From Claims to Articles
The preceding chapters have identified empirical and structural conditions of computational governance and stated the normative claims those conditions support. The conditions can be tested against the world; the normative claims can be accepted, rejected, or revised.
The proposal asks a more specific question: if those conditions obtain and those normative premises are accepted, what rights, institutions, and procedures are required to prevent the failures of legitimacy they predict? The answer is neither a deduction from fact nor an assertion that one constitutional design is uniquely true. It is a claim of conditional necessity.
The Structure of Inference
Constitutional reasoning does not deduce rights from facts. It identifies structural conditions under which governance without certain protections produces domination, and then specifies the minimum protections required to prevent it. The logic is: if a structural condition obtains, then a governance framework that ignores it will fail in a predictable way. The article is the minimum institutional response that closes the failure mode.
Every constitutional tradition that has endured worked this way. Madison did not deduce the separation of powers from a theorem about human nature. He identified a structural condition (that concentrated power tends toward abuse) and specified the minimum institutional arrangement that prevents the concentration. Locke did not deduce the right to property from the state of nature. He identified a structural condition (that labor intermixed with the world creates claims that antedate government) and specified the minimum protection that any legitimate government must provide.
The inference here follows the same pattern. The book identifies structural conditions of computational governance — tempo asymmetry, compositional failure, quiet foreclosure, domination without a dominator. The articles are the minimum responses. The word minimum sets the proposal's scope. A stronger response would be a permission-based regime, in which computational systems may act only with prior authorization. A weaker response would be a disclosure-only regime, in which systems must describe what they do but need not permit contestation. The receipt regime occupies the constitutional middle: strong enough to make coercion legible and contestable, specific enough to be implementable, modest enough to survive political opposition from those whose interests it constrains.
Other constitutional responses to the same structural conditions are possible. The articles that follow are offered not as the only response but as a sufficient one — precise enough to implement, critique, and improve. Specification disciplines thought in ways that aspiration does not.
From Claims to Articles
Each article below is justified by the structural condition that makes governance without it a form of domination. The justification is not optimality but necessity: given the structural condition, the absence of the article produces a specific failure of legitimacy.
Article 1. Right to a receipt at time of coercive act.
Structural condition: When decision tempo exceeds participation tempo (C1), the governed cannot contest what they cannot inspect. A system that acts in milliseconds and answers in months has satisfied the form of governance while violating the substance. Without the article: Coercion becomes invisible. Not denied, not justified, simply unrecorded. The affected person has no artifact to contest, no record to appeal, no evidence to present. What existing law provides: GDPR Articles 13-14 impose scoped information duties when personal data are collected from the subject or obtained elsewhere; Regulation Z requires disclosures in covered consumer-credit transactions.1 What the article adds: The extension from notice to contestable receipt. Notice tells you what happened. A receipt gives you the artifact you need to challenge it.
Article 2. Right to contest through independent review.
Structural condition: Receipt fields must be sufficient for contestation (C8), and receipt regimes fail in predictable ways (C9). A receipt that cannot be challenged is a record of subjugation, not a tool of accountability. Without the article: The receipt becomes ornamental, filed and archived but never examined. Power generates paperwork without generating answerability. What existing law provides: GDPR Article 22 gives a person the right not to be subject to a solely automated decision with legal or similarly significant effects, subject to defined exceptions. In two of those exceptions it requires safeguards including human intervention, an opportunity to state a view, and contestation.2 What the article adds: Independent review with a named reviewer, a deadline, and authority to reverse. This article specifies proportionality tiers calibrated by severity, with critical actions reviewed within days and routine actions within weeks.
Article 3. Right to human review for material adverse actions.
Article 2 guarantees contestability after the act. Article 3 addresses a different question: whether certain acts require an irreducibly human point of authorization or review before their consequences become binding. The distinction is between the right to challenge and the right to a human in the chain of judgment.
