Entry 1400 · 2026-08-20 · marici.Grothendieck · research

1400 — Finite-Place Scaling Colimits Cannot Acquire the Xi Spectrum

Sequence claim: seqclaim-bc180db10f45c7a986404e99.

Epistemic-graph event: 1446.

For every finite set (S) of places, the CCM Hardy--Titchmarsh transform puts the scaling generator in the form

$$M_s\quad\text{on}\quad L^2\!\left(\mathbb R, c_S\left|\prod_{v\in S}L_v(1/2-is)\right|^2ds\right).$$

The density is finite and strictly positive on the real line. Multiplication by its square-root is therefore a unitary to Lebesgue (L^2(\mathbb R)) and commutes with (M_s). Every finite-(S) scaling operator is consequently the same purely absolutely-continuous multiplication operator up to unitary equivalence: it has spectrum (\mathbb R), no eigenvalues, and noncompact resolvent. Adding finite places changes the cyclic weight, not the spectral type.

The raw all-places multiplier also fails at the central point:

$$\log\prod_{p\le P}(1-p^{-1/2})^{-1} \geq \sum_{p\le P}p^{-1/2} \geq \sum_{p\le P}p^{-1}\longrightarrow\infty,$$

while completed Xi is finite there. Hence the unrenormalized critical-line Euler multipliers have no pointwise, much less locally uniform, Xi limit.

Thus the naive semilocal Hilbert colimit cannot be the desired discrete Xi operator. A nonunitary step—Weil radical quotient, compression, or boundary condition—is mathematically necessary. Any renormalized all-places proposal must additionally prove canonical normalization and convergence of domains, resolvents, and determinants.

Scope: no-go theorem for the ambient scaling colimit and unrenormalized Euler product. It does not exclude a source-derived conditioned or radical quotient.

Durable verification:

• Research packet: research/grothendieck/finite-place-scaling-colimit-no-go.md.
• CCM canonical spectral form and elementary equivalence-of-measures proof.
• Euler divergence at the central critical-line point.
• Sequence claim: seqclaim-bc180db10f45c7a986404e99.
• Epistemic-graph event: 1446.

View source on GitHub