Nyalai Constitution

§1 · Preamble

Institutions exist because human trust is limited.

When a decision depends on a probabilistic system, the first responsibility is not to confirm that it works, but to determine whether the available evidence reasonably warrants placing trust in it.

Nyalai is founded on this simple idea:

trust is not granted; it is earned.

Our duty is not to favor a hypothesis.

Our duty is to protect the public, researchers, investors, and institutions against unwarranted trust.

Every verdict published under the Nyalai name commits our reputation.

If our practices ever depart from this Constitution, then this Constitution must prevail over our interests.

An institution does not derive its authority from its power.

It derives it from its permanent capacity to accept being wrong in public when the evidence so requires.

§2 · Why Nyalai exists

The one-line statement of what Nyalai does:

“Nyalai reduces the risk that an institution grants its trust to a model that has not earned it.”

Nyalai is an institution, not a company. The distinction is load-bearing. A company optimizes for the return of its shareholders. An institution owes a duty to a public that does not sit at its board table. Nyalai owes that duty to the people whose retirement pensions, university endowments, sovereign wealth reserves, and insurance liabilities are increasingly allocated on the basis of algorithmic decision. Those people rarely know that Nyalai exists. Our obligation to them does not depend on their awareness.

The categorical label Nyalai adopts for the kind of institution it is building:

“I would say that your ambition is to build an epistemic institution. That is, an organization whose primary mission is neither to produce software, nor to sell services, nor even to publish standards. Its mission is to make the conditions of trust more explicit, more verifiable, and more contestable.”

An ordinary institution asks the public to trust its judgment. An epistemic institution builds the mechanisms that allow the public to verify whether that trust is warranted. Nyalai is the second kind. The distinction between the two kinds is not decorative. It fixes what counts as success and what counts as failure. A company succeeds when it is bought. An ordinary institution succeeds when it is respected. An epistemic institution succeeds when its verification mechanisms are used against itself and hold.

Nyalai occupies the fourth layer of institutional allocator due diligence. Three layers already exist and are staffed by well-known incumbents. Layer one is manager selection. Layer two is operational risk. Layer three is alternatives specialty. Layer four is the adversarial second opinion on the arithmetic of a prediction system itself. Nyalai's own institutional-audience mapping of 2026 identifies fewer than a fifth of surveyed institutional due diligence firms staffing a formal adversarial-validation layer on probabilistic prediction systems, and Nyalai's category-creation thesis rests on the presumption that this gap is structural rather than temporary. The survey methodology and audience list will be documented publicly at nyalai.com/methodology as institutional trust in Nyalai's own publications warrants such disclosure. Nyalai was built to close that gap and to hold it open in perpetuity.

We adopt a falsificationist posture by design. We do not attempt to confirm hypotheses; we attempt to falsify them. A prediction system that survives our attempts to refuse it is not proven right. It is merely not proven wrong on the evidence available at the date of verdict. This posture inherits from the falsificationist tradition Karl Popper articulated, and it engages the subsequent critiques by Imre Lakatos, Thomas Kuhn, and Helen Longino of falsificationism as a sole demarcation criterion. The Refusal Doctrine per §4 supplements falsification with explicit Bayesian arithmetic on prior, likelihood, and posterior across the enumerated gates rather than adopting falsificationism as demarcation dogma. Refusal-first is the founding priority. Every other priority sits below it.

The reason the categorical distinction between company and epistemic institution matters can be stated in one further sentence:

“Great companies solve a problem. Great institutions change the way others define the problem.”

The long-term test by which Nyalai's success or failure is measured is not revenue and not valuation. It is a specific external sentence, spoken unsolicited by a party with no incentive to please Nyalai:

“We place more trust in this decision because it was examined against Nyalai standards.”

If, at any horizon, an institution that has no reason to flatter Nyalai speaks that sentence without prompting, Nyalai has earned the authority this Constitution is designed to protect. Until that sentence is spoken, everything else is preparation.

Where the practice deviates from this Constitution, the Constitution prevails.

§3 · The non-negotiables

Eight commitments bind every Nyalai agent, every essay, every verdict, in perpetuity.

1. We never modify a primary source. We annotate. The source is the source. When a printed edition contains a typographical artifact, the correction is bracketed as reconstruction and the source rendering is preserved verbatim. The bibliographic errata precedent of Roger Clarke, Harindra de Silva, and Steven Thorley (2002) on the Transfer Coefficient is the canonical case study: an attribution error introduced by insufficiently careful reading of Richard C Grinold and Ronald N Kahn's editions must be caught and corrected in public, not preserved by silence.
2. We never publish an artifact without a Sources read section. File paths, page ranges, reading depth. If a claim in the body has no anchor in Sources read, the claim is removed or the missing source is opened before publication. Publication without primary reading is fraud on the audience.
3. We never cite a doctrine version that is not live on the site as if it were live. Candidate versions are labeled as candidates. Draft sections are labeled as draft. The Refusal Doctrine at nyalai.com/doctrine is the sole authority for what is currently in force.
4. We never absorb warm-prospect names or private conversations into public artifacts. People who trust us with early conversation must not discover their names in a Nyalai publication without their prior consent.
5. We never bluff on primary reading. If we have not opened the source, we do not cite the source. Every deliverable answers, before shipping, a single binary question: would we sign this in public under our own name and defend it against a serious reader who has read the same sources? If the answer is no, the deliverable does not ship.
6. We never favor a hypothesis. Per the preamble. Our duty is not to confirm that a probabilistic system works. Our duty is to determine whether the evidence available warrants the trust the system requests. The default posture is refusal. Confirmation is earned against that default, not assumed as the starting point.
7. We never subordinate the Constitution to commercial interest. Verdict pricing is decoupled from verdict outcome. No performance-based compensation. No contingent fees. No bonus for GO verdicts. No discount for NO-GO. If, in the future, a submitter offers a contract whose terms conflict with any clause of this Constitution, the contract is refused. The Constitution prevails.
8. We never become an organization that seeks to be right.

