OctoInformBeta

OpenAI preprint, September 2026 · §2. Real realizations and moment-graph sheaves · §2.3

Geometric reflection orders and normalized R-polynomials: Positive roots ordered by distinct slopes f(α)/h(α)f(\alpha)/h(\alpha), with positive combinations in between, and Rx,y(v2)=vℓ(y)−ℓ(x)R̃x,y(v−v−1)R_{x,y}(v^2)=v^{\ell(y)-\ell(x)}\tilde R_{x,y}(v-v^{-1})

Slopes of positive roots and the normalized R-polynomial.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Let (W,S)(W,S) be a Coxeter system of finite rank, with Coxeter matrix M=(mij)M=(m_{ij}) indexed by a finite set II. This item builds two objects that enter Dyer’s path formula. The first orders the positive roots of the geometric representation of WW by a ratio of two linear functionals, the denominator being the height, the sum of the coordinates of a root. The second is the normalized R-polynomial R̃x,y\tilde R_{x,y}, read off from the inverse of a standard basis element of the Hecke algebra and tied to the R-polynomial Rx,yR_{x,y} by a change of variable.

The Geometric Representation

A realization of MM is a real vector space VV with a bilinear form BB and vectors αi∈V\alpha_i\in V, i∈Ii\in I, such that B(αi,αj)=−2cos⁡(π/mij)B(\alpha_i,\alpha_j)=-2\cos(\pi/m_{ij}) for all i,ji,j, the value being −2-2 when mij=∞m_{ij}=\infty. The geometric simple reflection σi\sigma_i of index ii of a realization (V,B,(αi)i∈I)(V,B,(\alpha_i)_{i\in I}) of MM is the linear automorphism of VV given by σi(v)=v−B(αi,v)αi,\sigma_i(v)=v-B(\alpha_i,v)\,\alpha_i ,(1) and the geometric action of WW on VV for this realization is the homomorphism W→GL(V)W\to GL(V) that sends the simple reflection of index ii to σi\sigma_i.

The Tits space is ℝI\mathbb R^{I}, whose ii-th standard basis vector is the simple root αi\alpha_i.✦ For i,j∈Ii,j\in I the Coxeter Gram entry is gij=−2cos⁡(π/mij)g_{ij}=-2\cos(\pi/m_{ij})✦; the Tits normal functional of index ii is the linear functional ni(v)=∑j∈Igijvjn_i(v)=\sum_{j\in I}g_{ij}v_j on ℝI\mathbb R^{I}✦, and the Tits form is the bilinear form on ℝI\mathbb R^{I} given by B(u,v)=∑i∈Iuini(v)=∑i,j∈Igijuivj.B(u,v)=\sum_{i\in I}u_i\,n_i(v)=\sum_{i,j\in I}g_{ij}\,u_iv_j .(2)✦ The Tits action of WW on ℝI\mathbb R^{I} is the geometric action of the realization (ℝI,B,(αi)i∈I)(\mathbb R^{I},B,(\alpha_i)_{i\in I}).✦

A root of the Tits action is a vector of ℝI\mathbb R^{I} of the form w(αi)w(\alpha_i) with w∈Ww\in W and i∈Ii\in I✦; a positive root is a root whose coordinates are all nonnegative, and Φ+\Phi^{+} is the set of positive roots.✦ The height is the linear functional h(v)=∑i∈Ivih(v)=\sum_{i\in I}v_i on ℝI\mathbb R^{I}.✦ For a linear functional ff on ℝI\mathbb R^{I}, the slope of a positive root α\alpha is f(α)/h(α)f(\alpha)/h(\alpha).

Bruhat Order and the R-polynomials

Let ℓ\ell be the length function of (W,S)(W,S). The Bruhat order ≤\le on WW is the reflexive transitive closure of the relation x→txx\to tx, for x∈Wx\in W and tt a reflection with ℓ(x)<ℓ(tx)\ell(x)<\ell(tx); such a step is an upward edge of the Bruhat graph, labelled by tt.✦ For x,y∈Wx,y\in W the Bruhat interval [x,y][x,y] is the set of z∈Wz\in W with x≤z≤yx\le z\le y✦, and the length difference is d(x,y)=ℓ(y)−ℓ(x)d(x,y)=\ell(y)-\ell(x).✦

The R-polynomials Rx,yR_{x,y} and the Kazhdan-Lusztig polynomials Px,yP_{x,y}, for x,y∈Wx,y\in W, are the pair of families of polynomials in ℤ[q]\mathbb Z[q] selected by the Kazhdan-Lusztig normalization conditions✦: Rx,x=Px,x=1R_{x,x}=P_{x,x}=1; Rx,y=Px,y=0R_{x,y}=P_{x,y}=0 unless x≤yx\le y; 2deg⁡Px,y<d(x,y)2\deg P_{x,y}<d(x,y) when x<yx<y; for every simple reflection ss with ℓ(sy)<ℓ(y)\ell(sy)<\ell(y), Rx,y={Rsx,sy,ℓ(sx)<ℓ(x),(q−1)Rx,sy+qRsx,sy,otherwise;R_{x,y}=\begin{cases}R_{sx,sy}, & \ell(sx)<\ell(x),\\ (q-1)\,R_{x,sy}+q\,R_{sx,sy}, & \text{otherwise;}\end{cases}(3) and, when x≤yx\le y, the reciprocity qd(x,y)Px,y(q−1)=∑z∈[x,y]Rx,zPz,y.q^{d(x,y)}P_{x,y}(q^{-1})=\sum_{z\in[x,y]}R_{x,z}\,P_{z,y}.(4)✦