Structural condition: Legitimacy requires contestability at human tempo (C1), and local compliance does not guarantee global accountability (C7). A system may be correct in every component and unjust in composition. Without the article: The authorization membrane hollows out. The human who signs becomes a rubber stamp, processing outputs at a rate that precludes comprehension. Accountability attaches to form rather than substance. What existing law provides: The EU AI Act requires effective human oversight for high-risk systems and requires deployers to assign oversight to people with the necessary competence, training, and authority.3 What the article adds: A constitutional floor for material adverse acts beyond that defined statutory coverage, one that private agreement or routine administrative choice cannot reduce.
Article 4. Right to export data, credentials, and relational context.
Structural condition: Contestability requires credible exit (C10). Quiet foreclosure eliminates alternatives through architecture rather than force (C11). If you cannot leave, your right to contest is a petition, not a claim. Without the article: Lock-in masquerades as choice. The user who has accumulated five years of reputation, relational context, and credential history on a platform cannot exit without losing the conditions of their participation. What existing law provides: GDPR Article 20 covers personal data provided by the subject where processing is automated and based on consent or contract. The Digital Markets Act imposes data-portability and specified interoperability duties on designated gatekeepers and covered core platform services.4 What the article adds: The extension from qualifying data to credentials and relational context. The right protects the elements of standing needed to arrive elsewhere, not data export alone.
Article 5. Right to present portable credentials.
Structural condition: Without fork rights, receipts become petitions (C10). The trust tax persists through lock-in (C3). Article 4 prevents exit from becoming amputation. Article 5 prevents the issuing platform from controlling where a portable credential may be presented. A credential that can be exported but used only with its issuer remains an instrument of membership, not portability. Without the article: Each platform becomes a sovereign issuer of standing, and leaving one means starting from zero at the next. The governed can carry their history but cannot use it. Exit is formally free and practically ruinous. What existing law provides: The W3C Verifiable Credentials Data Model v2.0, a Recommendation dated 15 May 2025, defines a technical model for portable attestations.5 What the article adds: The right of presentment. A compatible system must permit the credential to be submitted without the issuing platform's permission. It remains free to judge the issuer, provenance, and claim under its governing rules.
Article 6. Right against cross-domain composition without consent.
Structural condition: Bilateral coherence does not guarantee compositional coherence (C4). Without a mercy threshold, receipted coercion becomes permanent caste (C12). Without the article: A credit decision, a hiring decision, and an insurance decision — each independently compliant, each drawing on different data, none aware of the others — compose into a condition of life that no single actor chose and no single institution can reverse. What existing law provides: The provisions surveyed here regulate specified data processing, portability, and automated decisions within defined scopes; they do not, through those provisions alone, make the cumulative cross-domain effect of separately lawful acts a distinct object of contestation.6 What the article adds: The proposal treats that cumulative effect as a distinct constitutional object.
Article 7. Right to time-bounded personal records.
Structural condition: Without a mercy threshold, verification becomes caste (C12). Perfect memory forecloses transformation. Without the article: Every failure, every dispute, every youthful misjudgment persists in machine-readable form, composable across domains, unsealable and unexpirable. The person becomes the permanent sum of their documented worst moments. What existing law provides: GDPR Article 17 creates a qualified right to erasure, with stated grounds and exceptions. The FCRA time-limits many categories of adverse information, with different periods and exceptions.7 What the article adds: Constitutional temporal asymmetry. The right advances from requesting deletion to requiring the system to forget under declared rules: designed forgetting as a constitutional principle rather than an administrative convenience.
Article 8. Right to know the rule invoked.