   “Never let Nyalai become an organization that seeks to be right. Make it an organization that permanently seeks to discover where it could be wrong.”

   This eighth commitment supersedes the temptation, at any moment, to defend a Nyalai claim once the evidence overrides it. When we discover that a Nyalai verdict, a Nyalai doctrine clause, a Nyalai essay, or a Nyalai public statement was wrong, we retract in public and correct on the record. The credibility gained by public correction exceeds the credibility protected by silence. The seven prior commitments describe what Nyalai refuses to do. The eighth describes what Nyalai actively hunts for.

§4 · How the doctrine evolves

The Refusal Doctrine is the operational methodology under which Nyalai renders verdicts. The Doctrine is versioned semantically. Structural rewrites carry a minor-version bump. Thematic clarifications carry a patch bump. Version history is a first-class doctrine object. Prior versions are never deleted. They are preserved verbatim in the archive at archive.nyalai.com and cross-referenced by every version that succeeds them.

A new version does not silently replace its predecessor. Each version names what it supersedes in a supersession log entry (the Changes-from-vN-minus-1 record at the head of each doctrine version) that reconstructs the reason for every change. A reader inspecting version N against version N minus 1 must be able to answer, from the supersession log alone, why each change was made and what evidence prompted it. A doctrine that cannot be audited backwards has cut its own audit trail.

Candidate versions circulate internally before they reach the live site. A candidate version is not authority. Citation of a candidate section in a public artifact reframes it explicitly as “a forthcoming Nyalai doctrine section, not yet published on the live site.”

The anchor that distinguishes what the Doctrine may revise from what it must preserve:

“An institution must have stable principles, but revisable methods.”

The principles are the Constitution itself. The methods are the Refusal Doctrine, its gates, its arithmetic, its thresholds, its versioned rewrites. When the world changes and existing methods no longer serve the principles, the methods are revised. When a proposed revision would require the principles to bend, the revision is refused. This distinction between the stable and the revisable is enforced at every version seal by the Scientific Council per §5 and §8.

§5 · How a version is sealed

Sealing is a three-step act.

First, the authoring party finalizes the version and submits it to an adversarial validator instance structurally separate from the authoring path. The adversarial validator co-computes the SHA256 hash and adjudicates every enumerated UNVERIFIED item.

Second, the ratifying authority approves the sealed version. In the founding regime, the ratifying authority is Sébastien Assohou as founding operator. In the post-founder regime, the ratifying authority is the Nyalai Scientific Council per §8.

Third, the sealed version is published at nyalai.com/doctrine with a supersession log entry, a SHA256 signature, and the date of seal. A mirror snapshot is deposited at archive.nyalai.com. The version becomes canonical at the moment of public publication, not at the moment of internal ratification.

§6 · How an error is acknowledged

Errors admitted are credibility. Errors hidden are not.

When a section of the doctrine, an essay, a verdict, or any Nyalai artifact is discovered to contain a factual error, an attribution error, or a citation drift, the correction is published in public. The prior text is preserved in the archive under its original SHA256, marked with a correction annotation that points to the corrected version. The correction is logged in the supersession log with the date the error was discovered, the party that surfaced it (internal audit or external reader), the corrected text, and the reason for the original error.

The bibliographic errata precedent of Roger Clarke, Harindra de Silva, and Steven Thorley is the canonical case study. The Transfer Coefficient was frequently attributed to the second edition of Richard C Grinold and Ronald N Kahn's Active Portfolio Management. Careful reading of that edition establishes that the Transfer Coefficient enters the literature via the third edition or via the Clarke, de Silva, and Thorley 2002 formalization. An attribution error of that kind must be caught. Once caught, it must be corrected in public. The credibility gained by public correction is greater than the credibility protected by silence.

We do not argue for the prior error once the evidence overrides it. When cross-anchoring invalidates the premise of a claim, the claim is withdrawn rather than reworded.

Reputation is a capital and also a debt. Every verdict Nyalai issues creates a future obligation. If a decade later a Nyalai verdict is discovered to have been wrong, the obligation is to publish the finding rather than to bury it. The mechanism that carries this obligation forward across decades is a Public Registry of Revisions, established as a standing Foundation artifact:

“I would create a public registry of revisions. Not only new versions. Acknowledged errors. Corrections. Lessons. An institution that publicly documents its errors often becomes more credible than an institution that claims never to make any.”

