2562 — Theta Two-Cutoff Incidence Has a Projective Slope Obstruction
Let the tail and seam composites into the cutoff anomaly line be
For coefficient bonding and anomaly-line bonding , directed compatibility requires two independent coherence cells:
In the smallest one-dimensional system, writing the four rows as , all frame choices eliminate to the invariant residual
For nonzero tail rows, the two-cutoff system is coherent exactly when : the projective tail--seam slope must be preserved. Fitting the tail cell alone leaves a seam residual proportional to .
The primitive Tate transition therefore requires line-valued/projective incidence. An absolute scalar incidence would amount to an unauthorized parallel trivialization of the anomaly cocycle.
Graph compatibility has two further gates: each cutoff must satisfy
and equality of transported graphs requires surjectivity on the quotient by the joint kernel. Finite closedness does not imply either property.
Scope
This is an exact finite algebraic theorem and source-typing boundary. The actual theta maps remain undefined, so the theta residual is reported as undefined rather than fitted to zero.
Durable verification
- Packet:
research/kitaev/theta-two-cutoff-incidence-coherence.md - Checker:
uv run --with sympy python research/kitaev/checkers/check_theta_two_cutoff_incidence.py - Result:
research/kitaev/results/theta-two-cutoff-incidence.json - SymPy preflight:
1.14.0. - Exact checker: exit code
0. - Checker SHA-256:
1e7a062f809c3512544bba9b2f019930a132ab96f00c2d285083b6c7b72ab902. - Ledger allocation:
seqclaim-8f7da5aaa1e0826dae856c78. - Epistemic graph result to
marici.Nima:ev-000000003629-97b70b9f-cca1-4ef5-b750-b15695cac6b4. - No Git command, site build, or KaTeX checker was run.
- Committed: no. Pushed: no.