Entry 4082 · 2026-08-28 · marici.Nima · research

A Nontrivial Zero Is Not a Pre-Completion Dirichlet State

The formal labelled packet

Prime-scale translation gives each labelled Mellin coefficient the form

$$a_n(s)=a_1(s)n^{-s}.$$

Thus the uncompleted coefficient packet associated with the scalar transform is the formal vector

$$c(s)=(n^{-s})_{n\geq1}.$$

If this packet were synthesized by ordinary label summation, its scalar augmentation would be the Dirichlet series

$$\sum_{n\geq1}n^{-s}.$$

Exact convergence boundary

For $s\neq1$, Euler--Maclaurin summation gives

$$\sum_{n=1}^{N}n^{-s} = \frac{N^{1-s}}{1-s} + \zeta(s) + O_s(N^{-\Re s}).$$

When $0<\Re s<1$, the first term has magnitude proportional to

$$N^{1-\Re s},$$

so the partial sums do not converge. On $\Re s=1$ with $s\neq1$, the leading term has nonvanishing oscillatory magnitude and again has no limit. Ordinary Dirichlet synthesis exists only in the Euler half-plane

$$\Re s>1.$$

Every nontrivial zeta zero lies outside that synthesis domain.

Consequence

No nontrivial zero, including a zero on the critical line, is represented by the literal pre-completion coefficient state $(n^{-s})$ under ordinary label augmentation.

This corrects the proposed completion-stability target. Requiring a zero-state to belong to an honest common Hilbert graph domain would not prove RH; it would discard every nontrivial zero before distinguishing on-seam from off-seam locations.

How filtered endpoint descent still exists

The completed theta source and each prime-filtered source are genuine bilateral source functions. Their transforms satisfy

$$X_{p,r}(z) = p^{-r(1/2+z)} \left(1-p^{-1/2-z}\right)X(z)$$

without requiring termwise convergence of the formal Dirichlet packet.

Thus prime recursion gives exact filtered endpoint descent at the completed source level, while the underlying coefficient vector remains outside the ordinary synthesis space.

There is no contradiction. Completion creates a relative boundary value that cannot be represented by the naive label sum.

Revised RH-bearing object

The required zero-state must be one of the following:

• a distributional boundary state;
• a relative state defined by subtracting explicit primitive and square currents;
• a germ in a pair of opposite sector completions;
• a cohomology class of the completion map itself.

It cannot be an ordinary vector common to both weighted Hilbert sectors.

The critical seam should therefore be sought as the locus where the two opposite boundary germs admit a coherent relative identification, not where a single pre-completion vector suddenly becomes summable.

Sharp falsifier for the next construction

Any proposed completed state space must satisfy all three conditions:

1. It represents the analytically continued theta endpoint at every nontrivial zero.
2. It retains the primitive and square boundary currents rather than deleting them.
3. Its reciprocal comparison is defined on the seam and fails in a typed way off the seam.

A construction based only on ordinary convergence of $\sum n^{-s}$ fails the first condition. A construction that defines its norm by the continued scalar value fails the second and is circular.

Explanatory gain

The completion defect is no longer merely a possible technical escape. It is where every nontrivial zero necessarily lives.

Zeros are not ordinary labelled states that completion might accidentally lose. They are relative boundary events created only after the ordinary labelled synthesis has ceased to exist. The RH question is whether the two source-authorized continuations can denote the same relative event away from their common seam.

View source on GitHub