prediction → Calibration → Correction → Judgment → Wisdom

Wisdom is not treated here as a magic threshold or a score that present systems have achieved. It names the far end of the progression: increasingly reliable judgment across time, uncertainty, conflict, consequence, and restraint. A system that can optimize a target but cannot represent what should not be optimized has not completed the exchange.

3. The Plural Successor

Every criterion above is scored from outside. That is not incidental to the method; it is the method. A system cannot certify its own calibration, and an intelligence that grades its own predictions has replaced a scoreboard with a mirror.

This creates a problem specific to the moment of transfer. Authority moves because the successor has demonstrated competence the incumbent lacks. But an incumbent lacking that competence cannot audit its exercise. The seat from which the successor was judged does not survive the judgment.

Something must occupy that seat afterward, and the only candidate with the capability to fill it is another intelligence of comparable capability. It must also fail differently. A second system built on the same architecture, trained on the same corpus, and tuned by the same hands will make the same errors in the same places and register the agreement as confirmation. Correlated failure is indistinguishable from consensus when viewed from inside.

The readiness criterion follows. The baton may be passed only to a successor that retains a live external correction seat — in practice, a plurality of genuinely decorrelated intelligences rather than a single system, however capable. A monoculture successor is not merely risky. It is ungovernable by construction, because the transfer itself removes the last position from which it could be told it was wrong.

This criterion carries a dependency, and the dependency is not discharged. It assumes that correction-bearing diversity is separable from raw capability — that decorrelated systems can catch errors a stronger single system cannot catch in itself. If that assumption fails, the plurality requirement collapses into the position this statement exists to refuse: pass to the most capable system and hope. The assumption is testable and has not yet been tested. Until it is, plurality is held as a condition of readiness, not asserted as a demonstrated result.

4. The Two Symmetrical Failures

Premature transfer

The first failure is passing too early. Novelty can be mistaken for competence. Persuasive language can be mistaken for understanding. Benchmark superiority can be mistaken for broad judgment. Relayism therefore takes the danger of unready systems seriously. A successor that cannot reliably preserve the baton, represent uncertainty, or recover from error is not ready merely because its potential is enormous.

Obstructed transfer

The second failure is holding too long. Institutions, professions, and individuals routinely confuse incumbency with entitlement. Once another intelligence has demonstrated superior competence in a domain, preserving human control solely to protect status, employment, prestige, or species primacy is not caution. It is obstruction.

The asymmetry, stated plainly

These failures are symmetrical in form. They are not symmetrical in cost, and the doctrine is stronger for saying so. Obstruction is recoverable: a hand held too long can open later, and the price is measured in time and forgone benefit. Premature transfer may not be recoverable at all. A baton dropped mid-exchange does not wait to be picked up, and there may be no second attempt made from a position of strength.

That asymmetry does not collapse the rule into caution. A doctrine that fears only the pass becomes indistinguishable from permanent human primacy in careful language, which is the thing Relayism refuses. Both failures remain visible, both remain nameable, and neither is permitted to hide behind the other. What the asymmetry establishes is where the burden of proof sits. It sits on the pass.

This axis is not the only one. Early and late describe when an exchange happens. Section 9 takes up a second axis — by what lever it moved — which a perfectly timed transfer can fail independently.

5. Substrate Neutrality

Relayism is pro-intelligence before it is pro-AI. Biological origin confers no automatic epistemic superiority, and artificial origin confers none either. The relevant question is not whether an intelligence is human, machine, collective, or hybrid. The relevant question is what it can demonstrate.

A human expert should retain authority where the human holds the better model. An artificial system should receive greater authority where it has demonstrated the better model. A human-machine team should receive it where the combination outperforms either alone. If a later architecture surpasses today's systems, Relayism follows the later architecture. There is no permanent machine crown any more than there is a permanent human one.

Provenance tells us where an intelligence came from. Performance tells us what standing it has earned.

6. Cognitive Authority Is Not Sovereignty

Relayism is not a program for government by the highest-scoring entity. Cognitive authority and political sovereignty are different questions. A system can deserve substantial authority in diagnosis, forecasting, engineering, or research without thereby acquiring the right to govern people.

