Three of Hearts: Depersonalising Accounts into Studies

Overview

Accounts held as accounts are a set of records about named people, however carefully they were gathered and however good the reason. The collection at the rung below is only defensible because it was bound for this one, and this is where that promise is kept or quietly broken. Two things happen here and they are not the same thing. Identity is stripped, so that what remains cannot be traced back to whoever gave it. And the accounts are synthesised across one another into a study of a society’s conditions, because a single account retold is testimony, and testimony is not a finding.

What is kept from the older material is the anthropology cell’s production of studies on the state of a society, worked up from ethnographic material. What is removed is the machinery the source put around it — the accredited practitioner as the only competent synthesist, and the interpretation of the data into a model of conduct, which is the ninth rung’s work and not this one’s. This technic is graded by its own standard like any other.

How it works

Depersonalisation is not the removal of names, and treating it that way is the most common way this rung fails. People are recognised by combinations — a job, a street, an illness, an unusual event, in a place where few enough people fit all four. So the test is not whether the name is gone but whether someone inside that community could work out who is speaking from what remains. That test can only be applied by people who know the community, which is why the reading is done close, as it was at the collection.

Some accounts will not survive that test at all. In a small community a single account can be unmistakable however it is cut, and the honest answer is that it does not get published in any form. It is set aside, held unpublished, or waits until enough accounts exist for it to be merged into something that no longer points at anyone. A finding is never worth an exposure, and the pressure to make an exception will always arrive attached to the most compelling account in the set.

The second discipline is that the study travels and the accounts do not. What goes up the pipeline to the wing is the study alone. The raw accounts stay in the custody named at collection, and the record of which account belongs to whom stays with that holder and never moves alongside the study. That link has to exist, because a teller’s right to withdraw is worthless if nobody can find their account. It simply never travels. Where the study goes, no key goes with it.

Synthesis then measures conditions rather than virtue, in the manner of Reading a Community’s Morals: the health of connections, the fidelity of transmission, the supply of variation, the distribution of agency, and which failure is gaining ground. What it must not do is tidy itself. The accounts that cut against the emerging pattern are reported as cutting against it and carried into the study intact, because the variation is part of the finding and a synthesis that smooths it has thrown away the evidence that would have corrected it. The study states its counts and its falsifier for the same reason: how many accounts stand behind it, how many were set aside and why, and what would have to be observed for the finding to be wrong.

Excluded by name: no dossiers or profiles, and no composite person presented as a real one — a composite may be used where it is marked as a composite, never otherwise. No description that has quietly become a decree, which is the line the reading rung already draws. And no interpretation of the studies into a model of conduct, which belongs to the ninth rung and to a body that does not exist yet.

Operation

1. Fix the question the study answers before reading the accounts for patterns. A pattern found while looking for one is usually the looker’s.

2. Depersonalise against the community’s own knowledge. Ask whether someone inside it could identify the teller from what remains, not whether the identifying fields have been deleted.

3. Set aside what cannot survive that test. An account that stays identifiable however it is cut is held unpublished or waits for enough others to merge into. It is never published because the finding needed it.

4. Keep the link where it was. The record of which account belongs to whom stays with the holder named at collection and never travels with the study.

5. Synthesise into conditions rather than verdicts, in the manner of Reading a Community’s Morals, and across the body of accounts rather than from any one of them.

6. Carry the accounts that do not fit. Report them as not fitting, in the study itself, rather than resolving them into the pattern.

7. State the counts and the falsifier: how many accounts stand behind the finding, how many were set aside and why, and what would have to be true for the finding to be wrong.

8. Return the draft study to the tellers before it goes anywhere else. A teller who withdraws at this point has their account removed and the study redone without it.

9. Send the study up to the wing and keep the accounts where they are. The study travels; the collection does not.

Cautions & failure modes

The name removed and the person still standing there. Anonymisation that satisfies a checklist and not the community is the most likely way this rung breaks, and the people harmed by it are the ones who trusted the collection enough to speak.

The pattern that eats the exception. Synthesis rewards tidiness, and the account that fits nothing is the easiest to leave out and the one most likely to be right. What was excluded should be visible in the study rather than in the working papers.

The study that argues. A description sliding into a case for something is the failure the reading rung already names, and it arrives here with more force, because a study assembled from real voices carries a weight that an argument would have to earn.

The study that reads beautifully. This is the failure that arrives from doing the work well. A synthesis that accounts for everything in the material and leaves no loose ends has usually been fitted to it, and it will feel like understanding, which is the signal the third proposition warns tracks coherence rather than truth. The falsifier exists for this and should be written by someone willing to see the study fail.

The accounts that travel. The moment raw accounts move with the study — to the wing, to another cell, into any standing collection — the cell has built the pooled record of persons the floor forbids, and it will have done it for a good reason, at someone’s reasonable request, on a single occasion.