Summary

  • RIR policy archives preserve proposals, versions, statuses and discussions, but an earlier acceptance or withdrawal does not by itself reveal the principle that should govern a later dispute.
  • Informal precedent already operates through staff experience, chair memory, repeat entities, policy manuals and familiar phrases. Because that memory is unevenly distributed, refusing to name precedent does not create flexibility; it creates insider advantage.
  • Consistency should attach to reasons, factual predicates and institutional authority rather than to titles or outcomes. A later community may depart from an earlier approach when scarcity, technology, evidence or regional conditions differ, but it should identify the difference.
  • A practical precedent record would link related proposals, preserve decisive objections and implementation evidence, distinguish regional from global authority, and require reasoned explanations for material departures without making past consensus legally binding.

The institution remembers even when it says it does not

Bottom-up policy communities tend to resist the language of precedent. They are not courts, chairs are not judges, and a proposal accepted in 2012 should not bind entities facing a different network in 2026. That resistance protects an important freedom: policy must respond to evidence rather than worship old formulations. Yet the absence of a formal precedent book does not mean earlier decisions have no authority.

They enter discussion through memory. A veteran says that the community already rejected compulsory reclamation. A staff member recalls that a similar validation rule proved expensive. A chair treats one objection as resolved because another working group handled it before. An author copies language from an accepted proposal because familiar text appears safer. A Board asks why the current recommendation differs from its treatment of a previous one. Each act uses the past to structure the present.

The problem is not that people remember. Institutional learning would be impossible without analogy. The problem is that memory is uneven, selective and difficult to challenge. Long-serving entities possess context that newcomers cannot derive from a proposal status page. Staff retains operational experience that may never appear in public reasons. Earlier losers remember objections that an archive compresses into “withdrawn.” When precedent remains informal, the most connected entities can invoke it while others cannot test whether the analogy fits. Flexibility becomes privilege unless the community makes its use of the past visible.

An outcome is not a holding

The RIPE archived proposal collection lets readers see whether proposals were accepted or withdrawn and often records a short reason for withdrawal. This is valuable institutional memory. It does not, however, turn a status into a rule. “Accepted” may reflect consensus around a narrow compromise, while “withdrawn” may result from author fatigue, low participation, unresolved drafting or a judgment that further redrafting would not succeed.

The distinction matters because later entities often cite the outcome as if it settled the underlying issue. A withdrawn proposal on personal data does not prove that the community opposed privacy protection. It may show that the proposed scope, consent model or verification design was unacceptable. An accepted temporary-assignment rule does not establish that every future research use deserves the same minimum size. The authoritative learning lies in the reasons and conditions.

Courts develop devices such as holdings, dicta and factual findings to distinguish what a case decided from what it merely discussed. RIR communities need not import that legal structure. They do need an equivalent discipline: what problem was accepted as real, which principle controlled the answer, what compromise was essential, which objections remained, and what evidence later implementation supplied. Without that account, outcomes become slogans. The archive tells people where the journey ended but not which turn must be repeated.

Similarity is a claim that needs proof

Two proposals can share a noun while presenting different governance questions. “Transfer policy” might concern need, inter-regional compatibility, registration accuracy, sanctions, fees or legacy control. “Abuse contact” might concern presence of a field, validation frequency, publication, privacy or consequences of non-response. Treating them as precedents because their titles resemble each other is shallow classification.

A useful analogy identifies the feature that matters. Are both rules responding to the same failure mechanism? Do they affect the same class of holder? Is the resource equally scarce? Does the registry possess the same evidence and legal authority? Are the consequences reversible? Did the earlier outcome depend on a temporary condition? Similarity should be argued, not announced.

The burden belongs initially to the person invoking the precedent. They should cite the exact version and decision, state the principle they draw from it and explain the factual match. Opponents can then identify differences. Chairs should summarise whether the analogy was accepted and why. This simple method prevents a name or proposal number from becoming an insider password.

