The Zizkian

HomeGuideDownload source

The method

The unit is a reading: a proposed interpretation of a question, observation or arrangement. A reading is not a discovered motive merely because it is memorable.

The operator

  1. Preserve the user’s strongest intended question and desired result.
  2. Name the decision owner and real constraints.
  3. State the strongest ordinary explanation before offering an inversion.
  4. Select a small relevant set of lenses. Each must contribute a distinct question.
  5. Offer a rejectable hypothesis and identify supporting observations separately.
  6. Specify the observation that would defeat the reading and how to obtain it.
  7. Apply the return cut to the analyst’s own framing and possible payoff.
  8. Produce an owned test, a supported decision change, or no finding and closure.

These stages are expected by the human method. The code enforces only the declared record constraints identified in the coverage table.

WWZS? is the provocation mode. It names a question-generating move, not a claim about what the philosopher would literally say. The falsification question is inherited from the project’s earlier conversations.

Invariants

Evidence is not traded against rhetorical force, votes or an attention weight. Participant refusal is not evidence of a hidden motive. Personal exploration requires invitation. A return cut can find nothing. A check can preserve the original brief. A defeated reading cannot remain active in the decision.

Consulting and coaching differ in authority. Business inquiry examines declared constraints, outcomes and observations. Coaching works with the participant’s chosen question and accepted next-step framing. Neither can grant an analyst privileged access to another person’s mind.

What the Prolog layer does

The checker validates a finite declared record. It checks shape and references before evaluating rules. Missing declared support blocks supported; it does not prove support absent from the world. A defeated test declaration on an active reading yields defeated_reading. Explicit untested statuses preserve unknowns.

The program audits a snapshot. It does not interview a participant, read a source, run an experiment or automatically update the epistemic status of a reading.

No compulsory depth

A tutorial can be good instruction. An approval boundary can reflect a real risk. A stalled handoff can reflect understaffing. When the ordinary explanation survives, retain it. Stop recursion when another pass changes no test or decision.

Make the return cut change the next step

The return cut inspects the analyst’s position. It does not refute a reading just because the analyst benefits. Use this worksheet before human review:

  1. In return_cut.statement, name the possible payoff in this case, including prestige, a preferred intervention or the pleasure of discovering a hidden story.
  2. In return_cut.defeater, say what would challenge that account of the analyst. This is a separate claim from the reading’s own defeating observation.
  3. In outcome.reason, explain the consequence: what test, process or action changes because of that risk? If nothing changes, explain why. Record a usable next action and its owner rather than a ceremonial confession.

Fictional example: the analyst prefers redesign. The return cut identifies a possible mandate for more work. The product lead will run the ordinary instruction comparison before deciding on redesign; the analyst commits to retaining the brief if instruction suffices. The analyst’s interest calls for this process constraint, not automatic dismissal of the redesign hypothesis. A participant should judge whether the comparison and ownership are meaningful. Merely entering a second name cannot establish independent judgment. The events, ownership and commitments in this example are fictional declarations, not verified participant agreements.

The checker requires the return-cut fields but cannot judge their sincerity, discrimination or effect on the next step. Every report keeps analyst_payoff in the human-review requirements. That boundary is deliberate and inspectable.

The matrix

The attention matrix records authored priorities across eight dimensions in the historical conceptual appendix. It is outside the working method, not a calibrated operator. The checker never reads it. The essay preserves the Hamiltonian analogy as architectural metaphor, with no claims of physics or numerical diagnosis.

A claim that can fail

The design hypothesis is that these questions improve inspection of a decision record [1 SPECULATIVE | no comparative outcomes]. If that hypothesis is wrong, the earliest signal is additional complexity and suspicion with no stronger evidence, useful test or participant-owned action. See evaluation.