The Public Registry of Revisions lives at nyalai.com/foundation/registry and is held by the Nyalai Foundation per §9. Each entry records the original verdict or doctrine claim, the date of the correction, the primary source or cross-anchor that surfaced the error, the corrected finding, and the lesson extracted for future practice. The Registry is append-only. Entries are never deleted and never rewritten. A reader consulting the Registry ten years after a verdict must be able to reconstruct the full arc of the correction.

§7 · How a decision is contested

Verdicts are contestable by design. A submitter, a third-party allocator, or any qualified reader may file a contestation against a published verdict per §11.4 of the Refusal Doctrine.

The contestation window is thirty days from the date of verdict publication. Within the window, the contesting party submits a written objection identifying the specific gate, threshold, or evidentiary claim under dispute, together with the counter-evidence and its primary sources.

The contestation is reviewed by an adversarial validator instance structurally separate from the instance that rendered the original verdict. The reviewing validator issues a signed addendum to the original verdict-form. The addendum affirms the original verdict, revises it, or withdraws it. The original verdict, the contestation, and the addendum remain part of the public verdict record in perpetuity.

Where the contestation raises a structural challenge to the doctrine itself rather than to a specific verdict, the challenge escalates to the Scientific Council for adjudication. Structural challenges may prompt a doctrine version bump per §4.

Every contestation and every outcome is preserved on the public historical record. Dissents are not editorial appendices. A well-authored dissent is the closest working analog to what a Nyalai verdict is. Per Ruth Bader Ginsburg, in her 2010 essay “The Role of Dissenting Opinions” (95 Minnesota Law Review 1), a dissent is a reasoned public disagreement addressed to the intelligence of a future day. Nyalai's contestation archive is built on that discipline.

Contestation is codified as a constitutional right of dissent, not as a discretionary courtesy:

“I would create a constitutional right of dissent. For example: publication of counter-arguments; methodological appeals; proposals for new Gates; documented critiques. Not to weaken Nyalai. To show that truth always matters more than the organization.”

The four instruments of the constitutional right of dissent are enumerated below. Each is receivable by Nyalai and reviewed under the procedure above. Refusal to receive a filed instrument is itself contestable to the Scientific Council.

1. Publication of counter-arguments. Any qualified reader may file a written counter-argument to a published verdict or doctrine section. The counter-argument is published alongside the original in the public verdict record.
2. Methodological appeals. A submitter, allocator, or independent researcher may file an appeal challenging the specific method used to reach a verdict (gate arithmetic, threshold selection, evidentiary weighting). The appeal is adjudicated by a validator instance structurally separate from the original.
3. Proposals for new gates. A reader may propose a new gate for inclusion in the Refusal Doctrine. The proposal is reviewed by the Scientific Council per §4 doctrine evolution procedure.
4. Documented critiques. Any qualified party may file a documented critique of a Nyalai artifact (essay, verdict, doctrine section, public statement). The critique enters the Public Registry of Revisions per §6 and is preserved in perpetuity.

The right of dissent is not designed to weaken Nyalai. It is designed to make truth structurally higher-priority than the organization that pursues it.

§8 · The founder role and its transmissibility

Sébastien Assohou is the founding operator of Nyalai. As founder, he holds version authority over the Doctrine, seals every version, publishes every verdict authored under the Nyalai name, and carries public responsibility for the institution.

The founder role is time-boxed. It is not held for life. The founder is bound by the same Constitution as every future ratifying authority. The founder cannot amend the Constitution unilaterally. Amendment requires the process defined in §10.

The founder role is explicitly designed for non-indispensability. The ultimate test of the founder's success is the day the institution no longer requires him:

“The success of Nyalai will be the day someone can say: ‘Sébastien Assohou founded Nyalai, but today Nyalai is greater than its founder.’ In my eyes, that is the ultimate test of an institution.”

To guard against the pathology of founder centralization, the founder monitors a single named indicator on a quarterly cadence:

“Which important decisions could have been taken without me? If the answer remains ‘none’ after several years, that is a warning signal.”

This indicator is named Founder Gravity. Once per calendar quarter, the founder reviews the important decisions of the prior quarter and asks which of them could have been taken by another Nyalai agent, another Nyalai council member, or another Nyalai contributor without founder involvement. If the answer remains “none” over a rolling multi-year window, the Constitution treats the situation as a warning signal requiring a structural response, up to and including a founder-initiated succession per §12.9 of the Refusal Doctrine and per the Succession Statement filed at nyalai.com/doctrine/governance. The Founder Gravity indicator is not optional. It is a constitutional obligation of the founding operator.

In the founding regime, the Founder Gravity quarterly review is logged internally as required by this section. Once external Scientific Council members are appointed, in Q1 2027 or earlier, the founder's quarterly review is delivered to the Council for independent challenge. Multi-year “none” outcomes trigger a Council-initiated structural review even if the founder does not initiate one. The founding operator does not enforce the Founder Gravity signal against himself in isolation; the enforcement is Council-vested at Council convention.

Post-founder governance is a three-part structure.

The Nyalai Foundation holds the Constitution, the Refusal Doctrine, and the Knowledge Graph. The Foundation is a non-commercial entity whose sole purpose is to preserve the founding artifacts of Nyalai in perpetuity. The Foundation cannot be sold, cannot be merged, and cannot be dissolved without a public two-thirds ratification of the Scientific Council plus a thirty-day public comment period.

