This proof establishes the stability of rooted triple relations across various procedural choices, suggesting robustness in mathematical frameworks.
Ebert, Lindeberg, and Hellmuth introduced a saturation procedure for required and forbidden rooted triples and asked whether its returned relation depends on the legal choices made during a run. We prove that it does not: arbitrary completed runs, even with different lengths and different choices, return literally the same final relation. The proof isolates a deactivation lemma yielding a strong diamond property for the exact saturation transition system. The accompanying Lean 4 development formalizes the definitions and confluence theorem. It also formalizes and proves all four clauses of the source paper’s Lemma 5.12, including its realizability statement through a canonical quotient-poset DAG construction. The proof-and-replay archive contains the manuscript and its LaTeX source, the Lean formalization, theorem-interface and axiom checks, an independently written Python semantics, bounded cross-checks and mutation tests, source-version records, and a SHA-256 manifest. From a clean extraction, run: python verify.py The formal declarations use only the standard axioms propext, Classical.choice, and Quot.sound. The confluence theorem itself requires neither finiteness nor non-emptiness of the leaf type; the realizability portion of Lemma 5.12 uses the source paper’s finite, non-empty leaf-set assumptions. This work was conducted within the conjecture-prover and autoformalization project of the Stanford Lean Club, led by Donald D. Poindexter, Jr. OpenAI GPT-5.6 Sol Ultra assisted with problem analysis, proof discovery, and proof development. Claude Code using Claude Opus 5 constructed and iteratively hardened the Lean formalization under adversarial audit. Version 0.1 file correction, 5 August 2026: the deposited files were replaced solely to correct the manuscript’s typesetting and PDF metadata. The mathematical statements, proofs, Lean sources, and claimed results are unchanged. Files and SHA-256 checksums: phylogenetic_saturation_confluence_v0.1.pdfa20c1cb46c4f59788c2132d7c16e81183ead944f01bda5daacc2dfd9c724b032 phylogenetic_saturation_confluence_v0.1_proof_replay.zip3b82c21199c5c048c84b7451ae992cf9c4b3585770f06d2220446d6e9c9c125b DOI: https://doi.org/10.5281/zenodo.21802298 The manuscript and documentation are licensed under CC BY 4.0. Lean and Python source code are licensed under the MIT License.
No takes yet. Share an insight, caveat, or question.
Poindexter, Donald D., Jr (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: