Definition Positive-root inversion sets
✓ definition formalized

Positive-root inversion sets

For \(w\in W\), its root-inversion set is \[ \operatorname{Inv}_{\Phi}(w) =\{a\in\Phi^+:\rho_0(w^{-1})a\text{ is coordinatewise nonpositive}\}. \] The inverse in this definition matches left multiplication by the represented reflection.

For the identity, the set is empty: a positive real root is nonzero and cannot also be nonpositive. For a simple reflection \(s_i\), the set is \(\{\alpha_i\}\), because \(s_i\) makes precisely its own positive root negative.

For each positive root \(a\), its root-inversion indicator is one when \(a\) belongs to this set and zero otherwise. The sign-and-count theorem The root-sign criterion and the inversion count establishes \[ a\in\operatorname{Inv}_{\Phi}(w) \quad\Longleftrightarrow\quad \ell(r_aw)<\ell(w), \qquad |\operatorname{Inv}_{\Phi}(w)|=\ell(w). \] It constructs the bijection to left-inversion reflections through the signed action and a reduced word.

The geometric plane weights use the change of these indicators when \(w\) is multiplied by a reflection. Their finite support follows from this inversion-count result.

References

The local definitions are listed below.

Lean formalization

OAI.KLInvariance.TitsSpace.RootInversions: Tits.lean:2039.

Used by