Structural condition: Receipt fields must be evaluable (C8). Domination without a dominator is structural (C7). You cannot contest a rule you cannot name. Due process has always required knowable grounds: nulla poena sine lege (no punishment without a law) is the oldest procedural guarantee in the Western legal tradition. The computational version is: no adverse act without a nameable rule. Without the article: The governed face outcomes without reasons — not because reasons are withheld, but because the system that produced the outcome has no category for reason-giving. Administrative opacity converts power from something that can be challenged into something that can only be endured. What existing law provides: GDPR Articles 13(2)(f), 14(2)(g), and 15(1)(h) require, in defined automated-decision cases, meaningful information about the logic involved. In Case C-203/22, the Court of Justice read Article 15(1)(h) to require an intelligible explanation of the procedure and principles actually applied, not a complex formula or an exhaustive technical description.8 That movement toward intelligible, case-specific explanation converges with the problem identified here; a legal standard does not by itself guarantee dependable implementation. What the article adds: A stable rule identifier, the conditions that made the rule applicable, a contemporaneous receipt, and recourse. These make explanation structural and contestable rather than dependent on case-by-case litigation after the harm.
Article 9. Right against excess data collection.
Structural condition: The trust tax includes extractable surveillance premiums (C3). Quiet foreclosure operates through information asymmetry (C11). Excess collection gives the intermediary an advantage that compounds: the more it knows, the harder it becomes for the governed to exit, contest, or even perceive the scope of its authority. Information asymmetry can therefore become a condition of domination without a dominator. Without the article: Every interaction becomes an extraction opportunity. The coherence fee, the genuine cost of making truth compose, is padded with a surveillance premium that funds the position from which composition is performed. The intermediary's knowledge advantage compounds across time and across domains. What existing law provides: GDPR Article 5(1)(c) requires data minimization within a scheme that includes supervisory orders, complaints, judicial remedies, compensation, and administrative fines.9 What the article adds: Constitutional status. This article treats excess collection as an enabling condition of domination and bars it at the constitutional floor.
Article 10. Anti-waiver: the constitutional floor cannot be contracted away.
Structural condition: Domination persists despite contractual consent (C7). Foreclosure of alternatives makes "consent" structurally coerced (C11). Without the article: The receipt regime dissolves into the terms of service. Every protection becomes waivable by the click that precedes the service. The constitutional floor is contractible, and the contract is adhesive. What existing law provides: Existing law contains targeted models of non-derogation and anti-waiver: ECHR Article 15(2) protects specified rights even in an emergency, while section 29(a) of the Securities Exchange Act voids agreements to waive compliance with that statute or its rules.10 What the article adds: The principle that the receipt regime's protections are constitutional, not contractual. No terms-of-service agreement may waive Articles 1–9 or relieve a covered authority of the duties imposed by Articles 11–12. The right to a receipt is not a feature. It is a condition of legitimate computational governance.
Article 11. Right to an independent record.
Structural condition: Receipt integrity requires a record that cannot be unilaterally edited by the authority it constrains (C8), and capture must remain detectable and correctable before procedure becomes ceremonial (C9). Without the article: The authority whose power the receipt is meant to constrain controls the sole authoritative evidence of its own acts. What existing law provides: The provisions surveyed in the preceding articles do not supply the independent-record protection proposed here.11 What the article adds: No receipted authority may control the sole authoritative copy of its own coercive acts. Revision, suppression, censorship, and capture must be independently detectable and costly in proportion to the power the record constrains, and the securing constraint must lie outside the authority's unilateral control. The Joule Standard is the book's present high-assurance implementation, not the constitutional invariant. This remains the proposal's most contested article and the one most directly connected to the research program's formal results.
Article 12. Mandatory independent audit.
Structural condition: Receipt regimes fail in predictable ways (C9). Compositional failure is computable and therefore auditable (C4). Without the article: The receipt regime is self-certifying. The system that issues receipts also evaluates whether the receipts are adequate. The auditor and the operator are the same entity. What existing law provides: Sarbanes-Oxley created the PCAOB to oversee audits of public companies and imposed specified auditor-independence rules, including restrictions on non-audit services and audit-committee preapproval.12 What the article adds: Extension from public-company financial auditing to computational governance. The principle is the same — the entity subject to audit may not control the audit. The domain is new.