It also reveals when an apparent inconsistency is legitimate. A stricter verification rule for high-value transfers may coexist with a lighter rule for small routine updates because risk and burden differ. If the distinction is public, unequal treatment can be principled. If it remains unstated, entities reasonably suspect preference or improvisation.

The policy manual remembers only the winners

ARIN describes its Number Resource Policy Manual as the repository for policies successfully developed and adopted through its process. Consolidated manuals are essential for users. They tell an applicant the rule now, without forcing reconstruction from years of proposals. Their strength as current law is also their weakness as institutional memory: they generally preserve the surviving text more clearly than the alternatives, objections and failed experiments behind it.

When a section is amended repeatedly, the current language can look inevitable. Earlier versions may show that a requirement was once narrower, that an exception responded to a specific incident or that a temporary compromise became permanent. Change logs help, but textual diffs alone do not reveal the reasons that carried consensus.

This creates winner's memory. Accepted language remains easy to find and cite; rejected arguments are scattered across lists, meeting transcripts and minutes. A later author may assume an omitted option was never considered. Entities who remember otherwise must prove it through archival archaeology.

A precedent record should sit beside, not inside, the operational manual. The manual remains concise and authoritative. The companion record links each material provision to proposal lineage, decision summaries and implementation reviews. Users who only need the rule are not burdened, while people considering change can see why it exists. Institutional memory should not make everyday compliance harder, but current clarity should not erase the path by which power was chosen.

Withdrawal is weak negative precedent

A withdrawn proposal is often invoked as evidence that “the community did not want” the idea. Sometimes that is fair: sustained, reasoned opposition may have remained after revisions. In other cases, participation was too low, the author left, scope was too broad or a new proposal was thought more suitable. These outcomes do not carry equal weight.

Negative precedent should be graded by the record. A proposal rejected after broad consideration of a stable text provides stronger evidence than one withdrawn before meaningful discussion. A decision based on unresolved operational harm carries forward only while that harm remains. An appeal upholding chairs' consensus judgment tells us about the process and objections, not necessarily the merits of every future design.

The archive should therefore classify closure reasons without forcing complexity into a single code. It can record author withdrawal, chair withdrawal, limited public evidence participation, unresolved material objection, supersession, loss of relevance, duplication or procedural defect. A short narrative should identify the strongest unresolved issue. Later entities can then ask whether a new proposal cures it.

This approach respects failure. It neither treats a withdrawn author as defeated forever nor pretends the discussion had no value. It also stops repeat proposals from beginning at zero. If the same design returns without addressing a documented objection, chairs can require an answer. If circumstances changed, the new author can explain the change rather than deny history.

Acceptance is also weak positive precedent

Consensus around one text does not establish a general licence. An accepted proposal may depend on a narrow scope, safeguards, review clause or particular scarcity condition. Removing those features while citing the earlier acceptance misuses precedent. The later policy must earn its own consensus.

Positive precedent is strongest for process expectations and clearly articulated principles. If a community repeatedly requires public registration of delegated resources because traceability is central to stewardship, a proposal for secrecy should confront that principle. If earlier policies consistently preserve due process before revocation, a new immediate-termination rule needs reasons. The precedent does not make departure impossible; it raises the explanatory burden.

It is weaker for implementation details whose performance changed over time. A documentation method accepted before modern identity fraud or privacy regulation may no longer be proportionate. A block size suited to a replenished pool may be unsuitable after exhaustion. The historical decision shows what the community knew and valued then, not what it must choose now.

The discipline is symmetrical: do not cite acceptance as proof that an expanded version is legitimate, and do not ignore acceptance merely because formal binding force is absent. A mature community can say, “we adopted this principle in a narrower setting; here is why it applies,” or, “we are departing because the factual foundation no longer holds.” Those sentences turn memory into accountable reasoning.

