OATH / AN EXPLORABLE PROPOSAL
Continuous candidate integrity & public accountability
OATH is a discussion proposal for voluntary candidate participation and public records that invite scrutiny of interests, commitments, interactions, and official decisions. Its preferred high-assurance direction includes continuing public financial histories and near-24/7 candidate-worn video, subject to unresolved privacy, completeness, governance, legal, and electoral-effect questions.
Open the evidence lab ↗Start with the idea ↓01 / THE PROPOSITION
The goal is to make credible, public-minded candidates easier to recognise and more able to compete without wealth or patronage. Whether this mechanism achieves that is a hypothesis to test.
A candidate voluntarily takes the Compact.
Records, sources and gaps become inspectable.
Voters judge the evidence and the candidate’s policies.
Test recruitment and electoral effects over time.
An intended causal chain, not an established effect. Every link needs testing. Source §2.2
Candidates may join before political institutions authorise reform. The proposal seeks low personal cost, does not require a bond as its foundation, and does not formally disqualify wealth; declining remains a choice, not evidence of dishonesty.
Source §1.1OATH is meant to examine disclosure, conflicts, commitments, and responses to evidence. Policy agreement, ordinary error, and competence remain separate questions.
Source §3.1The proposal favors ongoing capture and prompt recording over periodic assurances. Accepted records, corrections, provenance, and challenges should remain inspectable, while missing, delayed, excluded, or unverified intervals stay visible as limits on what can be answered.
Source §3.3; §3.4The accepted direction is historically queryable candidate finance and near-24/7 candidate-worn video, with provenance and protection for third parties. The scope and proof of coverage remain open.
Source §1.1A voluntary no-reward covenant is a preferred interim idea, while a disclosure tail and other approaches remain under review. The duration, scope, and enforceability of any approach are unsettled.
Source §4OATH can preserve evidence of conduct, omissions, gaps, and responses. It cannot establish honest intent, prove every real-world event was captured, or guarantee future conduct.
Source §3.2Family responsibilities, safety, confidentiality and practical burden can all make an honest person decline. A credible standard must measure who it excludes as well as whom it attracts.
Source §13 / §1602 / THE EVIDENCE LAB
Choose a question. Inspect the relevant entries. Then change who holds the history and see exactly which assurance changes.
This changes the hypothetical custody arrangement. It cannot create missing evidence.
08:00 UTC · declared
The candidate lists two accounts. No claim is made that these are every account or beneficial interest.
Candidate submission
Source §6.1 / §10.309:00 UTC · corroborated
A fictional institution-sourced feed records a payment from Northbank to the declared campaign account. Its purpose and any connection to a decision remain unestablished.
Illustrative institution feed
Source §8.510:00 UTC · documented
The candidate publishes: “I will oppose the Northbank permit unless an independent review supports it.” This is a conditional commitment, not an unconditional promise.
Public statement
Source §6.1 / §5.2 C-0811:00 UTC · unobserved
The demonstration has no authenticated video for this interval. A calendar entry says “constituent meeting”. That explanation is a claim; the gap proves neither innocence nor misconduct.
Gap notice + candidate calendar
Source §7.3 / §8.412:00 UTC · sealed
The record contains an illustrative commitment to a sealed segment. The public cannot infer its contents from a digest. A relevant release would need a governed review.
Illustrative custody manifest
Source §6.4 / §9.613:40 UTC · documented
An official record shows support for the permit, citing a published review. Readers must examine whether that review satisfies the stated condition; a payment nearby in time does not settle intent.
Illustrative official record
Source §3.1 / §7.415:00 UTC · corrected
An erroneous meeting time is corrected by adding a new record. The previous assertion remains inspectable. Correction is not itself evidence of intentional deception.
Linked correction
Source §5.2 C-03The recorded payment and decision are visible. A causal bargain is not established.
Known: A payment appears in one declared account’s feed. The commitment and later official decision can be inspected together.
Unknown: Other accounts, intermediaries, purpose, private arrangements and intent are outside this answer.
Next: Examine the source transaction, stated condition, independent review and any additional corroboration. Give the affected parties a chance to respond.
This record cannot answer that question for the unobserved interval.
Known: The gap is visible. A separate sealed segment exists, but its contents are not publicly available.
Unknown: A later segment cannot fill the earlier gap. A calendar entry is not independent verification of what happened.
Next: Record a specific, material question; seek corroboration or a narrowly scoped evidence review. Preserve the gap and protect unrelated third parties.
The record allows comparison of a commitment, a decision and a correction.
Known: The original and correction remain visible. A change in position can be examined against the original words and cited reasons.
Unknown: A visible change does not establish dishonesty. Whether the condition was met requires judgement and review.
Next: Distinguish a false statement, a corrected error, a fulfilled condition and a changed judgement before attributing misconduct.
A candidate has voluntarily joined under a stated Compact and has a recorded participation period.
Participation does not establish motive, character, complete disclosure, or future conduct; refusal does not establish guilt.
Source §1.1; §2.2A bounded claim about a specific property of the records or process, such as who operated a registry or whether a query covers a committed index.
An assurance tier is not a guarantee of honesty, full real-world completeness, or electoral success.
Source §9.1The current operating condition of a stream or service, such as current, delayed, interrupted, or suspended.
Operational status is not an assurance tier and does not by itself describe a person’s conduct.
Source §9.1A conclusion reached after evidence and competing explanations receive the defined review; its scope and appeal state remain visible.
An allegation, automated flag, correlation, or unexplained gap is not by itself a finding of intent or misconduct.
Source §7.4; §14.2A candidate’s substantive political position or decision, which voters may assess through ordinary democratic judgment.
Policy agreement is not an OATH integrity test, and an unpopular or mistaken decision is not automatically evidence of corruption.
Source §3.103 / PRESSURE TESTS
A gap, an allegation, a reward years later. Follow the choices and their consequences, compare the alternatives, and go back to take another path.
Fictional walkthrough. Explore how a candidate and registry could record a capture gap near a confidential meeting without treating the gap as evidence of what happened.
Source §6.4; §7.2; §15Imagine a candidate has a scheduled meeting with a constituent who needs confidentiality. The camera record shows an unverified interval nearby; it does not show why the interval exists or what happened during it.
The chosen review can compare only evidence that was retained and sources that are available. The constituent’s presence does not waive their privacy.
The candidate has chosen to publish a limited note. The next choice affects how much context is shared and how clearly its source is labeled.
In this fictional path, device and timestamp evidence is reviewed and preserved with its limits. The gap remains a gap, and no account of the meeting is inferred from it.
Outcome: The public can inspect the review scope and the resulting technical record.
Still unresolved: The review does not establish what occurred in the room, whether the capture field was adequate, or whether every relevant source was available.
In this fictional path, any adjacent retained footage remains sealed while a person may decide whether to provide limited context. No one is required to participate; missing footage is not recovered by this process.
Outcome: The review can proceed only with evidence that is actually offered and within the agreed scope.
Still unresolved: If no context is offered, the gap remains unresolved; that choice does not imply anything about the meeting.
In this fictional path, the candidate’s account is retained as an attributed statement, not as independently verified meeting evidence.
Outcome: Readers can see what the candidate reports and compare it with any later evidence or correction.
Still unresolved: The constituent’s account, the missing footage, and the candidate’s full field of view remain unknown.
In this fictional path, the registry records the meeting category and the capture gap while withholding personal details.
Outcome: The known recording limit remains visible with less disclosure about the constituent.
Still unresolved: The reason for the gap and the substance of the meeting remain unobserved.
Fictional walkthrough. Explore provenance and review choices after a small donation appears near an official decision, without inferring influence or guilt from timing alone.
Source §7.2; §7.4; §14.2; §15Imagine a public record shows a small donation near the date of an official decision. The source and context are unclear, and the timing alone does not establish a connection or motive.
The review distinguishes what the payment record says from any claim about why it occurred. Both records may remain incomplete.
The chronology contains dates, not a demonstrated causal relationship. The next choice determines whether more records are attached or interpretation is paused.
In this fictional path, reviewers wait for available source material before describing the payment further.
Outcome: The public can see the record, its provenance gap, and the fact that a review is pending.
Still unresolved: The transaction’s source, any relationship to the decision, and any motive remain undetermined.
In this fictional path, the registry publishes the source disagreement and links the evidence offered by each side.
Outcome: A reader can compare the material rather than treating a single allegation as a finding.
Still unresolved: The public release may still amplify a false association, and neither timing nor a source dispute establishes influence.
In this fictional path, the donation and decision remain separately sourced entries on one chronology.
Outcome: The records can be queried side by side while retaining distinct evidence and status labels.
Still unresolved: Their proximity can still be misread, and the donation’s source and purpose may remain unknown.
In this fictional path, the public sees that a claim was raised and that no reviewed link is established in the available record.
Outcome: A premature interpretation is withheld while the dated records remain preserved.
Still unresolved: No review outcome is guaranteed, and the source or relationship may remain unknown.
Fictional walkthrough. Explore how a former official might record a later consultancy related in subject matter to earlier decisions, while keeping any claim of a quid pro quo unresolved.
Source §11; §14.2; §15Imagine a former officeholder receives a consultancy offer from an organisation affected by a decision made during their term. The offer and timing may merit scrutiny, but they do not by themselves show that a reward was arranged earlier.