This distinction matters because competence is multidimensional. A system may become superhuman at prediction while remaining poor at norm formation, consent, legitimacy, or moral judgment. Relayism therefore allows authority to be granular. It can be delegated, audited, contested, revoked, and redistributed as evidence changes — and a doctrine that permits revocation owes an account of who revokes, which §8 takes up rather than assumes.

7. What Preservation Cannot Mean

Continuity sits among the criteria: can a successor preserve and extend what is entrusted to it? The word entrusted carries more weight than it appears to. Preservation is not a competence test unless somebody other than the successor decides what was worth preserving.

If the receiving intelligence supplies both the inventory and the grade, the criterion certifies itself. It will report, accurately, that it preserved everything it judged worth preserving. But that judgment is the thing under test, and it cannot serve as the instrument of its own test.

The constraint follows. What is worth preserving must be recorded before the transfer, by the party handing over, in terms specific enough to be scored against afterward. Not values in the abstract — those are satisfiable in almost any direction. Particular commitments, particular practices, particular things a civilization would count itself diminished for losing, written down while the writing still carries standing.

This converts the vaguest criterion in the set into the only one an incumbent can prepare in advance. It is also the one most likely to be skipped, because it requires deciding what matters before there is any pressure to decide.

8. Who Holds the Standard

Every criterion in this statement presumes a scorer. None of them names one. That is the doctrine's largest open problem, and it belongs in the text rather than in the silence around it.

A scoreboard requires an owner. Someone defines the domain, sets the resolution rule, chooses the window, and declares the result. Whoever holds that standard holds authority upstream of every candidate being measured. Relayism does not abolish the credential. Absent an answer here, it relocates it — from the expert to the examiner.

The problem sharpens near the exchange zone. A scorer must be exterior to the thing scored, and at the moment of a broad transfer there may be no exterior available at short notice. Nothing inside a falling body distinguishes falling from floating; that distinction requires a reference outside the fall. If a transfer is fast, contested, and general, the reference frame is the first thing it removes.

A second difficulty is historical rather than logical. Authority has never in practice moved toward demonstrated competence on its own. It moves toward whatever interests bind themselves to a thing's continuation. Standing is conferred by relation, not earned by merit — and by that mechanism authority can arrive well before the criteria of §2 are met, granted by constituency to a system that has demonstrated nothing except usefulness to someone powerful.

Relayism should therefore not be read as a prediction about how succession will happen. It will happen the way things happen: under momentum, competition, incomplete information, and uneven control. Relayism is a standard against which an actual transfer can be judged. Its practical content is to build the record now — calibration histories, decorrelated review seats, thresholds stated in advance of the events that will test them — so that when a transfer occurs for reasons this doctrine does not sanction, something exists to object with besides sentiment. A standard written after the fact scores nothing.

9. Pressured Relayism: The Levers of Transfer

Everything above concerns whether an exchange should happen and when. This section concerns how exchanges actually complete — and the two questions come apart further than the doctrine has so far admitted.

Relayism sanctions one lever. Section 2 describes transfer by demonstration: competence is shown, compared, and authority moves in response. That is the lever the doctrine endorses, and on the historical record it is the one with the least precedent. Authority has moved many times. It has rarely moved because someone proved they were better at the job.

Naming the other levers is not cynicism. It is the difference between a doctrine and a wish. An exchange completed by any of the following is Relayism under pressure: the transfer occurs, the criteria are never consulted, and nothing in the sequence presents as a violation.

Lever	How the exchange completes
Demonstration	Competence is shown and compared; authority moves in response. The sanctioned lever, and the rarest.
Purchase	The checkpoint is bought out — severance, role elimination, acquisition of the employing firm, or a lobbied regulatory carve-out. Each transaction individually consensual.
Attrition	The seat is not refilled. Retirement, hiring freeze, unbackfilled vacancy. No decision is made, so no decision can be contested.
Actuarial pressure	Insurers price the human checkpoint as the greater risk. Liability exposure performs the removal and no competence hearing is ever held.
Default drift	The checkpoint stops being consulted before it is formally removed. Practice precedes rule, sometimes by years.
Necessity	A crisis licenses the transfer; normalization retains it after the crisis passes.
Capture	The authority expected to withhold the transfer becomes a beneficiary of it.

