Kill Switches

KS-4 (AP15), standing. α ≈ 1/137. LIVE — HARD, where AP15 put it and where AP24 took it up. This paper does not move it.

KS-35 and KS-36 (AP24), inherited. Multiplicity and wrong value on ε = f(ε). Unchanged. They fire, if ever, on a computation run under §11.

KS-37 (AP24), standing. Independent variation of the constants: find G varying without c or α varying and the six-face identity is falsified. LIVE — EMPIRICAL. It and P-GLASS.1 are neighbours rather than limbs of one another: KS-37 tests whether the constants vary independently of each other, P-GLASS.1 whether α(0) varies across the record. This paper adds nothing to either.

KS-ALPHA.1 — the side/front switch. Claim: the two views read one cut fraction and differ by exactly one grain per now, the pupil, and by nothing more. Test: structural — if the number of windows firing at one now moves the cut fraction, or if the front-face event coarse-grained differs from the side-view event by anything other than that one grain, α does not belong to the one Actualization State. Status: ARMED. Recovery: §3 falls; §2’s definition survives as a per-window statement; the corpus is unaffected. The switch does not know 137.

KS-ALPHA.1, the empirical limb. In a laboratory the number of windows firing at one now is the local density of matter — how much is actualising in a region, not where the region is. Depth of potential is deliberately not part of the proxy, because potential depth is a function of position and position belongs to P-GLASS.1. Attribution is fixed here, before any measurement rather than after one: one datum does not fire two switches, and a firing is assigned to KS-ALPHA.1 only if the dependence tracks density or multiplicity rather than location in the record. Status of the limb: PROVISIONAL. The obvious comparison — a terrestrial atom against a white-dwarf photosphere, at parts in a hundred thousand — varies density and position together and cannot discriminate, so it fires neither switch as this rule stands. The limb is registered without a design, and it becomes live the day an observation varies density materially at one position in the record. Registered provisionally rather than dropped, because the attribution rule is the thing worth fixing in advance.

KS-ALPHA.2 — the uniqueness obligation. Claim: if a normalisation of the neck is ever constructed under §11, it is unique. Test: two inequivalent normalisations satisfying every stated condition and returning different couplings. Status: REGISTERED, not armed. It is a proof obligation and not a test nature can run, and it stays that way until a normalisation exists to be tested; it is registered so that the obligation is on the record rather than discovered later. Recovery: the picture does not derive α; §10 already says so.

P-GLASS.1 (AP51), cited. No variation of α(0) across the record. §3's invariance across windows at one now is what P-GLASS.1 needs and this paper adds nothing to its firing rule.