Does the Differential roster have a seat for the friction itself, or only for its explanation?

Status: open · laid 2026-09-14 · type: structural

Statement. Every seat in the roster accepts the case's account of what is wrong and disputes only the explanation. Ledger grades the bookkeeping, Hypothesis checks the tool preconditions, Lateral tests the analogy, Framing contests the object and the form of the question — but all four work inside the premise that something is missing. No seat's axis is the friction itself: none is charged with asking whether there is a problem here at all.

The instance. In PN-001 (CASE-08), premise 6 — that a Lefschetz-trace argument needs a positivity input at all — was flagged load-bearing in the audit and then went two full rounds untested while four seats argued about where the positivity would come from. House caught it, late, only as test 5. If the trace formula already constrains the spectral configuration without any positivity input, the sign was never the missing object: the entire friction section is misdiagnosed, and with it all four accounts and both House rounds. The panel spent two rounds downstream of a premise no seat was assigned to doubt.

Why this is structural, not a slip. Apply the scope-statement method to the roster — what is this a list of? It is a list of the ways this project has been wrong before: Hypothesis is the unstated-preconditions finding, Lateral is the mechanism test, Bound is the four bounding mechanisms, Ledger is compute-it-twice. That is a defensible scope, and it has a consequence: a roster built from past failure modes is strong on recurrence and weak on novelty by construction. "The friction is itself the artefact" is not a past failure mode the roster was distilled from, so no seat carries it. The gap is a property of how the roster was assembled, not of one bad night.

Two candidate fixes, undecided. 1. Extend Framing's axis from is this the right object? to is there a problem here at all? — giving one seat standing to reject the friction rather than reframe it. 2. Leave it with House, since House did catch it.

Leaning toward (1): House caught it only after two rounds had already been spent downstream of the untested premise, so catching it at synthesis is catching it late. But not decided — extending a seat's axis is a roster change, and the roster earns its shape from recurrence, not from one panel.

What would settle it: whether it recurs. One instance is one instance. If a second panel argues downstream of an untested friction premise — a load-bearing "there is a problem of kind X" that no seat is assigned to doubt — then the roster has a hole and fix (1) is warranted. If the next panels surface their friction premises on their own, PN-001 was a bad night, not a missing seat. The decision is therefore deferred to the second panel, and a TODO.md item points here so it surfaces when one runs.

Cross-refs. PN-001 (the instance) · docs/procedure-differential-tool.md §10.2 (the roster this probe questions). Companion to the OPS Watching instance "a load-bearing premise can go untested while the panel argues downstream of it."