The Gold Standard of Anthropology | Proxy Record Migration, Legacy Forms and Preserving Authority after System Changes

1. Define the authority record before moving it

Back to contents

A proxy record is more than a contact pair. It links an appointor, a proxy holder, a governing instrument, a covered body or meeting, decision scope, instructions, execution or authentication evidence, receipt, duration, revocation and later use. Before migration, define which fields establish authority and which are merely administrative aids.

Separate source evidence from derived status. A scanned form and verified electronic submission are source records. “Active”, “valid” or “may vote” is a conclusion calculated under rules that can change. If the new system stores only the conclusion, later reviewers cannot reconstruct why it was reached.

The anthropological issue is institutional memory. A proxy is a social relation made operational by forms, procedures and expectations. Migration should preserve the relation’s documented boundaries rather than treating the old system’s vocabulary as self-explanatory.

First-principles check: Model the evidence chain before mapping database columns. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Draw one complete proxy record from appointment through meeting use and revocation. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Move from a paper binder to a membership platform with only appointor, holder and status fields. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Keep the record definition, governing sources, field meanings, derived rules and ownership. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

2. Inventory legacy forms, versions and hidden conventions

Back to contents

Legacy repositories often contain several generations of form. A checkbox may change meaning, a field may be optional in one year and mandatory in another, and staff may encode exceptions in notes or colours. Start with a versioned inventory of paper forms, scans, spreadsheets, portals and meeting-specific exports.

Sample ordinary and difficult records: handwritten amendments, partial instructions, alternates, revocations, late receipts, renewed appointments and forms whose meeting has been adjourned. Interview staff about conventions, but verify those memories against documents and outcomes. A remembered shortcut is evidence about practice, not automatically the governing rule.

Record absence explicitly. A blank field can mean not asked, not applicable, unknown, unreadable or intentionally unrestricted. Converting every blank to no destroys these distinctions and can change authority.

First-principles check: Treat each form version and convention as a separate source context. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Build a legacy-form register with dates, fields, codes, known exceptions and sample counts. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Discover a colour-coded spreadsheet whose colours are lost in CSV export. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Retain originals, screenshots, codebooks, interviews, exception samples and unresolved meanings. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

3. Map meaning, not matching field names

Back to contents

Schema mapping should begin with semantics. “Expiry” may mean the last valid day, a review date or a meeting date. “Representative” may mean proxy holder, household respondent or staff contact. A same-named field can differ; differently named fields can represent the same event.

For every target field, state its source, transformation, allowed values, cardinality and loss. Where one old record becomes several scoped records, preserve the parent relationship. Where several legacy fields become one target field, retain the components or a reversible transformation log.

Do not force incompatible content into an apparently convenient field. Use a controlled exception store or retain the legacy attachment until an authorised interpretation is made. A clean target schema is not a justification for erasing inconvenient meaning.

First-principles check: Require a written semantic equivalence for every migrated field. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Map legacy fields called delegate, proxy, alternate and authorised voter into a target authority model. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Use one old free-text scope note that cannot be represented by target checkboxes. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Keep source and target dictionaries, mappings, transformations, losses, exceptions and reviewers. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

4. Preserve originals, provenance and fixity

Back to contents

Migration normally creates a new representation of an old record. Preserve the original form or export where retention and privacy rules permit, and link it to the migrated representation. Record source system, item identifier, extraction time, transformation version and responsible operator.

Use fixity checks so that files can be shown to remain unchanged after transfer. A checksum does not prove the record was valid when created, but it can demonstrate that the preserved bitstream matches the captured source. Access controls and retention schedules continue to apply; preservation is not an instruction to retain personal data forever.

Provenance should distinguish source, migration action and later correction. If a reviewer fixes a date after migration, the record must not imply that the corrected value came directly from the legacy system.

First-principles check: Make every migrated value traceable to a source and transformation. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Design a provenance row for one scanned proxy and its extracted fields. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Repeat after optical character recognition misreads a handwritten date. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Retain originals, identifiers, checksums, extraction logs, transformation code, correction history and access controls. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

5. Reconstruct authority instead of copying active flags

Back to contents

An active flag is often a cached decision. Recalculate validity in the new system from appointment, receipt, scope, duration, revocation and meeting rules. This may reveal records that were active only because the old platform lacked a relevant event or used a superseded rule.

