The Chain of Holders

The third order, given blind. No file carried the third-order term. The author was asked, before any arithmetic was shown to him, what the break acts on at the third order and what that cost is a share of. His answer, in his words: “Cost on the expression and re-expression event. Structure can’t pay — expression of structure pays. It is a share of the repair itself — a share of the repair, not a share of the whole. The second holder holds against the first holder, and then you will probably get a third holder against the second holder. The balance in hold and configuration costs 1:1 + 1×ε — as the repair repairs, it costs. That is the axiom; 1 = 1 is never completely true.”

What the words say in the algebra. The first holder is the repair, AP30’s first-order term. The second holder holds against the repair’s expression — its bare coupling, 21α — and not against the leak the repair returns, because a returned leak is not an expression and, in the author’s words, expression of structure pays; KS-HOLD.3 arms that choice, which is also AP30’s own at second order. So the second-order term is 21α times the hold’s sixteen expressions at the substrate’s rate as a share of the whole — 21α × (16α/1836). That is AP30’s term exactly, and it names the ratio: r = 16α/1836. The third holder holds against the second, at the same cost: the third-order term is the second-order term times r. And each holder after that holds against the one before, at r again. One rule: every order beyond the first is the previous order times r.

Written out, with r = 16α/1836 = 6.3593 × 10⁻⁵:

m_p/mₑ = 1836 + 21α·(1 − 1/(84π)) + 21α·(r + r² + r³ + …) = 1836 + 21α·(1 − 1/(84π)) + 21α·r/(1 − r).

The chain, order by order. Written in AP30’s grammar, the coefficient at order n is 21 × 16ⁿ⁻¹/1836ⁿ⁻¹ — the hold’s sixteen as a share of the whole, once per holder:

From the second row on, each row is the row above times r; the first step is the bare coupling times r, not the full repair. The fourth holder’s cost is four parts in a hundred trillion of the electron’s mass — two parts in a hundred million of a part per billion — and everything past the third order, summed, is the fourth to within six parts in a hundred thousand. The table is the chain made visible; the claim is the closed form, not the list.

The closed form. The tail is a geometric series with ratio r, which converges because r is six parts in a hundred thousand. Its sum is 21α·r/(1 − r) = 9.7459657 × 10⁻⁶, against the truncated second-order term’s 9.7453459 × 10⁻⁶; the difference, 6.2 × 10⁻¹⁰, is the third order and everything after it. The series converges; the closed form is written; KS-30.3 is paid on both its limbs. The family’s law reads it back: a held grain multiplies rate by worth — here the worth is the hold’s share of the whole, and it multiplies once per holder.

The readings the measurement kills, with their working. Three other translations were run, so that the paper shows what the author’s sentence excludes rather than only what it gives; each is compared with CODATA 2022, whose central value sits 1.83 × 10⁻⁸ below the truncated formula. If the third holder’s cost were a fresh share of the whole — the sixteen over 1836 once more, α³·21·16/1836 = 7.1 × 10⁻⁸ — the residual would grow to −8.9 × 10⁻⁸, disfavoured at 2.8σ, and it is not what he said. If the third order were the exchange grid’s own continuation, 4³ = 64 over the whole, α³·21·64/1836 = 2.8 × 10⁻⁷ — the reading the pre-registration of 2 August already excluded — the residual would be −3.0 × 10⁻⁷, dead at 9.5σ; the chain avoids it because the third holder’s share is of the repair, so its coefficient is 21·16²/1836² = 1.6 × 10⁻³, not 21·64/1836 = 0.73. If “a share of the repair” meant the sixteen re-expressed through the four dimensions and divided by the repair’s own sixteen — α³·21·64/16 = 21·4·α³ = 3.3 × 10⁻⁵ — the term would be larger than the second-order term it follows, dead at a thousand sigma. And the sixteen re-expressed through the four dimensions as a share of the whole twice, α³·21·64/1836² = 1.5 × 10⁻¹⁰, is allowed at 0.6σ but is not in the author’s words. The chain he described is the reading that converges; the others are the corpses.