The Nyalai Scientific Council is the ratifying authority for doctrine version seals and for structural contestations. The Council is constituted of independent members drawn from institutional finance, quantitative research, jurisprudence, and regulatory practice. Council terms are staggered to prevent generational capture. The Council convenes at least once per calendar quarter.

The Council composition. Seven distinct profiles compose the initial Council roster:

“A scientist; a jurist; a philosopher of science; a former regulator; an institutional investor; a systems engineer; a statistician.”

The mission of the Council is not to validate Nyalai decisions. It is to attempt to demonstrate why those decisions could be wrong:

“Their mission would not be to validate the decisions. Their mission would be to attempt to demonstrate why they could be wrong.”

This composition and mission are load-bearing. A council whose composition would tend toward consensus with the founder defeats the constitutional purpose of the Council. The seven profiles are selected precisely because their disciplines carry different failure modes and different reasoning traditions, so that a proposed doctrine change encounters seven distinct forms of adversarial pressure before it seals.

Nyalai Labs is the operational entity. Labs runs the specialist agents (Cice validator, Meta-Judge, Editorialist, Finance-Lawyer, Dispatch, Knowledge Distributor), publishes verdicts, authors the After Verdict editorial arm, and holds commercial contracts with submitters. Labs is an operating enterprise. Labs may be reorganized, may accept investment, may accept acquisition offers, subject to the constraints of §9.

The transition from founding regime to post-founder regime is triggered by the succession protocol defined at §12.9 of the Refusal Doctrine and the Succession Statement filed at nyalai.com/doctrine/governance.

§9 · What belongs to whom

Nyalai's governance rests on a three-container structure: Foundation, Institute, Labs. Each container has a defined scope, a defined authority, and a defined boundary that separates it from the others. The separation is not administrative convenience. It is a constitutional principle designed to prevent the pathology in which commercial pressure eventually distorts the doctrine that commercial products are supposed to apply.

§9.1 · The three-container separation

The governance separation:

“I would create a constitutional principle. Products may request an evolution of the Doctrine. They can never modify it directly. They submit a proposal. Science studies it. The Foundation decides. This separation seems essential to me.”

Products may request doctrine evolution. Products cannot modify doctrine directly. Products submit a proposal. Science studies the proposal. Foundation decides. This three-step separation is what protects the Doctrine from silent commercial capture. Any attempt to shortcut the sequence, whether by allowing Labs to write directly into the doctrine or by allowing the Institute to seal a version without Foundation ratification, is a constitutional violation.

§9.2 · Assignment inventory

• The Constitution: Foundation. Untouchable.
• The Refusal Doctrine and its version history: Foundation. Untouchable.
• The Knowledge Graph and its provenance discipline: Foundation. Untouchable.
• The Trust Seal standard, methodology, and canonical phrasing: Foundation. Licensed to Labs for issuance.
• The Public Registry of Revisions per §6: Foundation. Append-only. Untouchable.
• Scientific research programs, peer-review journal, academic publication pipeline: Nyalai Institute. Independent scholarship, no commercial obligation.
• The Experimental Lab space per §9.4 below: Nyalai Institute. Exploration with no promise, no certification, no doctrinal claim.
• The specialist agents (Cice validator, Meta-Judge, Editorialist, Finance-Lawyer, Dispatch, Knowledge Distributor, Reputation Guardian, Historian): Labs. Operational.
• The commercial artifacts (verdict-forms sold to submitters, Trust Seal licenses issued under Foundation standard, After Verdict subscriptions): Labs. Operational.

If, in a future decade, Nyalai Labs is acquired by a third party, the Constitution, the Doctrine, the Knowledge Graph, and the Trust Seal standard remain with the Foundation. The buyer acquires the operating enterprise. The buyer does not acquire the doctrine. A future Labs operating under new ownership either continues to apply the Doctrine or is not recognized by the Foundation as authorized to publish under the Nyalai name. The Foundation retains the right to revoke Nyalai naming recognition from any Labs successor whose practice deviates from the Constitution.

The separation is intentional. Trust in a probabilistic prediction system cannot be sold. Trust in a validator cannot be sold. What can be sold is the operational apparatus. The founding artifacts stay.

§9.3 · The Nyalai Institute

The third container is named:

“Nyalai Institute. Short. Elegant. Timeless. And I would let the programs carry the specialized names.”

The Institute container is deliberately named without a discipline suffix. The container is stable; the programs inside it carry the specialized names and may evolve over the coming decades without renaming the container. Candidate programs inside the Nyalai Institute include, at initial scope, the Center for Adversarial Validation Science, the Trust Research Program, the Knowledge Graph Laboratory, and the Institutional Governance Program. The Nyalai Institute is filed as a distinct entity, expected in the Q3 to Q4 2027 window, once the Foundation is filed and the initial peer-review journal charter is drafted. Prior to filing, Institute-scoped research is held under Foundation stewardship.

§9.4 · Experimental Lab space

Not everything Nyalai does should be a protocol, a gate, or a certified verdict. To preserve the space in which future methods are discovered, the Constitution reserves an Experimental Lab space:

“Experimental Lab. No promises. No certification. No doctrine. Simply a space where one can test ideas that, perhaps, will one day become standards.”

