A computer-assisted proof of the dimension threshold for turns in a coaxial resting curve
This paper gives a complete, computer-assisted proof of a result first reported in the note "When do new mathematical branches appear?" (10.5281/zenodo.24060223). For one restricted family of symmetric states of an entropic Bures-type gradient flow (10.5281/zenodo.23901246), with real dimension d of at least 4, the curve of resting points has no turns below a single threshold dimension, which lies strictly between 4222.5584838595 and 4222.5584838596, and exactly two turns above it. The proof combines short written analysis with six inequalities certified in exact rational arithmetic: no rounding, no sampling. Three independently written programs agree on every certified inequality, and every step, including the coaxiality lemma that links the curve to all stationary states in whole-number dimensions, was checked by someone other than its author. A related result, that at d = 10,000 there are at least three resting points, is machine-checked by the Coq proof assistant and was confirmed on two machines with two checker versions. To check it yourself: unzip the package, run "bash run_all.sh" (Python 3 and SymPy), and optionally "bash check_coq.sh" (Coq with the Interval and Coquelicot libraries). The README explains each file. Limits, stated plainly: this is computer-assisted mathematics checked inside the author's research group; it has not been peer reviewed, and the threshold theorem itself has not yet been formally checked by a proof assistant. It says nothing about dimensions 2 and 3, about stability, or about any physical system. A failed check on your machine is useful information: please keep the output and tell us. Paper licensed CC BY 4.0; code licensed MIT. Prepared with AI assistance (ChatGPT by OpenAI; Claude by Anthropic). Directed by Gervaseus E. Nerik. Integrity: package SHA-256 64a0e1a1178044c909dc6831b97c0df94a642eeb463698bfc9abb4e36be05308
Authors
- Gervaseus Nerik
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-12
- DOI
- https://doi.org/10.5281/zenodo.24173290
- Primary Topic
- Mathematical and Computational Methods
- Type
- preprint