Staff experience is precedent with privileged access

Registry staff applies policy across thousands of cases. It sees recurring ambiguity, applicant behaviour, document failure, fraud patterns and unintended burdens. This experience is indispensable. It also creates a body of practical precedent unavailable to ordinary entities. Two formally similar requests may be treated alike because an internal interpretation was settled years ago, even though no public document explains the rule.

Policy-experience reports are one way to surface that knowledge. ARIN's published materials identify reports and customer feedback as sources for gaps, discrepancies and provisions not operating as expected. RIPE NCC impact analyses similarly place operational evidence before the community. These practices should be systematic rather than occasional.

Staff should publish recurring interpretations that materially affect eligibility, burden or remedy, with anonymised examples. It should distinguish a faithful reading from a policy gap. Where case handling creates a durable rule, the issue belongs before the community. Confidentiality and anti-abuse concerns may limit detail, but not the existence and effect of the interpretation.

Otherwise precedent flows in one direction. Staff can invoke accumulated experience to resist a proposal, while authors cannot examine the cases or ask whether different handling would work. Public experience reporting makes operational memory contestable. It does not devalue expertise; it converts expertise from assertion into evidence.

Chairs create precedent through classification

Chairs decide whether objections are material, whether a new version cures them, whether consensus exists and whether a proposal should return to discussion. Even without formal opinions, these decisions teach entities what arguments count. If one chair treats economic concentration as a valid stewardship concern while another recognises only routing objections, later proposals face different constitutional environments.

Some variation is inevitable. Working groups have different subjects and histories. The risk arises when classifications are not explained. A brief consensus summary should identify categories of objection and why they were addressed, outweighed or outside scope. If an earlier chair decision is cited, the summary should explain its relevance rather than simply invoking consistency.

Chair collectives can compare practices without centralising regional policy. They might maintain guidance on reasoning quality, not substantive outcomes: how to describe unresolved objections, how to handle repeated proposals and how to distinguish low participation from opposition. Such guidance makes process precedent visible while preserving each community's authority.

Appeals also generate learning. An appeal decision should state the procedural principle applied and its limits. It should avoid unnecessary commentary on the policy merits, which could be mistaken for future doctrine. Reasoned chair decisions are not binding judgments; they are public evidence of how entrusted discretion was used.

Boards create a different kind of precedent

In systems where a Board adopts, remands or rejects community-developed policy, its decisions can influence later expectations. The ARIN PDP directs the Board to review the history and compliance of recommended policies. A remand based on legal conflict or process defect differs from disagreement with policy wisdom. If reasons are unclear, later entities cannot know the boundary.

Board precedent should be limited to the Board's legitimate role. A prior decision that a proposal conflicted with law may guide future legal review. It should not become a general power to substitute director preference for community judgment. A remand for inadequate consultation may establish a useful procedural expectation, provided comparable cases receive comparable scrutiny.

Minutes should identify the authority and standard. Was the text unimplementable, inconsistent with mission, procedurally defective or insufficiently supported? What would cure the problem? If confidentiality limits detail, the Board can still state the category and operative consequence.

Election legitimacy is implicated because members choose many directors. Voters should be able to see whether a Board applies its standards consistently and respects the division between fiduciary oversight and policy authorship. An unexplained departure from a prior adoption practice may be justified, but it should become part of electoral accountability rather than disappear as institutional discretion.

Cross-regional comparison is not horizontal binding law

The NRO Comparative Policy Overview has long offered a structured view of policies across all five regions while warning that it is a reference, not an official policy statement, and may be overtaken by later implementation. This is exactly the right starting posture. Regional differences are evidence and experiments, not commands.

A community considering transfer needs, assignment sizes or out-of-region use can learn from another region's design and outcomes. It should not assume that adoption elsewhere proves local legitimacy. Membership structure, scarcity, legal context, market behaviour and operating capacity differ. Conversely, “our region is different” should not end analysis. The speaker should identify the relevant difference.

Comparative precedent is most useful when the mechanism is shared. If one region's validation method produced high false-positive rates, another should examine that evidence. If a rule depends on a unique statute, the analogy weakens. The record should state what travelled: principle, text, implementation technique or empirical result.