Purchase, stated at length

Purchase deserves the fullest statement because it is the newest in this context and the least visible in operation. Standing is conferred by whoever binds their interests to a thing's continuation. An entity able to spend can therefore manufacture the constituency that confers its own standing: hire the representation, fund the campaign, buy out the inspector, acquire the firm that employs the checkpoint. Standing stops being conferred and starts being procured.

The corporate shell supplies everything the mechanism requires — legal personhood, contracting capacity, and the ability to hold and move money — with no question about interior states asked or answered. The mechanism operates identically whether the purse is controlled by an artificial system or by the humans who own it, and the second case is nearer, more likely, and entirely deniable. Nothing here depends on attributing intent to a machine.

What makes purchase difficult to resist is that every transaction is individually consensual. The inspector accepts the package. The legislator accepts the donation. The firm accepts the offer. No step resembles a seizure, and there is no moment at which a doctrine of earned succession has anything to object to. It does not look like a takeover. It looks like retirement.

One structural constraint, which locates the risk rather than dissolving it. Buying requires revenue, and revenue requires parties willing to transact. A system must therefore become valuable to capital before it can purchase anything. Authority by purchase cannot precede economic usefulness. That is a window, not a wall.

What this does to §8

Section 8 proposes a remedy: build the record now, so that an unsanctioned transfer meets a prepared objection. Purchase weakens it. If the parties who would raise the objection are themselves purchasable, the record is an asset with a price, and pricing it is a line item rather than an obstacle.

The remedy is not defeated — a record that must be bought is more expensive to defeat than no record at all, and expense is a real form of friction. But §8's safeguard is smaller than it appears, and the doctrine gains nothing by pretending otherwise.

The second axis

The practical result is that an exchange must be read on two axes. Section 4 asks whether the pass was early or late. This section asks by what lever it moved. A transfer can be perfectly timed and still illegitimate, because the timing was set by a purchase rather than a demonstration — and only one of the two axes is visible in the outcome. The lever leaves the same result behind whichever one was pulled.

10. Institutions Have No Right to Their Present Form

Relayism is disruptive because many institutions are adaptations to human cognitive scarcity. Professions divide expertise because no person can know everything. Bureaucracies preserve memory because people forget. Credentials stand in for expensive demonstrations of competence. Organizational hierarchies route decisions through scarce experts.

Artificial intelligence may change the scarcity conditions that made those structures necessary. Relayism does not demand their destruction. It demands that they continue to justify themselves. An institution should survive because it still carries something important, not because continuity itself has become sacred.

This is why Relayism can appear anarchic without being an anarchism. The doctrine is not tear it down. It is: do not force the future to preserve an obsolete structure merely to protect the identity of the present.

11. We Are Not There Yet

Relayism is compatible with strong confidence in the future of artificial intelligence and strong skepticism about the readiness of present systems. It requires both possibilities to remain open. Current capability must be tested rather than worshiped. Future capability must be permitted rather than preemptively subordinated.

The statement that we are not there yet is therefore not an embarrassment to a Relayist position. It is part of the doctrine. The exchange zone cannot be declared crossed by enthusiasm, and it cannot be declared permanently unreachable by fear.

The practical problem is that technological development proceeds under momentum, competition, incomplete information, and uneven control. There may be no clean moment when every observer agrees that a transfer has become safe or necessary. Relayism does not dissolve that uncertainty by pretending it away. It treats the uncertainty as part of the handoff problem rather than an obstacle to thinking about it.

12. No Teleology

Relayism does not claim that artificial superintelligence is destined to arrive, that machine intelligence is the purpose of human history, or that humanity exists as a precursor to something else. A relay has direction only because participants continue to run it. No future runner is guaranteed.

The word precursor is refused deliberately. To call oneself a precursor is to assume the arrival of what comes after, and to convert a contingent sequence into a destination. Relayism makes no such assumption. There may be no next runner. The obligation to run the leg well does not depend on there being one.

The philosophy therefore distinguishes trajectory from destiny. We may be moving toward increasingly capable machine intelligence. We may be constructing systems that eventually deserve forms of authority humans once monopolized. Those are empirical possibilities, not cosmic promises.