The offer and decision timeline are now visible. The proposal leaves the duration and legal effect of any post-office obligation open.
A pause can create room for review, but no private initiative can assume that a covenant is enforceable or that a reviewer can settle motive.
In this fictional path, a voluntary post-office reporting period is recorded as a proposal with a stated scope.
Outcome: Readers can see the offer and the intended period for reporting later roles or benefits.
Still unresolved: No duration or enforceability is assumed, and payments outside the period or through proxies may remain unseen.
In this fictional path, the former official publishes the offer and review record without accepting a broader continuing disclosure tail.
Outcome: The documented offer and its relationship in subject matter to earlier decisions can be inspected.
Still unresolved: The record does not show whether a prior promise existed or whether later benefits will occur through another route.
In this fictional path, the offer remains undecided while a bounded review records available sources and its own limits.
Outcome: The public can see what was reviewed and whether the offer proceeded after the pause.
Still unresolved: A review cannot establish motive from timing alone or rule out an undisclosed arrangement beyond its evidence.
In this fictional path, the former official gives up the disclosed consultancy and records that decision.
Outcome: The proposed role does not proceed under the disclosed offer.
Still unresolved: Declining this offer does not establish that no prior arrangement or other future benefit exists.
04 / BUILD IN STAGES
A modest candidate-operated pilot can begin before a mature witness network exists. Its public claims must match what has actually been tested.
Descriptive phase labels below are an editorial clarification of the source’s provisional names “Proven” and “Full”. They do not certify a person.
Can claim: A candidate has started publishing a time-linked evidence trail. Later changes can be checked where published copies were independently retained.
Cannot claim: Candidate operation cannot independently establish complete capture, complete finances, or candidate-independent custody and control.
Gate: Gate A: show basic device reliability, safe recovery after offline periods, and visible gap history before describing the mode publicly.
Source §9.2Can claim: The candidate cannot silently rewrite evidence already accepted by a separately operated registry; accepted entries can be checked against published checkpoints.
Cannot claim: Without independent monitoring, the registry operator may still omit entries or present different histories. Source completeness is also unresolved.
Gate: Gates B and C: verify exported records and proofs independently, retain visible corrections, and demonstrate candidate-independent hosting and gap notices.
Source §9.3Can claim: External witnesses and monitors can make incompatible revisions of an already committed registry history detectable under stated connectivity and independence assumptions.
Cannot claim: Witnesses do not prove that real-world events or financial accounts were all reported, and a compromised or unavailable witness group remains a risk.
Gate: Gate D: test deliberate forks, witness loss or compromise, and verification without trusting the registry website.
Source §9.4Can claim: For defined supported queries, a verifier may check that results include all matches in a specified committed and indexed corpus.
Cannot claim: This does not establish complete real-world reporting, prove an event did not occur outside the corpus, or settle complex interpretation.
Gate: Gate E: test source-to-index mapping and result proofs against deliberate omissions, while labeling unsupported queries and coverage limits.
Source §9.5Can claim: Committed capture evidence and governed release decisions can be reviewed against independent custody records and published procedures.
Cannot claim: Custody and review cannot guarantee complete surveillance, honest intent, lawful operation in every jurisdiction, or an electoral result.
Gate: Gate F: test both legitimate public-interest review and protection against disclosure forced by a mere hostile allegation.
Source §9.6TEST THE SOCIAL MECHANISM TOO
A proposed pilot should test whether a clearly labeled, candidate-operated evidence trail is usable and understandable before making broader claims. It should measure capture, provenance, candidate burden, public comprehension, and failure handling; a small pilot alone cannot establish broad electoral efficacy or complete real-world coverage.
Source §16.1; §16.2; §16.3; §18Compare public financial histories, institution-source labels, meeting and decision records, and independent challenge procedures without continuous personal video. This deliberately tests a lower-surveillance design against the proposal’s PREFERRED near-24/7 camera direction; it is an editorial comparator, not a replacement decision.
Test: Assess whether candidates and voters find the narrower record useful and acceptable, which questions remain unanswered, and whether reduced intrusion changes participation or verification quality.
Source §1.1; §17 R-09Keep the preferred finance and near-24/7 capture direction, but require independent registry ingestion and checkpoint witnesses before the first participant joins instead of beginning with OATH-Self.
Test: Compare candidate cost, time to launch, operational readiness, and the actual independence gained; determine whether the added entry gate delays learning without resolving source completeness.
Source §9.2; §9.4; §17 R-0705 / COMPACT & LIMITS
The Initiative would set the standard; the Compact is the participant’s pledge; the Registry preserves the evidence. Those roles should not collapse into one unchecked authority.
TEN WORKING COMMITMENTS
Record when participation or capture begins, changes, pauses, or ends, along with the resulting assurance limits.
