OctoInformBeta

OpenAI preprint, September 2026 · §3. Dihedral intervals and transported reflection orders

Lemma 3.6: wawi=wa−iw_a w_i = w_{a-i} for odd aa, and slopes of edge roots order their labels waw_a by 1≺3≺…≺−11 \prec 3 \prec \dots \prec -1, modulo 2ℓP(u)2\ell_P(u) when WPW_P is finite

Root slopes order plane dihedral reflections by signed index.

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 with root space ℝI\mathbb{R}^I, where II indexes SS, and PP a plane in ℝI\mathbb{R}^I spanned by the two outermost positive roots α\alpha and β\beta in it. The reflections whose roots lie in PP generate a dihedral reflection subgroup WPW_P with simple reflections sαs_\alpha and sβs_\beta. The question is in which order the roots of the reflections of WPW_P lie across PP, read from α\alpha to β\beta.

The lemma answers it through a signed numbering of the elements of a dihedral group by the integers: it says how left multiplication by a reflection acts on the numbers, and it gives the order of the roots as an order on the odd numbers, for WPW_P infinite and for WPW_P finite.

Dihedral Coxeter Systems

Let ss and tt be elements of a group GG with s2=1s^2 = 1 and s≠ts \neq t. The dihedral Coxeter matrix of the pair is the Coxeter matrix on a two-element index set whose diagonal entries are 11 and whose off-diagonal entry is the order mm of stst, with m=∞m = \infty when stst has infinite order.✦ If also t2=1t^2 = 1, the pair lift is the group homomorphism from the Coxeter group of this matrix to GG that sends its two simple generators to ss and tt.✦ If moreover s≠1s \neq 1 and t≠1t \neq 1, the dihedral Coxeter system of the pair is the Coxeter system on the subgroup ⟨s,t⟩\langle s,t \rangle generated by ss and tt whose Coxeter matrix is the dihedral Coxeter matrix of the pair and whose simple reflections are ss and tt.✦

Let a Coxeter system have two simple reflections, ss for the first index and tt for the second. The rotation of this Coxeter system is the element ρ=st\rho = st of its group, and the signed dihedral word of an integer ii is the element wi={ρi/2,i even,ρ(i−1)/2s,i odd,w_i = \begin{cases} \rho^{i/2}, & i \text{ even},\\ \rho^{(i-1)/2}\, s, & i \text{ odd}, \end{cases}(1)✦ so that wiw_i is the alternating word sts⋯sts\cdots with ii letters for i≥0i \geq 0, w−iw_{-i} is the alternating word tst⋯tst\cdots with ii letters, and the waw_a with aa odd are the reflections.

The signed order is the relation ≺\prec on ℤ\mathbb{Z} given by a≺b⇔0<a<b or a<b<0 or b<0<a;a \prec b \iff 0 < a < b \ \text{ or } \ a < b < 0 \ \text{ or } \ b < 0 < a ;(2)✦ on the odd integers it is the chain 1≺3≺5≺⋯≺−5≺−3≺−11 \prec 3 \prec 5 \prec \cdots \prec -5 \prec -3 \prec -1.

Roots and Slopes

Let the Coxeter matrix of (W,S)(W,S) be M=(mij)M = (m_{ij}) over a finite index set II, with S={si:i∈I}S = \{s_i : i \in I\}. The Coxeter Gram entry at i,j∈Ii, j \in I is the real number gij=−2cos⁡(π/mij)g_{ij} = -2\cos(\pi/m_{ij}), equal to −2-2 when mij=∞m_{ij} = \infty and to 22 when i=ji = j.✦ The Tits normal functional of index ii is the linear functional ni(η)=∑jgijηjn_i(\eta) = \sum_j g_{ij}\eta_j on the root space ℝI\mathbb{R}^I✦, and the Tits bilinear form is the real bilinear form on ℝI\mathbb{R}^I B(ξ,η)=∑i∈Iξini(η)=∑i,j∈Iξigijηj.B(\xi,\eta) = \sum_{i \in I} \xi_i\, n_i(\eta) = \sum_{i,j \in I} \xi_i\, g_{ij}\, \eta_j .(3)✦ The simple root αi\alpha_i is the coordinate vector of ii in ℝI\mathbb{R}^I.✦

