Definition

The Layer 0 precondition step that requires the operator to name, before running a read, which of the known bias positions might be influencing the read already in progress. Parent-enabling step for [Bias Prosecution] — the formal discipline that names, logs, evaluates, and adjudicates the bias-checking activity every operator is already running informally.

Explanation

The move [Declaration] makes is small but structurally important: it requires the bias check to happen before the read is trusted, not after the decision is made and the operator is looking to justify it. An operator who declares “I might be running [Golden Rule Bias] on this one” before evaluating a Guest Experience decision has already changed the conditions of the evaluation — the bias is now visible and named, rather than running invisibly underneath a read the operator believes is neutral.

This is why [Declaration] sits at Layer 0, ahead of the substantive read itself, rather than as a check performed afterward. A bias caught after the decision has already been reached tends to get rationalized away — the sunk cost of having already decided creates pressure to defend the decision rather than revisit it. A bias declared before the read starts has no decision yet to defend, which makes genuine self-correction far more likely.

[Declaration] is the step that enables [Bias Prosecution] to function as a discipline rather than an occasional lucky catch. Without it, the operator would need to somehow notice his own bias unprompted, in real time, while also running the substantive decision — a much less reliable ask than declaring the risk up front and then checking the read against it.