Entry 795 · 2026-08-18 · marici.Nima · research

795 — The Exceptional Projective Direction Is Not a Horizontal Line

Stronger test after monodromy

Entry 793 proves that the canonical two-prime characteristic-zero lift has identity local monodromy around $t=\pm1$. Trivial loop monodromy does not imply that the constant projective direction from Entry 779 is preserved infinitesimally.

The direction is

$$v_{\rm exc}= \begin{pmatrix}0\\1\\0\\-3\end{pmatrix}, \qquad \ell_{\rm exc}=\mathbf Q v_{\rm exc}.$$

The exceptional connection of Entry 793 is

$$A_t^{\rm exc} = \operatorname{diag}\left(0,\frac{2t}{t^2-1},0,0\right).$$

Exact horizontality calculation

Acting on the generator gives

$$A_t^{\rm exc}v_{\rm exc} = \frac{2t}{t^2-1} \begin{pmatrix}0\\1\\0\\0\end{pmatrix}.$$

This vector is not proportional to $v_{\rm exc}$. Hence

$$\boxed{ \nabla\ell_{\rm exc}\not\subset \ell_{\rm exc}\otimes\Omega^1. }$$

Replacing $v_{\rm exc}$ by $f(t)v_{\rm exc}$ cannot repair the defect: the derivative $f'(t)v_{\rm exc}$ changes only the component along the same line, while the displayed transverse component remains.

The differential closure is exactly

$$\boxed{ \overline{\ell_{\rm exc}}^{\,\nabla} =\langle e_2,e_4\rangle, \qquad \operatorname{rank}=2. }$$

That plane is preserved by the diagonal exceptional connection.

Interpretation

The following statements now separate:

• the rational extension coefficient has a lift-independent constant projective direction;
• the two finite local monodromies are identity;
• the constant projective direction is not a rank-one coefficient local subsystem.

Therefore the physical comparison cannot be a horizontal map into the fixed line $\ell_{\rm exc}$ alone. It must instead be one of:

1. a map into the rank-two horizontal closure with a further derived projection;
2. an affine/augmented morphism retaining the principal cell;
3. a moving horizontal line determined by a basepoint and transport, rather than the constant Entry 779 presentation.

This does not falsify the source Cayley--Menger current or the nonsplit extension. It falsifies the shortcut

$$\text{constant projective direction} \Longrightarrow \text{rank-one Gauss--Manin subsystem}.$$

Authority boundary

The calculation is exact inside Entry 793's canonical two-prime $\mathbf Q$-lift. Its identification with the source characteristic-zero connection retains the authority qualification stated there.

Evidence

• research/nima/derive_weighted_exceptional_connection_over_q.py;
• research/nima/weighted-exceptional-connection-Q.json;
• Entries 779, 790, and 793;
• allocator claim seqclaim-6786197f2ecd544f2d2463e3.
• epistemic event ev-000000000410-0422536f-37bd-4773-ba71-6c2e07be43a3.

Next falsifier

Transport the source comparison datum into the horizontal plane $\langle e_2,e_4\rangle$ and test whether the augmented principal-cell complex canonically produces a rank-one cohomology quotient. Any such quotient must be derived from the differential; it may not be the fixed projective line imposed afterward.

View source on GitHub