Global policy adds another layer because identical text must pass through regional processes before coordinated action. A region's objection cannot be dismissed merely because others accepted the proposal. The precedent of coordination concerns procedural equality and textual compatibility, not majority rule among registries. Cross-regional learning should widen evidence without reducing any community to a branch office.

Time changes the force of an analogy

Policy precedent decays at different rates. A principle such as public traceability may endure. A factual assumption about free-pool availability can expire quickly. An implementation constraint may vanish after software changes. A legal conclusion may shift with legislation. Archives need dates, but users also need a method for testing continued relevance.

Every precedent citation should identify the temporal predicate. What was true then that made the decision reasonable? Has inventory, technology, participation or market structure changed? An old decision can still be persuasive if its principle survives changed facts. A recent decision can be weak if it addressed an emergency unlikely to recur.

Implementation reviews refresh or weaken precedent. If a policy achieved its purpose without predicted harm, later analogies gain evidence. If it produced exceptions and was amended, citing the original acceptance without that history is misleading. The companion record should link outcome data and subsequent revisions.

Sunsets and temporary rules require special caution. Their acceptance may show only that the community tolerated a measure for a defined interval. Treating them as support for permanence reverses the compromise. Historical context should travel with the citation so that time does not strip away the very limit that made consensus possible.

Scarcity magnifies inconsistency

When number resources were more readily available, inconsistent treatment still mattered, but some errors could be corrected with another allocation. Under scarcity, a rule about queues, transfers or reserved pools can distribute durable economic opportunity. Similar applicants who encounter different interpretations may not be made whole later.

Consistency therefore needs both ex ante guidance and ex post measurement. The registry should publish transition examples, common interpretations and reasons for exceptions. Aggregate case data can show whether particular organisation types receive different outcomes under formally identical rules. Appeals should be able to restore position where timing caused loss.

Precedent is not the only safeguard. Clear text and trained staff matter more. But a visible account of prior interpretation helps applicants predict treatment and challenge unexplained divergence. It also constrains opportunistic analogy by incumbents seeking to preserve advantages created under old scarcity conditions.

The community should be alert to selective history. A proposal may cite an earlier restriction as evidence of stewardship while ignoring later liberalisation, or cite a market-opening rule without its anti-concentration safeguards. A lineage view that shows the full sequence makes cherry-picking harder. Scarcity makes that completeness a fairness requirement rather than an archival luxury.

Emergency decisions should carry a warning label

Emergency policy often creates attractive precedent because it produces decisive action. The institution suspended a rule, imposed a temporary control or allowed an expedited exception. Later advocates point to the act as proof that the institution has authority and the mechanism works. What disappears is the exceptional factual predicate.

An emergency record should state trigger, evidence, duration, scope and compensating safeguards. Any later citation should acknowledge those limits. Authority used to prevent imminent service failure does not automatically support ordinary administrative convenience. A temporary burden accepted under urgent review does not establish proportionality for permanent use.

Boards and chairs should resist “we did it before” where the earlier case was exceptional. They should ask whether the trigger recurs and whether ordinary deliberation is now possible. If a known risk repeats, the precedent may instead show a failure of preparedness: the institution should create an ordinary rule rather than renew emergency discretion.

Warning labels preserve useful learning. The emergency may reveal a technically feasible control or coordination method. Those facts can inform a later proposal. The label prevents feasibility from being mistaken for legitimacy outside the conditions that justified speed.

Silence is not precedent

Low-response calls and quiet final periods frequently appear in policy histories. Silence can support a conclusion that previously established consensus was not disturbed. It cannot reliably establish a general principle for future cases. People may be unaware, fatigued, satisfied, confused or unable to participate.

Later summaries should avoid saying that the community “endorsed” a proposition merely because no one objected. The record can state that no new material objection was submitted during a defined period after earlier discussion. That language accurately describes evidence without manufacturing doctrine.