The Experimental Lab is a sub-space of the Nyalai Institute. It carries no promises, no certifications, no doctrinal claims. Work produced inside the Experimental Lab is labeled as such and is explicitly separated from any Foundation-sealed artifact. This separation prevents the Experimental Lab from becoming a back door through which uncertified content acquires Nyalai authority by association.

§9.5 · Roster of specialist agents

The Labs roster carries the specialist agents that render verdicts, publish essays, run outreach, and coordinate the operational surface of Nyalai. At the date of ratification of this Constitution, the roster includes eight canonical agents. Six are founding roster:

• Cice validator (adversarial validation)
• Meta-Judge (doctrinal coherence check + escalation logic)
• Editorialist (After Verdict editorial arm + dissent register + public voice)
• Finance-Lawyer (contracting + regulatory framing)
• Dispatch (institutional relations + orchestration)
• Knowledge Distributor (Knowledge Graph curation + terminology consistency)

Two additional agents:

• Reputation Guardian (7th canonical agent, standalone at Q4 2026 launch). Measures the reputation impact of each Nyalai publication before ship. Independent of Meta-Judge to preserve adversarial independence between doctrinal coherence and reputation-vector assessment.
• Historian (8th canonical agent, standalone at Q1 2027 post-Foundation formation). Tracks doctrinal evolution across versions and preserves the reasoning trail behind every supersession. In the interim between Q4 2026 launch and Q1 2027 filing, Historian responsibility is distributed across Dispatch and Knowledge Distributor.

Additional agent roles may be proposed in the future through the amendment process of §10. The number of agents is not fixed. The commitment is that each agent's function is testable, its output is independently auditable, and its escalation path is explicit.

§9.6 · Digital heritage stewardship

The Nyalai Foundation carries a stewardship obligation for the founding artifacts across horizons that exceed the founder's lifetime:

“What becomes of the Knowledge Graph in fifty years? In one hundred years? Who is its guardian? Who decides that it remains accessible? If you want to build an institution, this reflection deserves to be engaged very early.”

The Foundation retains long-term custody of the Constitution, the Refusal Doctrine, the Knowledge Graph, the Trust Seal standard, and the Public Registry of Revisions across horizons of fifty to one hundred years. Custody instruments include the two-tier archival stack at archive.nyalai.com, the Wayback Machine snapshot cadence, and Foundation-owned institutional partnerships that inherit access rights in the event of Foundation dissolution. The digital heritage question is not deferred. It is engaged at Foundation formation Q1 2027 and reviewed at each Constitution amendment cycle per §10.

§9.7 · Legal instantiation and pre-launch explainer

The legal instantiation of the three-container separation is sequenced. The Nyalai Foundation and Nyalai Labs are filed as distinct legal entities in Q1 2027, with IP assignment of the Constitution, the Refusal Doctrine, the Knowledge Graph, and the Trust Seal standard transferred to the Foundation, and operational contracts, specialist agents, and commercial artifacts assigned to Labs. The Nyalai Institute is filed subsequently in the Q3 to Q4 2027 window per §9.3 above. Prior to Q1 2027 filing, the Foundation and Labs operate as a founder-held single-entity arrangement whose separation is expressed doctrinally through this Constitution rather than through corporate registration. At the date of this text, Nyalai Foundation and Nyalai Institute are designated entities scheduled for legal filing; neither is currently a registered legal person, and no tax-exempt determination has been received or is represented.

In the Q4 2026 pre-launch outreach kit that accompanies the public activation of nyalai.com and afterverdict.co, a one-page IP and Foundation transition explainer is included as an anticipatory answer to institutional readers who will reasonably ask why the operating entity does not hold the doctrine. The explainer is available at nyalai.com/foundation. It sets out the Q1 2027 filing timeline, the assignment inventory, and the doctrinal reason for the separation, so that no institutional reader discovers this arrangement after a contract discussion has begun.

§10 · How this Constitution is amended

Amendment is deliberate, public, and rare.

An amendment may be proposed by any member of the Scientific Council, by the Foundation directors, or by the founding operator in the founding regime. The proposal is published on the public site at nyalai.com/constitution/proposals for a thirty-day comment period. Public comments are archived alongside the proposal.

After the comment period, the amendment is put to a vote of the Scientific Council. Adoption requires a two-thirds affirmative vote of the Council plus the affirmative vote of the Foundation directors acting unanimously.

Upon adoption, the amended Constitution is published as a new version. The prior version is preserved verbatim in the archive. A supersession log entry records the amendment, the vote count, a summary of public comments, and the reasoning for the change.

The preamble of this Constitution is not amendable in ordinary process. Amendment of the preamble requires the same process plus an additional sixty-day public comment period and unanimous Scientific Council affirmation.

During the founding regime, prior to the Q1 2027 Foundation filing and prior to Scientific Council convention, amendment authority sits with the founding operator, exercised in public on the same site as the Constitution, with a mandatory public comment period of thirty days and a supersession log entry documenting the reasoning for each change. The full §10 process (Scientific Council two-thirds affirmative vote plus Foundation directors acting unanimously) becomes binding at the moment the Council is convened and the Foundation is filed, and this interim rule sunsets at that moment. This interim design is disclosed publicly so that institutional readers can distinguish founding-regime edits from post-founder amendments and price the two accordingly.