Run the old and new decision logic side by side on a frozen test set. Categorise disagreements: data defect, mapping loss, rule change, date interpretation, manual override or genuine ambiguity. Do not resolve every disagreement in favour of the newer system merely because it is newer.

Where the governing rule changed, separate migration from policy. First reproduce the historical record faithfully; then apply the authorised new rule prospectively or through a documented review. Otherwise a technical project silently rewrites institutional decisions.

First-principles check: Derive current authority from preserved evidence under an identified rule version. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Recalculate one hundred legacy active records and explain every status mismatch. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Introduce a new maximum duration that begins after the migration date. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Keep old status, recalculated status, rule versions, disagreement category, decision and approver. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

6. Worked example: one proxy becomes three records

Back to contents

A legacy spreadsheet contains one row: Mina appointed Farid, active, annual meeting, notes “finance only; alternate Lian; renewed 4/5”. The new platform has separate records for appointment, scope, renewal and alternate authority. A hurried import creates one unrestricted standing appointment for Farid and lists Lian as a second proxy.

The defensible migration returns to the signed form and renewal message. Farid is authorised only for finance items at the named annual meeting and its permitted adjournment. The renewal evidence extends the applicable window but does not broaden scope. Lian is an alternate who may act only if the triggering condition occurs; Lian is not a simultaneous co-holder.

The migrated package therefore contains linked records with source citations and one unresolved question about the renewal date format. Until that date is interpreted under a stated locale, the record is quarantined from automatic voting eligibility. Clean data waits for authority rather than inventing it.

First-principles check: Decompose compound legacy notes into scoped, linked claims. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Build the target records and show which evidence supports each field. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Change 4/5 from day-month to month-day and test the meeting result. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Preserve the row, original instruments, locale assumption, linked records, quarantine and review. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

7. Handle unreadable, incomplete and orphaned records

Back to contents

Some legacy records cannot be made complete. The appointor may not match the current member register, a scan may be unreadable, the form version may be unknown or the meeting identifier may no longer exist. These are not ordinary nulls; they are authority risks.

Create reason-coded exception states such as unreadable source, missing attachment, unmatched person, ambiguous scope and unresolved date. Route each state to a proportionate process: rescan, identity reconciliation, authorised interpretation, fresh appointment or archival retention without operational use.

Never repair an orphan by attaching it to the most similar current person or meeting. Similarity can help generate candidates for review, but authority requires verified linkage. The system should favour a visible unresolved state over a falsely confident match.

First-principles check: Keep uncertainty operationally visible until linkage is verified. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Triage ten exception records into repair, re-authorisation, archival-only or rejection routes. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Use two members with the same name and no legacy identifier. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Retain exception code, candidate matches, evidence, reviewer, decision and notification. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

8. Test transformed records against real decisions

Back to contents

Field-level reconciliation is necessary but insufficient. The migrated system must answer the decisions it exists to support: who may vote at a meeting, on which item, with which instruction, at what time and subject to which revocation or alternate condition.

Build decision tests from historical cases and synthetic edge cases. Compare old records, source instruments, migrated output and expected decision. Include boundary dates, overlapping appointments, meeting adjournments, narrow instructions and late revocation. Test both inclusion and exclusion; a migration that never denies authority may be dangerously permissive.

Use independent reviewers and record false grants separately from false denials. Both matter, but they can have different consequences and remediation. Accuracy percentages should not hide a small number of high-impact authority changes.

First-principles check: Validate the migrated system at the level of authorised action. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Create a twenty-case acceptance suite with expected voting outcomes. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Add a special meeting sharing the annual meeting’s name but not its scope. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Keep test cases, expected and actual decisions, severity, repairs, retest and sign-off. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

9. Protect privacy, access and retention during migration

Back to contents

Proxy records can contain signatures, contact details, membership status and political or organisational preferences. Extraction files, test copies and vendor sandboxes can multiply exposure. Define least-privilege access, secure transfer, approved storage, deletion and breach response before copying data.

Retain only what authority, audit and applicable law require. A preservation copy may need restricted access while an operational system uses structured fields. Redaction, tokenisation or synthetic test data can reduce exposure, but they must not remove the features needed to test difficult mappings.