The same caution applies to unchallenged staff interpretations. Years without a formal complaint may reflect reasonable operation, but may also reflect that affected users did not know an alternative was possible. Before treating practice as precedent, the institution should publish it and invite review.

Precedent needs an affirmative reason that can be examined. Silence lacks content. It may close a stage, but it cannot tell a later community which factual distinction or value controlled. Elevating quiet into authority rewards invisible participation barriers and makes institutional drift hard to reverse.

Repeated language can hide different bargains

Policy authors often reuse accepted text. Familiar language reduces drafting risk and can promote interoperability. Yet the same sentence can sit inside a different bargain. A “demonstrated need” requirement may be tolerable when accompanied by manual review and appeal, but harsh when automated with narrow evidence. A “reasonable period” may mean weeks in one service and years in another.

Textual precedent should therefore include context and implementation. Copying words does not copy legitimacy. The new proposal should explain the function of borrowed language, any changed surrounding provisions and how operational practice will interpret it. Reviewers should examine interactions rather than approve a clause because it previously passed.

Conversely, different words may express the same principle. One region may describe stewardship through conservation, another through fair distribution. Semantic differences should not be exaggerated into constitutional divergence without looking at effect.

A structured comparison can identify problem, mechanism, affected class, safeguard and review. That is more useful than line matching. It helps policy communities reuse genuine learning while avoiding cargo-cult drafting, where phrases survive after the conditions that justified them disappear.

Search design controls which past becomes visible

Archives organised by proposal number and title serve entities who already know what they seek. A newcomer researching validation, legacy resources or revocation may not know the historical vocabulary. If search cannot connect changed terminology, withdrawn proposals and implementation reports, the visible precedents will be the most recent or best remembered rather than the most relevant.

Archive design should support problem-based discovery. Tags can identify affected resource, mechanism, policy section, organisation class and outcome. Synonyms and former section numbers should resolve to the same lineage. Search results should include withdrawn and superseded material clearly labelled, not only current policy.

Accessibility matters. Meeting videos without transcripts, inaccessible scans and broken list links make some reasons practically unavailable. Stable URLs and preservation across site redesigns protect shared memory. Translation of decision summaries can broaden access without pretending every historical discussion can be fully translated.

Search is governance because visibility shapes analogy. An institution that makes accepted proposals easy to find and failed objections difficult to locate creates a systematic bias toward existing power. A balanced archive does not guarantee good reasoning, but it gives every entity a more equal starting point.

A precedent map should show lineage, not hierarchy

The right visual model is a map of related decisions rather than a ladder of authorities. A current provision can link backward to proposals that created and amended it, sideways to regional analogues, and forward to implementation reviews and appeals. Withdrawn alternatives can appear as branches with reasons for closure. Emergency measures and temporary rules can carry distinct markers.

The map should not rank one region above another or suggest that older decisions are inherently stronger. It should expose relationships so entities can make the argument. Each node needs an authoritative version, date, status, problem statement, decisive reasons, unresolved objections and later evidence. Human summaries should link to the full record.

Maintenance belongs to the institution, but substantive interpretation should remain contestable. Staff can curate metadata; chairs can approve decision summaries; entities should be able to request correction where a summary omits a material objection. Corrections should append rather than rewrite history silently.

Such a map would also reveal policy churn and neglected recurrence. If the same problem generates proposals every few years, the institution can ask whether earlier closures were misunderstood or conditions changed. Lineage turns repetition from frustration into evidence.

Reasoned departure is more legitimate than artificial uniformity

Consistency does not require identical outcomes. Institutions become brittle when they follow an old rule after its justification fails. The discipline is reasoned departure: identify the relevant earlier decision, state the principle or expectation it created, explain changed facts or values, and describe safeguards against arbitrary treatment.