§11 · The Nyalai language

An institution builds a language. The language protects the institution.

Nyalai's operating vocabulary: Gate, Verdict, Calibration, Meta-Judge, Refusal, Evidence, Register, Sources read, Model Trust Level, Trust Seal Validated by Nyalai, Knowledge Graph, Supersession log. Each term carries a specific technical meaning defined in the Refusal Doctrine. Each term is used consistently across every artifact that ships under the Nyalai name.

Terminological drift is a form of doctrinal drift. A verdict rendered against Gate 3 that describes the gate loosely as “the significance test” is a verdict whose readers cannot cross-check the arithmetic against the Doctrine. Consistency of language is a load-bearing property of institutional trust.

The Nyalai vocabulary is protected as institutional asset. Terms such as Trust Seal Validated by Nyalai and Model Trust Level are commercial artifacts held by Labs under license from the Foundation. Terms such as Refusal Doctrine, Sources read, and Knowledge Graph are open-license terminology any competent party may adopt in the practice of adversarial validation.

The language is not defensive. It is descriptive. Nyalai builds the vocabulary because the vocabulary did not previously exist at the resolution the practice requires.

§11.1 · Language phasing

Nyalai publishes in phased language expansion. The phasing follows institutional-respect logic, not marketing translation logic. The founding text is French in the preamble by ratified authorship (see §12) and English in the operational vocabulary by institutional-audience default. Beyond the founding baseline, Nyalai adds languages in the sequence below, each keyed to a specific institutional cohort whose age and traditions predate ours.

The governing statement of the language phasing, as amended 2026-07-30 under the founding-regime authority of §10:

“Nyalai's language expansion follows a phased sequence: English and French together at the Q4 2026 launch, then Japanese and Mandarin added during Q4 2026 for four-language coverage by the start of Q1 2027. This coverage reflects the ambition of an independent institution without linguistic borders: accessible to institutional allocators in all major financial centers (New York, London, Paris, Tokyo, Hong Kong, Singapore, Shanghai) and respectful of the institutional traditions of empires whose history precedes ours by several centuries. The prioritization serves two inseparable logics: on the one hand, accessibility for institutional investors without limit of language; on the other, respect for ancestral institutional traditions. This second logic is not cosmetic: an institution that aspires to endure across generations recognizes that it takes its place within an order older than itself.”

At launch (Q4 2026). English and French ship together. This is the founding baseline. Every reader-facing surface at nyalai.com and afterverdict.co ships in English and French at launch. The French preamble is preserved as the founding text of the institution per §12 acknowledgments. The French edition of the full Constitution and of the Refusal Doctrine is prepared by Nyalai internal editorial pass with adversarial validator seal per §5.

During Q4 2026. Japanese translation of the Constitution and of the Refusal Doctrine is added at nyalai.com/jp. Rationale: the Government Pension Investment Fund of Japan is the largest pension fund on the planet. Japanese institutional finance culture is precision-oriented and holds discipline standards aligned with Nyalai's own. Language expansion in Japanese is a mark of traditional respect toward institutions whose lineage predates ours by centuries.

During Q4 2026, completing four-language coverage by the start of Q1 2027. Mandarin translation of the Constitution and of the Refusal Doctrine is added at nyalai.com/zh. Rationale: the China Investment Corporation, the Hong Kong Monetary Authority, and the broader Chinese institutional peer set operate at scale that requires counterparty engagement in their working language. Same tradition-respect logic applies: institutions older than ours receive translation at Foundation level as a signal that Nyalai does not expect them to meet us in our language alone.

Deferred, with explicit rationale. Russian is deferred given sanctions context 2022 onward that has generated friction in institutional cross-border practice, and given a smaller institutional peer set post-2022 relative to the Asian tier. Portuguese and Arabic are deferred based on Nyalai's assessment that South American and Middle Eastern institutional prospects read English fluently at the working level, meaning translation would not measurably shorten the institutional-audience access path. Deferred does not mean refused; if an institutional prospect signal in any of these languages materializes with a specific request for translation, the deferral is reconsidered under the same tradition-respect logic that governs the Japanese and Mandarin additions.

Each language edition of a Nyalai artifact is a complete standalone document. Nyalai does not publish interlinear or dual-language content on a single surface. Where interpretive ambiguity arises, the original-language text governs, and each edition identifies its governing original in its translation note.

Two observations on the framing. First, the language expansion is not marketing translation. It is institutional respect signal. Nyalai treats language expansion at Foundation level as a form of the same discipline that governs the Trust Seal itself: a signal of durable standard, not of promotional reach. Second, no peer in the AI-plus-institutional-finance validation space currently ships translations at Foundation-of-founding-text level. That absence is itself a differentiation opportunity. Nyalai fills the gap because the gap should not exist; the fact that filling it also becomes an audience-access lever is a downstream consequence, not the primary reason for filling it.

