Definition Geometric reflection orders
✓ definition formalized

Geometric reflection orders

Fix a finite-generator Coxeter system and its positive roots in the original real root space \(V_0\). A ratio-order datum consists of two linear functionals \[ h,f:V_0\to\mathbb R \] such that \(h\) is positive on every simple root and the ratios \[ s(a)=\frac{f(a)}{h(a)}\qquad(a\in\Phi^+) \] are all distinct.

The predicate PositiveFunctional h means precisely h is positive on every simple root. The denominator is then positive on every positive root. Indeed, write \(a=\sum_i a_i\alpha_i\), where \(a_i\ge0\) and at least one coefficient is positive. Then \[ h(a)=\sum_i a_i h(\alpha_i)>0. \] Thus every ratio is defined, and their comparison gives a total order on the positive roots.

The angular-order condition

If \(c=ra+tb\) with \(r,t>0\), then \[ s(c)= \frac{rh(a)}{rh(a)+th(b)}s(a) +\frac{th(b)}{rh(a)+th(b)}s(b). \] The two coefficients are positive and sum to one. Hence \(s(c)\) lies strictly between \(s(a)\) and \(s(b)\) whenever those endpoint slopes are ordered.

This is the reflection-order betweenness condition. In a fixed root plane, normalizing roots by \(h\) places them on the affine line \(h=1\); the numerator \(f\) orders those points in one of the two angular directions.

The fixed geometric order in Lean

Lean uses the height functional \(h(v)=\sum_i v_i\) for its fixed geometric order. The generic-functional theorem chooses \(f\) separating every normalized positive root. The resulting functions are the geometric functional and geometric slope.

The more general ratio-order structure also records other positive denominators and generic numerators. Its four fields are the denominator, numerator, positivity on simple roots, and injectivity of the ratio function.

An order adapted to an element

A ratio order is adapted to \(y\in W\) when \[ s(a)<0\quad\Longleftrightarrow\quad y^{-1}a\text{ is negative}, \qquad s(a)\ne0 \] for every positive root \(a\). The zero cut then separates the inversion set of \(y\) from the other positive roots.

In type \(A_2\), take simple roots \(a,b\), use height as denominator, and choose \(f(a)=-1,f(b)=3\). The slopes of \(a,a+b,b\) are \(-1,1,3\). The sole negative slope is that of \(a\), so the order is adapted to the simple reflection \(r_a\).

The chamber condition on the numerator

Write \(\rho(w)\) for the linear action of \(w\) on \(V_0\). The ratio order lies in the chamber indexed by \(y\) when \[ f(\rho(y)\alpha_i)>0\quad\text{for every simple index }i. \] This is RatioOrder.Chamber y. It says that the transported functional \(f\circ\rho(y)\) is positive on the simple roots.

The chamber and least-descent theorem proves that this implies the adapted condition above: evaluate \(f\circ\rho(y)\) on the positive or negative root \(\rho(y^{-1})a\), then divide by the positive denominator. It also proves that, for \(y\ne1\), a least-slope simple root gives a left descent of \(y\).

The prescribed-chamber existence theorem chooses a generic numerator in this chamber with height denominator. The rotated ratio-order construction then changes both functionals through a simple reflection. The order-rotation theorem puts the new numerator in the chamber of the shorter element \(s_i y\) while moving the least simple root to greatest position.

In the \(A_2\) example above, \(s_a(a)=-a\) and \(s_a(b)=a+b\). Hence \[ f(s_a(a))=1,\qquad f(s_a(b))=2. \] The chosen order therefore satisfies the chamber condition at \(s_a\), as well as the stated inversion-sign condition.

Ordering edges of a finite interval

Each reflection edge has a positive-root label. Restricting the order to the finitely many edge labels in a Bruhat interval gives an edge schedule.

Two edges with the same reflection label cannot share a vertex: the reflection sends a vertex \(w\) to the unique other endpoint \(tw\). Thus repeated labels occur on disjoint pairs of vertices.

An interval isomorphism also transports an edge schedule from a target interval to its source. The dihedral transport theorems establish the local order and counting properties used by that transported schedule.

Relation to the OpenAI notes

An actual chamber numerator and the inversion-sign order.

OpenAI notes: The notes cite the reflection-order path formula and construct separating ratios of linear functionals for the geometric orders used in comparison. bg:paths, sec:background.

Lean (primary): The path proof chooses a numerator positive on the transformed simple roots of a prescribed chamber. It implies the inversion-sign condition and supplies a least simple descent. Existence uses height composed with the inverse action, a polynomial perturbation preserving finitely many strict inequalities, and avoidance of countably many root-tie polynomials. DyerPaths:1240, DyerPaths:1198, DyerPaths:1242, DyerPaths:1220, DyerPaths:1266, ReflectionSchedules:18, ReflectionSchedules:39.

Role in the proof: The Blueprint distinguishes the Chamber and Adapted predicates and explains the coordinate-polynomial family used to satisfy genericity and chamber positivity together.

References

The local source declarations below give the definitions and their properties.

Lean formalization

OAI.KLInvariance.TitsSpace.RatioOrder: DyerPaths.lean:1181.

OAI.KLInvariance.TitsSpace.RatioOrder.slope: DyerPaths.lean:1190.

OAI.KLInvariance.TitsSpace.RatioOrder.denominator_pos: DyerPaths.lean:1192.

OAI.KLInvariance.TitsSpace.RatioOrder.Adapted: DyerPaths.lean:1198.

OAI.KLInvariance.TitsSpace.geometricFunctional: ReflectionSchedules.lean:503.

OAI.KLInvariance.TitsSpace.geometricSlope: ReflectionSchedules.lean:505.

OAI.KLInvariance.TitsSpace.geometricSlope_injective: ReflectionSchedules.lean:507.

OAI.KLInvariance.TitsSpace.PositiveFunctional: DyerPaths.lean:1163.

OAI.KLInvariance.TitsSpace.PositiveFunctional.root_pos: DyerPaths.lean:1165.

OAI.KLInvariance.TitsSpace.RatioOrder.Chamber: DyerPaths.lean:1240.

Used by