Stellar subdivision injects a complex's minimal non-faces into those of its refinement
ProvedHirsch.stellar_injects_minimal_nonfacescombinatoricshirsch-conjecturesimplicial-complexes
Let be a finite simplicial complex, a nonempty face of , a new vertex used by no face of , and the stellar subdivision of at . Then there is an injective map from 's minimal non-faces to 's minimal non-faces, given by if and otherwise.
So a single stellar subdivision step never decreases the number of witnessed minimal non-faces — the quantity that must shrink to zero for a complex to become flag. This depends on the corrected IsMinimalNonface in Hirsch_simplicial_complex_flag (restricted to 's own vertex support); without that restriction an unused vertex gives a vacuous minimal non-face of whose image becomes a genuine face of , falsifying injectivity.
Preamble
import Mathlib import Definitions.Def_Hirsch_simplicial_complex_flag
Formal statement
namespace Hirsch
theorem stellar_injects_minimal_nonfaces {V : Type*} [DecidableEq V] [Fintype V]
(K : SComplex V) (σ : Finset V) (hσ : σ ∈ K.faces) (hσne : σ.Nonempty)
(a : V) (ha : ∀ F ∈ K.faces, a ∉ F)
(K' : SComplex V) (hK' : K'.faces = SComplex.stellar K σ a) :
∃ φ : Finset V → Finset V,
(∀ M, K.IsMinimalNonface M → K'.IsMinimalNonface (φ M)) ∧
(∀ M₁ M₂, K.IsMinimalNonface M₁ → K.IsMinimalNonface M₂ → φ M₁ = φ M₂ → M₁ = M₂) := by sorry
end HirschSource
Campaign research notes (2026-09-13/17), Prove2Me mission 'The Polynomial Hirsch Conjecture'; plans/attempt_thin.md + plans/referee_thin.md (cyclic-polar family, referee-verified SOUND); plans/attempt_flag.md + plans/referee_flag.md (flag/stellar subdivision, referee-verified SOUND) (Lemma 10 of attempt_flag.md, referee_flag.md confirms SOUND)