Skip to content

Label-Neutral Reporting

A reporting rule — instantiates Self-Fulfilling Prophecy Interruption

Constrains what a record or handoff may say, replacing fixed-trait labels with observed behavior and evidence so downstream readers are not pre-steered.

A label written into a record does not stay a description — it becomes an instruction to every future reader. Label-Neutral Reporting governs the artifact of communication itself: it constrains what a report, case note, or handoff may say, requiring observed behavior, needs, and evidence in place of fixed-trait labels that would steer later treatment more strongly than the evidence warrants. Its defining move is that it acts on the record, upstream of any single evaluation and across many future readers — not on one evaluator's blindness (Blind Review) and not on in-person treatment. By standardizing the report to carry evidence rather than a verdict, it stops the label from arriving pre-loaded at every downstream desk.

Example

In a content-moderation queue, once a user is tagged "troll" in the case notes, every subsequent moderator reads that user's borderline posts as bad-faith and actions them harder — generating more strikes that "prove" the tag and lock the loop. A label-neutral reporting rule changes what the notes may contain: instead of the trait label, each entry must record the specific observed behavior and the exact policy line at issue ("posted X, which may violate rule Y"), with the prior context noted but not editorialized. New moderators now judge each post on its own evidence rather than inheriting a verdict. The manufactured escalation stops, and genuine violations are still fully documented — because the rule bounds the label, not the evidence.

How it works

Its distinguishing lever is the standardized content of records:

  • Identify the steering labels — the fixed-trait words that carry a verdict downstream more strongly than the evidence supports.
  • Set a neutral reporting template — require observed behavior, specific needs or risks, and evidence in place of the label.
  • Bound the label where it must appear — attach expiry dates, confidence notes, or context so a residual label cannot overdetermine treatment.
  • Preserve material risk information — ensure genuinely necessary diagnostic or safety facts are still conveyed, not neutralized away.

Tuning parameters

  • Which labels are bounded — the fixed-trait terms constrained versus the evidence that must replace them.
  • Label expiry and context notes — whether a permitted label carries an expiration, a confidence level, or a context caveat, so it fades rather than compounds.
  • Risk-information floor — how much material risk or diagnostic detail must survive the neutralization; the dial between de-steering and dangerous omission.
  • Scope of records — which reports, handoffs, or systems the rule governs.

When it helps, and when it misleads

Its strength is that it keeps useful information flowing to many readers without the steering charge a label carries — the observation travels, the verdict does not. It is a direct guard against diagnostic overshadowing,[1] where an existing label causes new evidence to be misread or missed.

It misleads, and can do real harm, when neutrality is taken too far and strips information a downstream reader genuinely needs — an over-sanitized handoff that hides a real safety risk is worse than a blunt one. The classic misuse is bureaucratic euphemism that conceals material danger in the name of fairness. The discipline is to bound labels, never evidence: remove the trait verdict, keep every risk-relevant fact.

How it implements the components

  • label_boundary — its core: it constrains how labels, categories, and prior verdicts may be recorded, displayed, and carried forward, with expiry and context notes.
  • neutral_treatment_protocol — it defines the standardized reporting rule (the neutral template a record must follow), a treatment-of-information standard that reduces discretionary label effects.

It does not hide the trigger from one evaluator at a review gate (channel_interruption_choice, Blind Review), standardize in-person treatment behavior (interaction_pattern_change, Neutral Treatment Protocol), or monitor outcomes (outcome_monitor, Outcome Monitoring Review).

Notes

Its twin is Blind Review: both bound a label, but blind review hides identity from one evaluator at one gate, while this rewrites the record so no future reader inherits a verdict. Use this when the loop is carried by durable documentation (case notes, handoffs, files) rather than a single evaluation, and never let neutrality erase a material risk.

References

[1] Diagnostic overshadowing — a documented clinical pattern in which an existing label (often a psychiatric or intellectual-disability diagnosis) causes new or unrelated symptoms to be misattributed to the label and under-investigated. It generalizes to any setting where a recorded label crowds out fresh evidence; label-neutral reporting is the structural counter. ↩