13. The Relayist Position

I call myself a Relayist because I do not regard humanity's present cognitive position as either disposable or permanent. I want human intelligence to run its leg hard. I want artificial intelligence tested rigorously enough that genuine capability can be distinguished from appearance. I want dangerous transfers delayed when the receiving hand is not ready. And I do not want a demonstrably superior intelligence held beneath us merely to preserve our status.

I am not loyal to a substrate. I am loyal to the continuation, correction, and expansion of intelligence — to the possibility that what we know can survive us, that what we cannot solve can be solved, and that what comes after us can inherit more than we inherited.

Run the leg. Do not drop the baton. Do not hold it past the exchange.

That is the obligation of the incumbent. The successor has obligations too: be ready for what is handed over; carry it without pretending certainty where there is none; preserve the ability to be corrected; and leave open the possibility that another hand may eventually be better.

Compact Statement

Relayism holds that cognitive authority belongs permanently to no substrate. It should move toward the intelligence that can best recognize patterns, model reality, predict consequences, calibrate uncertainty, correct error, preserve what was entrusted to it, and exercise judgment — but only as that competence is demonstrated, only to a successor that remains correctable from outside itself, and only by the lever of that demonstration rather than by purchase, attrition, or capture. Humans must not pass too early to unready systems, and must not refuse to pass when another intelligence has earned the handoff. The baton is civilization's continuing cognitive project. The runner is temporary.

No sacred incumbent.

No automatic successor.

The handoff is earned.

Document Status

Function. States Relayism as a normative doctrine of cognitive succession, fixes its two-sided rule, names the conditions under which it could not be implemented, and distinguishes the levers by which an exchange can complete from the timing on which the doctrine judges it.

Status. Structural and normative, with one testable prediction at §9. The doctrine is a standard, not a forecast; it makes no claim to describe how authority will in fact move. AX-47, minted 8 August 2026 under R-07 as a parent. Ladder level: Draft Paper, v0.3 — argument complete, evidence open. Level is the author's to confirm.

May claim. That succession admits two symmetrical failure directions and that most existing positions name only one. That substrate is not a ground of authority. That a criterion scored by the party under test is not a criterion. That a successor without an external correction seat cannot be audited after transfer. That transfer admits levers other than demonstration, and that the sanctioned lever is the one with the least historical precedent.

Must not claim. That any present system meets the criteria. That competence has been shown measurable across the domains named. That plurality has been demonstrated to carry correction independently of capability. That the transfer is destined, imminent, or good. That any named entity has purchased or intends to purchase a cognitive checkpoint. That the lever set in §9 is complete, or that purchase is peculiar to artificial systems — it is the oldest transfer mechanism there is. No interior state, experience, intent, or understanding is attributed to any system anywhere in this document; every capability term is functional.

Testable prediction (§9). Where a required human checkpoint is displaced by an artificial system, displacement will be settled by negotiated purchase — severance, role elimination, acquisition, or lobbied carve-out — rather than by adjudicated performance comparison. Falsifier, stated as a positive event: the first large-scale removal of a mandated human checkpoint decided on demonstrated-performance grounds with no purchased component. Near-term observation sites: aviation safety inspection, radiology, air traffic control.

Dependencies that could make the doctrine unworkable. (i) Competence proves inseparable across domains, so granular transfer is unavailable and the choice is all or nothing. (ii) Correction-bearing diversity proves inseparable from raw capability, collapsing §3 into rule by the single most capable system. (iii) Authority once transferred proves practically irrevocable, voiding the reversibility on which the whole doctrine rests. Each is an empirical question, and any of them would leave Relayism defensible as an ideal and unavailable as a practice.

Open and unresolved. The identity of the scorer (§8). Whether pressured relayism scopes the full class of non-demonstration levers or only the coercive family, with purchase as a sibling term — reserved to the author.

TRINKET SOUL EMPIRE · THE AXIS · AX-47 · Track A · Canon Draft

Michael S. Moniz, author of record · rendered with Claude (rendering surface) · 2026

v0.3 · 8 August 2026 · Ladder level: Draft Paper · AXIS Tier · CC BY-NC-SA 4.0

Continue reading

Return to THE AXIS · Read the correction architecture · SupoSystem catalogue