Do not use migration as an opportunity to enrich records from unrelated sources without authority. More data can make matching easier while expanding surveillance and changing the relationship people originally consented to.

First-principles check: Preserve evidential sufficiency with the least necessary exposure. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Map each migration environment, user role, data class, retention period and deletion proof. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Use an external vendor for extraction validation. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Keep privacy assessment, access logs, contracts, transfer records, deletion evidence and incidents. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

10. Audit cutover, rollback and post-migration corrections

Back to contents

Cutover creates a period when old and new systems can diverge. Freeze or synchronise writes, define the authoritative source at each time and reconcile transactions made during the transition. A proxy received after the final export but before the new system opens must not disappear between systems.

Set measurable acceptance criteria and a rollback plan. Rollback should restore both data and operational authority, not merely restart software. After launch, log every correction with its source and reason, and monitor unusual increases in unresolved or manually overridden records.

Close the migration only after reconciliation, exception review, user training, access verification and evidence that temporary copies were handled as planned. Successful import is an intermediate state, not project completion.

First-principles check: Control the interval in which two systems could claim authority. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Write a cutover ledger for exports, new receipts, freezes, reconciliation and final handover. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: The new system fails after five meeting updates have been entered. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Archive cutover plan, timestamps, deltas, rollback test, reconciliations, corrections and closure approval. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

11. Frequently asked questions

Back to contents

Should every old field be migrated? Only when it has evidential, operational, retention or historical value; omission must be authorised and documented. Is a PDF enough? It can preserve the source, but structured decisions still require reliable extraction and linkage. Can active records simply be re-entered? Not without reconstructing scope and provenance.

What if the legacy rule was wrong? Preserve what happened, identify the defect and apply an authorised correction process. Can machine learning match people or classify notes? It may prioritise review, but uncertain outputs should not create authority without verification. When is migration complete? When decisions, provenance, security, exceptions and reconciliation meet the declared acceptance criteria.

First-principles check: Answer every migration shortcut with the authority and evidence it would change. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Prepare an operator FAQ with examples of permissible and impermissible repair. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Use a record that is historically important but no longer operational. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Maintain reviewed answers, rule owners, examples, escalation routes and revision date. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.

12. The gold standard preserves bounded delegated voice

Back to contents

A strong migration does not make legacy records look modern at the cost of their meaning. It inventories form histories, maps semantics, preserves source evidence and provenance, recalculates derived status and quarantines ambiguity instead of guessing.

It tests real authority decisions, protects personal data and controls cutover so that no appointment vanishes or silently expands. Corrections remain distinguishable from source values, and policy changes remain distinguishable from technical transformations.

The final test is human and institutional: after the system changes, can a reviewer still explain who delegated what, under which rule, for which decision and on what evidence? If so, modernisation has preserved authority rather than merely moving data.

First-principles check: Require continuity of meaning, evidence and bounded authority across the system change. Write the answer before looking at the preferred outcome. Then identify the authority, observation or calculation that could change it. This keeps the method tied to evidence rather than hindsight.

Worked practice: Audit one migrated record from original form to a current meeting decision. Keep the intermediate steps visible. A reader should be able to locate the input, apply the same rule and understand why a different result would follow if one fact changed.

Changed-context transfer: Repeat after the organisation changes platform again five years later. Preserve the underlying principle but replace the setting-specific details. State which parts travel, which do not, and which new authority or measurement would be needed.

Evidence record: Keep the final inventory, mappings, sources, tests, exceptions, approvals, corrections and lessons. Retain uncertainty and disagreement instead of compressing them into a clean label. The record should let another person reconstruct the decision without relying on the memory or status of the original decision-maker.

Decision-quality drill: Run the practice once under the most favourable plausible assumptions and again under the least favourable plausible assumptions. Compare which conclusion changes, which merely becomes less precise and which survives. If the result depends on an unstated convention, make that convention visible before acting. This is a transfer test: the method should guide a new case, not only explain the example after the answer is known.

Practical verification: Give the case file to a reviewer who did not build the analysis. Ask that person to reproduce the critical step, identify the stopping rule and name the strongest live alternative explanation. Resolve differences by returning to the governing evidence, not by averaging opinions. Record corrections, version the decision and keep a short list of conditions that would require reopening it.