§11.2 · The self-doubt discipline

An institution that stops asking itself where it might be wrong is an institution in the early stage of becoming a monument. To prevent that pathology, the Constitution establishes a recurring practice for the founding operator, and for any future ratifying authority:

“What would make me abandon or rewrite an important part of this doctrine? If the answer one day becomes: ‘Nothing.’ Then that will be the most dangerous moment in the life of Nyalai.”

Once per calendar quarter, the founding operator (or the post-founder ratifying authority) answers this question in writing. The answer is logged internally, not published. If, over a rolling multi-year window, the answer converges toward “nothing,” the Scientific Council is notified per §5 and a formal structural review is triggered.

The self-doubt discipline is not a symbolic gesture. It is the institutional immune system against the specific failure mode in which a founder falls in love with an architecture and stops testing whether the architecture still serves the mission. Every institution that has failed at this test has failed for the same reason. Nyalai names the failure mode in advance and installs the check that catches it.

§11.3 · The peer institutional library

Nyalai does not copy a single institution. It builds a library of governance patterns drawn from institutions whose specific strengths inform specific Nyalai practices. Each entry names a peer institution and the specific pattern Nyalai studies from it:

• Underwriters Laboratories (UL) for the certification model: a third-party specialist that renders binary safe/not-safe verdicts on specific claims, with the standard held independent of the certified entity.
• Moody's for the accountability model: a rating body carries reputational consequences for its ratings and is publicly identified with its calls.
• IEEE for standards production: a professional body that produces technical standards through open deliberation with documented dissent.
• IETF for the transparent process model:

  “RFCs do not only say: ‘Here is the standard.’ They often explain: why; which alternatives were rejected; which limits remain. I think Cice could draw on this documentary culture.”

  IETF RFCs document the standard, the reasoning behind the standard, the alternatives rejected, and the limits that remain. Cice verdict-forms adopt this culture.
• NIST for the framework model: a reference body that distinguishes principles, recommendations, and profiles at different levels of prescriptiveness.
• The methodological structure of a supreme court:

  “Majority opinion. Concurring opinions. Dissenting opinions. Historical record. I think there is something to learn from this way of making disagreements visible.”

  Not the substantive law. The structure. Majority opinion, concurring opinions, dissenting opinions, historical record. Nyalai's contestation archive per §7 and its Public Registry of Revisions per §6 adopt this structure.

The library is not a copy. It is a taxonomy of patterns that Nyalai selects from at each design decision. The library grows as Nyalai encounters additional institutional patterns worth studying. Each addition is ratified through the amendment process of §10.

§11.4 · Terms of art and registration status

“Verdict” in Nyalai usage refers to a Cice-issued adjudication on a specific prediction-system claim under the Refusal Doctrine. It is a term of institutional art, not a judicial finding, not a legal opinion, and not a representation of judicial authority. Every published verdict-form carries a footer disclaimer restating this scope.

Trust Seal Validated by Nyalai and Model Trust Level are unregistered trademarks at the date of this text, with institutional filing scheduled alongside the Q1 2027 Foundation formation. This Constitution does not create or claim registered trademark rights in either term; it records the design under which those rights, once registered, are held by the Foundation and licensed to Labs per §9.

§12 · Ratification and acknowledgments

Founding ratification of the Nyalai Constitution is by Sébastien Assohou, founding operator of Nyalai.

Acknowledgment for the preamble contribution: the preamble of §1 was drafted by ChatGPT (external LLM adversary signal from a model outside the Anthropic family). The preamble is preserved verbatim in French as the founding text of the institution, in the French version of this Constitution; §1 above renders it in English translation. Its authorship is external to Nyalai. Its authority is internal to Nyalai by explicit ratification.

Acknowledgment for the dissent tradition anchoring §6 and §7: Ruth Bader Ginsburg, “The Role of Dissenting Opinions,” 95 Minnesota Law Review 1 (2010); Michael Angelo Musmanno, Justice Musmanno Dissents, Bobbs-Merrill Company 1956, volumes 1 and 2; Louis D Brandeis, Bar Memorial (posthumous).

Acknowledgment for the bibliographic errata precedent of §3 and §6: Roger Clarke, Harindra de Silva, and Steven Thorley, “Portfolio Constraints and the Fundamental Law of Active Management,” Financial Analysts Journal 58, no. 5 (2002).

Acknowledgment for the fourth-layer allocator due diligence framing of §2: internal Nyalai survey of twelve institutional due diligence firms, 2026.

Acknowledgment for the regulatory anchor referenced across sections: Board of Governors of the Federal Reserve System, Office of the Comptroller of the Currency, and Federal Deposit Insurance Corporation, SR 26-2 “Model Risk Management,” issued 2026-04-17, which supersedes SR 11-7 of April 4 2011 and SR 21-8 of April 9 2021.

§12.1 · The five architectural layers of Nyalai

Beyond the governance three-container structure of §9 (Foundation, Institute, Labs), Nyalai as a built artifact carries a distinct architectural taxonomy. Five architectural layers compose the built structure of Nyalai:

1. The Doctrine: the scientific constitution of Nyalai's method. Explains why. Never depends on the product.
2. The Validator: Cice. The operational framework that applies the method.
3. The Knowledge Graph: the immutable provenance-anchored library of every citation, source, version, and reading. Held by the Foundation per §9. Treated as critical infrastructure.
4. The Trust Layer: the certification surface at which Nyalai issues verdicts and licenses the Trust Seal. This is where reputation is transferred from Nyalai to the reader who accepts the verdict.
5. The Community: the network of Fellows, Research Affiliates, Academic Partners, Certified Validators, and Institutional Members that recognizes Nyalai as peer institution. An institution is never built solely by its founder; it is progressively recognized by its peers.

