Documentation

DerivedAlgGeo.CategoryTheory.Triangulated.StabilityCondition.Weak.Tilting.TorsionPair.WeakStabilityFunction

The torsion pair of a weak stability function at a slope cutoff #

For a weak stability function W on an abelian category with the Harder--Narasimhan property, and a cutoff μ₀ : WithTop ℝ, the two classes of WeakCutoff.lean

T μ₀ = {E | every weak HN slope of E exceeds μ₀}   F μ₀ = {E | every one is at most μ₀}

are a TorsionPair. Both axioms are now proved:

What took so long, recorded because it is the interesting part #

The Hom-vanishing axiom ported from the strict theory almost verbatim in #789. The splitting did not, and the obstruction was two levels below where it looked: not in Cutoff.lean and not in the torsion-pair assembly, but in Uniqueness/SubobjectLattice.lean, whose pullback_imageSubobject_eq and cokernelPullbackIso were proved from semiClosedUpperHalfPlane_ne_zero — a nonzero object has nonzero charge. That is exactly the implication weak stability drops, and it fails on exactly the object the surface case of Bridgeland's Lemma 6.2 is about. Both statements are true in any abelian category; CategoryTheory/SubobjectCorrespondence.lean proves them without a charge, WeakTail.lean ports the truncation on top of them, and this file is what that unblocks.

Cutoff by slope, not by phase #

The strict pair (TorsionPair/StabilityFunction.lean) is cut by a phase β : ℝ. This one is cut by a slope in WithTop ℝ, which is what lets a rank-zero object — slope ⊤, no phase at all — sit in T μ₀ at every finite cutoff (WeakSlopeData.mem_hnTors_of_rank_zero, #801).

The torsion class at a weak slope cutoff, as an ObjectProperty.

Equations
Instances For

    The torsion-free class at a weak slope cutoff, as an ObjectProperty.

    Equations
    Instances For

      The two classes at a weak slope cutoff are a torsion pair.

      Equations
      • One or more equations did not get rendered due to their size.
      Instances For