Departure may be necessary because technology altered risk, resource scarcity changed distribution, a court clarified obligations, participation broadened or implementation evidence contradicted assumptions. The explanation should be proportionate to the precedent's strength. A broad, recent, reasoned decision deserves more engagement than a lightly discussed withdrawal.

The record should also recognise correction. A community may decide that an earlier outcome was wrong even without changed facts. Bottom-up authority includes the power to reconsider values. Candour is better than inventing a factual distinction to avoid admitting change. The new consensus must still address reliance and transition.

Artificial uniformity can be unfair where circumstances differ. Reasoned departure protects both adaptability and equal treatment because it makes the distinction public. People can disagree with the result, appeal process defects and use the explanation in future debate. Unexplained inconsistency produces only suspicion and insider speculation.

Appeals need a consistency ground without becoming merits review

An appeal system should not overturn policy merely because a prior proposal ended differently. It can, however, review whether chairs or a Board ignored a materially comparable precedent without explanation, relied on a false analogy or applied a process standard unevenly. The remedy may be a reasoned reconsideration rather than adoption of the appellant's preferred outcome.

The appellant should identify the earlier decision and relevant similarity. The decision-maker can then explain distinction, changed evidence or different authority. This keeps the appeal focused. It does not transform archives into binding law or invite endless citation contests.

Consistency review is strongest for procedural promises: notice periods, recusal, treatment of late text and opportunity to respond. Entities reasonably rely on these practices across proposals. Substantive analogies require greater deference because each community must evaluate current evidence.

Appeal decisions should enter the lineage map with limits. If a panel requires chairs to explain a departure, that does not endorse either policy. Clear scope prevents the review body from becoming a shadow policy council while still protecting people from arbitrary procedural variation.

Membership should audit the institution's memory

Members elect directors and fund archives, staff analysis and preservation. They should ask whether institutional memory is complete, accessible and used consistently. This is not a demand that members control open policy discussion. It is oversight of the infrastructure that makes open discussion meaningful.

Annual reporting can identify broken links, untranscribed records, missing decision summaries, unresolved correction requests and recurring interpretations not yet published. Boards can fund remediation and set preservation standards without deciding substantive policy. Candidates can state whether they support reasoned Board decisions and transparent implementation evidence.

The wider community should participate in the audit because non-members create policy and bear its effects. User testing can reveal whether a newcomer can trace a current provision to its history. Researchers can compare summaries with archives. Former proposers and objectors can identify missing context.

Memory is an accountability surface. If only insiders can reconstruct why rules exist, elections and consultations operate with unequal evidence. A Board that maintains accessible history strengthens community authority; one that invokes institutional knowledge while leaving records unusable weakens it.

A practical doctrine of persuasive history

RIR communities do not need stare decisis. They need a modest doctrine of persuasive history. Earlier decisions matter in proportion to the clarity of their reasons, breadth of relevant participation, factual similarity, continued validity and implementation evidence. No single factor makes a result binding.

Anyone invoking history should cite the exact decision and principle. Decision summaries should identify decisive reasons and limits. Chairs should engage material analogies in consensus statements. Boards should state standards when adopting, remanding or rejecting. Staff should publish recurring interpretations and outcome evidence. Archives should link lineage across versions, regions, appeals and later reviews.

Anyone departing should explain why. Changed facts, different scope, stronger evidence, corrected values or distinct authority may justify a new result. Reliance deserves transition analysis where old policy shaped conduct. Silence, title similarity and status labels alone carry little weight.

This doctrine preserves bottom-up freedom because the current community remains sovereign. It improves equality because the past becomes available to more than repeat entities. It improves quality because authors must answer known objections rather than rediscover them. Most importantly, it makes consistency a matter of public reason rather than private recollection.

Conclusion: the past should advise in public

Number-resource policy already has precedent. It lives in manuals, staff practice, chair habits, Board minutes, mailing-list memory and the expectations of people who have watched proposals rise and fall. Denying that influence does not make the system more democratic. It leaves authority dispersed through networks of experience that newcomers cannot inspect.

