When written records are lost, a community keeps only what living people still carry, so it should sort its knowledge by how badly it is needed without the book and how easily it can be checked, and keep the dangerous-if-wrong items in several heads. To rebuild a lost record, interview carriers one at a time and separately, write down where they agree, and mark where they differ instead of forcing a single answer.
Records fail by fire, damp, vermin, theft and plain disuse; what survives is what people carry, and recovery depends on how that carried knowledge is gathered and cross-checked.
Paper Fails In Ordinary Ways Long Before A Catastrophe
A community’s records are usually kept in one place, often one copy, because making copies takes labour and materials. That makes every record a single point of failure. Fire, a leaking roof, mould, mice, a flood, or the death of the one person who knew where the book was kept are all ordinary events. After a collapse, replacing a lost record also costs more, because paper and ink are scarce (Paper Making, Carbon Ink).
The same logic applies to knowledge held by a specialist; the corpus treats that case in Knowledge Preservation (governance), and this page does not repeat it.
Sort What Must Live In Heads
Not everything should be carried by memory. Ask three questions of each item:
| Question | If yes | If no |
|---|---|---|
| Will it be needed when the book cannot be fetched (in the field, at night, in an emergency)? | Carry it | Writing is enough |
| Is a small error dangerous or expensive? | Carry it in a fixed, checked form with several carriers, and keep a reference | Free-form telling is acceptable |
| Can it be shown by demonstration? | Teach by doing as well as by telling | Needs words or a diagram |
Items that fail the first question belong on paper alone. Items that pass all three are the community’s core and go into the canon (Community Canon). Exact quantities, tables and drawings are poor material for memory alone; the existing Oral Tradition page lists which content suits each medium.
Method: Rebuilding A Lost Record From People
Step 1: List who might know. Include people who used the record, not only the people who wrote it. A farmer may remember field boundaries the surveyor wrote down.
Step 2: Interview separately. Ask each carrier alone, before they hear anyone else’s version. Group recall pulls everyone toward the loudest or most confident account.
Step 3: Record what each person says in their own words. Note who said it and how sure they were. Do not tidy it yet.
Step 4: Compare. Where independent carriers agree, mark the item as recovered. Where they differ, mark it as disputed, keep both versions, and write who holds which.
Step 5: Check against anything physical. Boundary stones, building dates, tool marks, a surviving fragment, the ground itself. A physical check beats any number of agreeing memories.
Step 6: Write the recovered record at once, in at least two places, and have the original carriers read it back.
Common Mistakes
- Interviewing in a group — the first confident speaker sets everyone’s version.
- Forcing one answer on a disagreement — two honest carriers who differ are information; the difference may mark the one point where memory is least reliable.
- Treating agreement as proof — people who learned an item from the same teacher agree because they share a source, not because they independently confirmed it (Verification & Version Control).
- Leaving the recovered record unwritten — the interviews cost many hours and the carriers will not be around forever.
What’s Next
- Learn next: Memory as Infrastructure — keeping a register of what is carried and by whom
- Combine with: Recording & Backup Systems — so the next loss does not need an interview campaign