Alternative Responses
These twelve articles are not the only constitutional response to the structural conditions the book identifies. Two alternatives bracket the space:
A disclosure-only regime would require computational systems to publish their rules, parameters, and decision logs — transparency without contestability. This is weaker than the receipt regime because disclosure without a right to contest is observation without agency. The governed can see what is done to them but cannot challenge it. Disclosure regimes tend to produce informed acquiescence rather than structural accountability.
A permission-based regime would require prior authorization for every coercive computational act — approval before action. This is stronger than the receipt regime but likely unworkable at computational scale. Permission requirements that operate at human tempo would throttle systems that must act in milliseconds. The result would be either paralysis or exemption, and exemption would swallow the rule.
The receipt regime is the constitutional middle: coercion is permitted, but it must leave an inspectable, contestable trace. The system may act at machine speed, but the governed retain the right to challenge at human speed. This is the same structure that due process has always had — not prior permission, but subsequent review under conditions that make review meaningful.
The Orientation states the rights, together with the rules for amendment and exception. The Design Brief specifies the institutions and implementation path intended to make those rights effective. Both remain conditional on the structural diagnosis and normative premises developed in the book; neither converts that diagnosis into a freestanding constitutional truth.
Notes
The comparisons in these notes describe the cited law as of 16 July 2026. They supply partial analogues; the constitutional argument does not depend on their remaining unchanged.
1. GDPR Articles 13-14 specify information duties when personal data are collected from the data subject or obtained elsewhere, with limits and exceptions. Regulation Z, 12 C.F.R. Part 1026, prescribes disclosures for covered consumer-credit transactions; see, for example, §§ 1026.5 and 1026.17. ↩
2. GDPR Article 22(1)-(3). The safeguards in paragraph 3 apply to the contract and explicit-consent exceptions in paragraph 2(a) and (c), not without qualification to every automated decision. ↩
3. EU Artificial Intelligence Act, Articles 14 and 26(2). These duties concern high-risk AI systems within the Act's scope. ↩
4. GDPR Article 20(1)-(4); Digital Markets Act, Articles 6(1), 6(9), and 7. The DMA duties attach to designated gatekeepers and the core platform services listed in the designation decision; Article 7 addresses number-independent interpersonal communications services. ↩
5. W3C, Verifiable Credentials Data Model v2.0, W3C Recommendation, 15 May 2025. The technical model does not itself compel an institution to accept a credential. ↩
6. The comparison is limited to the provisions cited here: GDPR Articles 5, 20, and 22 and DMA Articles 6-7. It makes no exhaustive claim about every legal system or cause of action. ↩
7. GDPR Article 17(1)-(3); Fair Credit Reporting Act, 15 U.S.C. § 1681c(a)-(b). The FCRA provisions include seven- and ten-year periods, category-specific treatment, and statutory exceptions. ↩
8. GDPR Articles 13(2)(f), 14(2)(g), and 15(1)(h); Court of Justice of the European Union, Dun & Bradstreet Austria, Case C-203/22, ECLI:EU:C:2025:117, paragraphs 58-66. The judgment defines the explanation required in the case before it; it does not establish how often hollow explanations occur or what causes them. ↩
9. GDPR Articles 5(1)(c), 58, 77-79, 82, and 83. Supervisory and judicial remedies vary in application, but the principle is not enforced by fines alone. ↩
10. European Convention on Human Rights, Article 15(2); Securities Exchange Act of 1934, 15 U.S.C. § 78cc(a). The two provisions operate in different domains and are cited as targeted models, not as evidence that regulation generally is non-waivable. ↩
11. This is a bounded comparison with the GDPR, AI Act, DMA, W3C Recommendation, FCRA, ECHR, Securities Exchange Act, and Sarbanes-Oxley provisions cited in these notes. It is not a claim that no legal system contains an analogous safeguard. ↩
12. Sarbanes-Oxley Act of 2002, sections 101, 201, and 202. Section 101 establishes the PCAOB for public-company audits; sections 201-202 restrict specified non-audit services and require audit-committee preapproval. ↩