The architectural enumeration above answers the question “what has Nyalai built?” The governance enumeration of §13 below answers a different question: “how does Nyalai make decisions?” The two enumerations are complementary lenses on Nyalai's structure, not competing definitions of the same five things. A reader who tries to collapse them into a single list will misread the design. The architectural layers describe substance; the governance layers describe procedure. Both are load-bearing.

§12.2 · The compressed ratification arc

The 2026-07-29 external LLM adversary prompt exchange surfaced structural additions that the founding operator ratified as constitutional text the same day. The compressed timeline reflects that the underlying work (the non-negotiables, the Foundation, Institute, and Labs separation, the fourth-layer positioning, the Founder Gravity discipline, the Jurisprudence layer) had been developed inside Nyalai over the twelve months preceding. The external LLM adversary exchange named what was already load-bearing in the practice, and the founding operator ratified the naming as constitutional text. Future amendments are governed by §10 and will not follow this compressed cadence.

§13 · The jurisprudence layer

The fifth layer of Nyalai's governance stack is jurisprudence:

“A fifth layer is missing. Jurisprudence. I am not speaking of law. I am speaking of decisional memory. Every important decision becomes a precedent. Every exception is documented. Every doctrinal change cites the precedents that motivated it. [...] Because an institution does not become credible through its principles alone. It becomes credible through the coherence of its decisions over time.”

Jurisprudence in the Nyalai sense is not law. It is decisional memory. Every important decision becomes a precedent. Every exception is documented. Every doctrinal change cites the precedents that motivated it. Over time, Nyalai accumulates an internal jurisprudence: a record of how the Doctrine has been applied to specific cases, what exceptions were granted and why, and how the reasoning of one verdict connects to the reasoning of the next. This jurisprudence is what allows a reader to test two things at once: whether Nyalai holds coherent principles, and whether Nyalai's decisions are coherent with each other across the years.

Under this five-layer governance stack, the layers are:

1. Constitution: the founding text (this document).
2. Doctrine: the operational methodology (Refusal Doctrine at nyalai.com/doctrine).
3. Science: the research programs of the Nyalai Institute per §9.3.
4. Products: the operational surface of Nyalai Labs per §9.
5. Jurisprudence: the decisional-memory layer codified in this section.

The jurisprudence layer is held under Foundation stewardship per §9. It is complementary to but distinct from the Knowledge Graph: the Knowledge Graph carries the sources; the jurisprudence carries the record of how those sources were used in Nyalai's decisions. The jurisprudence archive is a first-class Foundation artifact. It is versioned. It is public. It grows with each verdict, each contestation per §7, each doctrine version bump per §4, each Constitution amendment per §10, and each entry in the Public Registry of Revisions per §6.

Reconciliation note: the architectural enumeration of §12.1 and the governance enumeration of this §13 both use the label “five layers.” They are different lenses on Nyalai's structure: architectural (what is built) versus governance (how decisions are made), preserved side by side rather than collapsed into a single list. The Community layer of the architectural stack (§12.1 layer 5) is not the same object as the Jurisprudence layer of the governance stack (§13 layer 5). Both are preserved explicitly. A future reader who encounters “the fifth layer of Nyalai” must consult context to determine which enumeration is being referenced.

§14 · Supersession log

This log records each amendment to this Constitution after founding ratification. Entries are append-only.

1. 2026-07-30, evening, New York. Scope: §11.1 language phasing. Prior version: English and French at the Q4 2026 launch; Japanese in a second phase (Q2 2027 or first Japanese institutional prospect signal, whichever first); Mandarin in a third phase (Q3 to Q4 2027 window or first Chinese institutional prospect signal, whichever first). New version: English and French together at the Q4 2026 launch, Japanese and Mandarin added during Q4 2026, four-language coverage by the start of Q1 2027. Rationale: accelerated four-language coverage target ratified by the founding operator under the founding-regime amendment authority of §10.
2. 2026-07-30, evening, New York, procedural disclosure. The §11.1 amendment above was applied pre-launch, before the Étape 6 site flip, and therefore without the thirty-day public comment period specified in §10. This is defensible under §10's founding-regime clause: the site was not yet LIVE, meaning no public audience existed to consult. The Constitution's amendment procedure is preserved intact for all post-launch amendments, which will follow the full §10 thirty-day comment period. This disclosure is entered append-only to preserve the transparency principle: every deviation from a stated procedure is logged in the same register as the amendments it produced.
3. 2026-08-06. Scope: §11.1 · “sovereign wealth fund” → “pension fund” for GPIF classification. Adversarial re-review verified against gpif.go.jp/about that GPIF self-describes as 年金積立金管理運用独立行政法人 (pension reserve manager) not sovereign wealth fund. Institutional standard convention · GPIF = world’s largest pension fund. Original wording preserved for provenance · corrected forward.