The Hecke Algebra

Let CC be a commutative ring and a∈Ca\in C. The standard module is the free CC-module ℳ\mathcal M whose basis is formed by the standard basis vectors HwH_w, w∈Ww\in W.✦ For a simple reflection ss, left multiplication by HsH_s in the Hecke algebra with Hs2=1+aHsH_s^2=1+aH_s is the CC-linear operator LsL_s on ℳ\mathcal M given by LsHw={Hsw+aHw,ℓ(sw)<ℓ(w),Hsw,otherwise,L_sH_w=\begin{cases}H_{sw}+a\,H_w, & \ell(sw)<\ell(w),\\ H_{sw}, & \text{otherwise,}\end{cases}(5)✦ and the inverse generator Ls−1L_s^{-1} is the operator Ls−aL_s-a.✦ The word product of a word s1⋯sks_1\cdots s_k in SS is the operator Ls1⋯LskL_{s_1}\cdots L_{s_k}✦, and the standard unit LwL_w of w∈Ww\in W is the word product of a reduced expression of ww.✦ For y∈Wy\in W, the bar image of HyH_y is the vector Hy¯=Ly−1−1H1.\overline{H_y}=L_{y^{-1}}^{-1}H_1 .(6)✦

The Claim

Definition 1. Let Φ+\Phi^{+} be the set of positive roots in the Tits space ℝI\mathbb R^{I} and hh the height. There is a linear functional ff on ℝI\mathbb R^{I} whose slopes separate the positive roots: f(α)h(α)≠f(β)h(β)for all α≠β in Φ+.\frac{f(\alpha)}{h(\alpha)}\neq\frac{f(\beta)}{h(\beta)}\quad\text{for all }\alpha\neq\beta \text{ in }\Phi^{+}.(7)✦ The geometric functional φ\varphi is fixed as one of the functionals given by (7): φ∈{f∈(ℝI)*:f(α)h(α)≠f(β)h(β)for all α≠β in Φ+}.\varphi\in\Bigl\{\,f\in(\mathbb R^{I})^{*} : \frac{f(\alpha)}{h(\alpha)}\neq\frac{f(\beta)}{h(\beta)} \ \text{for all }\alpha\neq\beta\text{ in }\Phi^{+}\Bigr\}.(8)✦ The geometric slope of a positive root α\alpha is sl⁡(α)=φ(α)h(α).\operatorname{sl}(\alpha)=\frac{\varphi(\alpha)}{h(\alpha)}.(9)✦ For every linear functional ff on ℝI\mathbb R^{I}, all positive roots α,β,γ\alpha,\beta,\gamma and all real r,s>0r,s>0 with γ=rα+sβ\gamma=r\alpha+s\beta and f(α)/h(α)<f(β)/h(β)f(\alpha)/h(\alpha)<f(\beta)/h(\beta), f(α)h(α)<f(γ)h(γ)<f(β)h(β).\frac{f(\alpha)}{h(\alpha)}<\frac{f(\gamma)}{h(\gamma)}<\frac{f(\beta)}{h(\beta)}.(10)✦ The normalized R-polynomial R̃x,y∈ℤ[T]\tilde R_{x,y}\in\mathbb Z[T] of x,y∈Wx,y\in W is the coefficient of HxH_x in the bar image of HyH_y, in the standard module over C=ℤ[T]C=\mathbb Z[T] with a=−Ta=-T: Hy¯=∑x∈WR̃x,y(T)Hx.\overline{H_y}=\sum_{x\in W}\tilde R_{x,y}(T)\,H_x .(11)✦ For every field KK, every nonzero v∈Kv\in K and all x,y∈Wx,y\in W, Rx,y(v2)=vℓ(y)−ℓ(x)R̃x,y(v−v−1).R_{x,y}(v^{2})=v^{\ell(y)-\ell(x)}\,\tilde R_{x,y}(v-v^{-1}).(12)✦

In the paper, the result reads: the positive roots are ordered by f(α)/h(α)f(\alpha)/h(\alpha), for hh positive on the simple roots and ff making these ratios distinct, and R̃x,y(T)\tilde R_{x,y}(T) counts label-monotone paths, that is, paths in the Bruhat graph whose labels are monotone for this order, with q−d(x,y)/2Rx,y(q)=R̃x,y(q1/2−q−1/2)q^{-d(x,y)/2}R_{x,y}(q)=\tilde R_{x,y}(q^{1/2}-q^{-1/2}).

What Is Checked

The slopes are stated over the positive roots of the Tits action of a Coxeter system (W,S)(W,S) of finite rank, with the height and arbitrary linear functionals on ℝI\mathbb R^{I}; the normalized R-polynomial is stated over the bar images in the standard module over ℤ[T]\mathbb Z[T], and its comparison with Rx,yR_{x,y} over the R-polynomials selected by the Kazhdan-Lusztig normalization conditions, evaluated in an arbitrary field.

The library proves that Ls−aL_s-a is a two-sided inverse of LsL_s. In every realization it proves B(αi,αi)=2B(\alpha_i,\alpha_i)=2✦ and that the geometric simple reflections σi\sigma_i satisfy the Coxeter relations of MM✦, and on the Tits space it proves that the Tits form takes the values B(αi,αj)=−2cos⁡(π/mij)B(\alpha_i,\alpha_j)=-2\cos(\pi/m_{ij}) on the simple roots.✦