wiki/concepts/comparator

Rendered from markdown source. Open raw source on GitHub.

Comparator

Working definition: the disclosed evaluative standard by which an attention pattern, artifact, action, or outcome is judged.

A comparator names what counts as better, worse, faithful, distorted, loving, harmful, safe, costly, or successful inside a particular frame. It is not the same as a metric: a metric records a signal, while a comparator states how that signal is being interpreted and what it is being used to decide.

Core claim

No attention claim is grounded until its comparator is named.

The comparator is the contract that makes evaluation contestable without pretending that “the field” or “reality” is a self-evident authority. A participant may accept, reject, revise, or fork a comparator; the ability to do so is part of autonomy and consent.

Hidden comparators

A hidden comparator makes a situated judgment feel like fact. It can turn vigilance into care, intervention into protection, compliance into safety, or optimization into success without disclosing whose expectations, risks, time horizon, or protected conditions define the result.

The Attention Engine calls the resulting diagnostic wrong-side error: the system spends effort repairing the visible symptom while the disagreement about frame or standard remains unnamed.

Comparator and expectation set

An expectation set is the implied or explicit set of expectations under which a comparator operates: what should happen, what response is owed, which risks matter, what counts as adequate care, and what would count as repair. The comparator and expectation set should be disclosed together, especially when an action recruits another center’s energy or future possibility.

Related invariants and pages

• P10 — Distinction / Comparator · Quantum Invariants
• C11 — Illegitimate Constitutional Comparator · Quantum Invariants
• Semantic Compression Layer
• Semantic Doppler
• Telic Fields
• Attention
• Consent to Context
• Surface of Mutual Projection
• Attention Engine: Book Series Plan and Semantic Compression
• The Real Is a Ritual

Open questions

• Who has standing to propose, accept, reject, or revise a comparator?
• How should plural comparators be held without collapsing them into one utility function?
• What witness record is sufficient to show that a comparator was disclosed before evaluation began?