Proposed Source §5.2Publish the agreed accounts, holdings, liabilities, income, interests, and transactions with source labels and known feed limits.
Proposed Source §5.2Add corrections, withdrawals, recording gaps, and evidence conflicts without silently replacing the earlier record.
Proposed Source §5.2Use the contemplated near-continuous personal capture regime, while recording interruptions and protecting third parties under rules still to be defined.
Proposed Source §5.2The candidate does not control the evidence registry, witnesses, contested-event review, or the rules used to assess their own compliance.
Proposed Source §5.2Preserve evidence-backed questions and competing interpretations. A machine-generated suspicion is not by itself a finding of intent or misconduct.
Proposed Source §5.2Record business arrangements, gifts, offers, advisers, collaborations, and interested-party contacts under a defined protocol; the Compact does not itself require universal divestment.
Proposed Source §5.2Keep a traceable account of public commitments and material decisions where practicable, distinguishing factual claims, judgment, changes of view, and policy disagreement.
Proposed Source §5.2Recognize deferred compensation as an unresolved vulnerability and assess disclosure tails or other mechanisms without presenting them as settled or enforceable.
Proposed Source §5.2Do not describe candidate-run records, witnessed logs, reconciled accounts, and reviewed findings as though they establish the same property.
Proposed Source §5.2Each safeguard leaves a trust dependency. Keep the residual risk visible.
The attack: A participant omits an account, uses a nominee or associate, or routes a material interaction through an unmonitored channel.
Proposed response: Label declared and institution-attested sources separately; reconcile feeds with available ownership records, official records, and evidence-backed challenges.
Still open: No one logged account proves that no other account or informal arrangement exists. Cross-border, proxy, and unsupported sources may remain unknown.
Source §10.3The attack: A device is disabled, substituted, covered, poorly aimed, or made unavailable while relevant conduct occurs outside the recorded field of view.
Proposed response: Publish sequence counters, device health, independent receipts, explicit gap categories, and reviewable exception decisions; preserve sensitive segments under governed custody where available.
Still open: A hash verifies committed bytes, not that a camera was worn, aimed appropriately, or operating throughout an interval.
Source §7.2The attack: An operator hides accepted entries, presents incompatible histories to different audiences, or quietly changes a correction or search result.
Proposed response: Use append-only logs, signed checkpoints, open verification, external mirrors, independent witnesses, and preserved correction events.
Still open: Witnesses may be unavailable or collude, and a log cannot expose an event that was omitted before the registry accepted it.
Source §8.3The attack: A small adversarial transaction, fabricated material, selective edit, or temporal coincidence is framed as proof of improper influence.
Proposed response: Keep suspicion, source disagreement, record contradiction, reviewed finding, appeal, and resolved explanation separate; publish provenance and preserve response rights.
Still open: Review can limit unsupported labels but cannot prevent every rumor, reputational harm, or malicious use of public records.
Source §7.4The attack: A donor, operator, witness, custodian, or reviewer gains enough influence to shape rules, suppress evidence, or favor one candidate.
Proposed response: Disclose funding and conflicts, separate registry functions, version rules publicly, provide open exports and verifiers, and preserve appeal histories.
Still open: Formal separation does not prevent coordinated pressure or capture of several supposedly independent parties.
Source §14.3The attack: A favorable decision is followed later by a consultancy, appointment, payment, or benefit routed through an associate or intermediary.
Proposed response: Assess the proposed interim no-reward covenant, a post-office disclosure tail, or other low-burden approaches, with terms and limits stated openly.
Still open: Rewards may outlast a disclosure period or use proxies. A later benefit alone does not establish an earlier quid pro quo.
Source §1106 / SOURCE & OPEN QUESTIONS
This is a visual interpretation of the supplied OATH working specification, v0.1, prepared 13 October 2026. It presents a proposed social mechanism, not a functioning candidate registry.
The complete original includes the rationale, candidate pledge, threat model, architecture, research agenda and unresolved decisions. Section references throughout this page point to that document.
Download original specification ↓Core identifies the proposer’s objective. Preferred is a favoured direction awaiting validation. Proposed is a design candidate. Open remains unresolved. Limitation bounds what can be claimed.
The diagrams, fictional evidence, scenarios, descriptive phase names and comparator framing are editorial presentation aids. They do not silently amend the original.
Does this standard actually help credible candidates compete, at an acceptable cost to them and everyone around them—and does it make opportunistic concealment harder? A beautiful record is insufficient if the selection mechanism fails.
Source §16 / §17 R-01, R-08, R-09Legal scope, public financial detail, capture exceptions, witness independence, custody, funding, appeal rules and post-office obligations remain decisions for a real pilot. No jurisdiction has been selected or cleared by this presentation.
← Return to the visual model library