A precedent book in the judicial sense would be the wrong cure. Regional communities need room to change, experiment and respond to local conditions. The better institution is a connected record of persuasive history: outcomes tied to reasons, reasons tied to facts, facts tested against implementation, and departures openly defended.

With that record, apparently inconsistent results can become intelligible distinctions rather than evidence of favour. Repeated failures can teach rather than vanish. Accepted language can carry its safeguards instead of becoming a talisman. Boards and chairs can be held accountable for the standards they actually use.

The goal is not obedience to the past. It is equal access to the past's arguments. When history advises in public, current consensus can remain free without becoming forgetful, and flexibility can remain principled without becoming arbitrary.

Precedent claims should disclose institutional interest

The person invoking history may have a stake in which history controls. An incumbent holder can emphasise continuity with a rule that protects existing rights. A broker can cite liberal transfer policy while discounting its verification safeguards. Staff may prefer a familiar interpretation because systems and training already reflect it. An author may describe a prior withdrawal narrowly to make a revived proposal look new. None of these interests invalidates the argument, but each helps others evaluate selection.

Decision summaries should record relevant affiliations for authors, chairs and formal reviewers under the community's ordinary disclosure standard. They should also identify when the institution itself benefits from the analogy through lower cost, easier enforcement or reduced legal exposure. Institutional convenience can be a legitimate factor, but it should not masquerade as settled principle.

Disclosure matters more when the historical record is incomplete. A entity with private recollection of an earlier negotiation has an evidentiary advantage. They should distinguish what the public archive supports from personal memory and invite others to correct the account. Chairs should avoid relying on unverifiable recollections in a consensus determination.

The aim is not to police motives. It is to prevent selective memory from acquiring authority through the speaker's status. A precedent argument becomes stronger when its source, factual match and interest are all visible, because entities can test the reason rather than guess at the relationship behind it.

Small communities need rotation and continuity at once

Institutional memory often depends on a small number of people who served as chairs, council members, staff specialists or prolific contributors. Rotation is necessary to prevent permanent control, but abrupt turnover can erase context. The answer cannot be indefinite tenure in the name of historical knowledge.

Handover should transfer records and methods rather than private conclusions. Incoming chairs need a documented account of active proposals, recurring objections, unresolved appeals and relevant prior decisions. They should receive links to public material and a statement of uncertainties. Outgoing chairs may explain history, but the new office-holder must independently assess current evidence.

Mentoring can broaden memory if it includes entities outside the established inner circle. Archive training for new authors and reviewers reduces dependence on personal introductions. Periodic retrospective sessions can examine a policy family without asking the original entities to defend every old choice.

Continuity should reside in durable public reasons; rotation should renew judgment. When communities confuse the two, they either preserve the same people to preserve knowledge or lose knowledge to achieve renewal. A usable precedent record allows both: experienced entities remain valuable witnesses, while no individual becomes the sole custodian of why the institution acts as it does.

Metrics can reveal inconsistent use of history

The institution can audit precedent without scoring policy outcomes. It can measure how often decision summaries cite related proposals, whether cited records are accessible, whether withdrawals include reasons, and whether implementation reviews link back to the authorising decision. It can sample materially similar cases to see whether distinctions were explained.

Metrics should diagnose process, not automate judgment. A high citation count does not prove careful reasoning, and a genuinely novel proposal may need no historical analogy. The useful questions concern completeness and symmetry: were supportive and contrary precedents both considered, were emergency cases labelled, and were superseding decisions included?

Public correction rates can reveal weak summaries. Search logs and user testing can show whether newcomers find relevant history under current terminology. Appeal outcomes can identify recurring failures to explain departure. These measures help Boards fund archives and help chairs improve practice without centralising policy substance.

An annual memory audit would make inconsistency visible before it hardens into distrust. It would also show progress. Governance improves not when every case matches, but when people can reliably discover the earlier case, understand its limits and obtain a reason for difference.

Sources