Let VV be a real vector space with a bilinear form BB and vectors αi\alpha_i, i∈Ii \in I, such that B(αi,αj)=gijB(\alpha_i,\alpha_j) = g_{ij}. The geometric simple reflection σi\sigma_i is the linear automorphism of VV σi(ξ)=ξ−B(αi,ξ)αi,\sigma_i(\xi) = \xi - B(\alpha_i,\xi)\,\alpha_i ,(4)✦ and the geometric reflection action is the homomorphism from WW to the linear automorphisms of VV that sends each sis_i to σi\sigma_i.✦ The Tits action of WW on ℝI\mathbb{R}^I, written ξ↦w(ξ)\xi \mapsto w(\xi), is the geometric reflection action for the Tits form and the simple roots.✦

A real root of (W,S)(W,S) 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 real root is a real root whose coordinates are all nonnegative, and Φ+\Phi^+ is the set of these.✦ A vector ξ∈ℝI\xi \in \mathbb{R}^I represents an element r∈Wr \in W when r=wsiw−1r = w s_i w^{-1} and ξ=w(αi)\xi = w(\alpha_i) for some w∈Ww \in W and i∈Ii \in I.✦ The positive root αr\alpha_r of a reflection rr is a positive real root chosen among those that represent rr✦, and the reflection sγs_\gamma of a positive root γ\gamma is the reflection of WW that γ\gamma represents.✦

The height of ξ∈ℝI\xi \in \mathbb{R}^I is ht(ξ)=∑iξi\mathrm{ht}(\xi) = \sum_i \xi_i, a linear functional on ℝI\mathbb{R}^I.✦ The geometric functional is a linear functional ff on ℝI\mathbb{R}^I, fixed once and for all, such that f(γ)/ht(γ)f(\gamma)/\mathrm{ht}(\gamma) takes distinct values at distinct positive real roots γ\gamma✦, and the geometric slope of γ∈Φ+\gamma \in \Phi^+ is μ(γ)=f(γ)ht(γ).\mu(\gamma) = \frac{f(\gamma)}{\mathrm{ht}(\gamma)} .(5)✦

Plane Subgroups

The root plane of two positive roots γ\gamma and δ\delta is the subspace of ℝI\mathbb{R}^I spanned by γ\gamma and δ\delta.✦ For a subspace PP of ℝI\mathbb{R}^I, the plane reflection subgroup WPW_P is the subgroup of WW generated by the elements represented by a real root lying in PP.✦

A plane angular basis of PP is a pair (α,β)(\alpha,\beta) of distinct positive roots, its left and right endpoints, whose root plane is PP and such that every positive root γ\gamma in PP satisfies μ(α)≤μ(γ)≤μ(β).\mu(\alpha) \leq \mu(\gamma) \leq \mu(\beta).(6)✦ The plane dihedral matrix of a plane angular basis (α,β)(\alpha,\beta) of PP is the dihedral Coxeter matrix of the pair sα,sβs_\alpha, s_\beta. The Coxeter system on WPW_P determined by (α,β)(\alpha,\beta) is the Coxeter system (WP,{sα,sβ})(W_P,\{s_\alpha,s_\beta\}) whose Coxeter matrix is the plane dihedral matrix of (α,β)(\alpha,\beta) and whose simple reflections are sαs_\alpha, first, and sβs_\beta, second.✦ The length function of the Coxeter system (WP,{sα,sβ})(W_P,\{s_\alpha,s_\beta\}) is denoted ℓP\ell_P, and its signed words wiw_i are those of (1) with s=sαs = s_\alpha and t=sβt = s_\beta.

An edge x→yx \to y of the Bruhat graph of a Coxeter system joins two of its elements with ℓ(x)<ℓ(y)\ell(x) < \ell(y) and y=rxy = rx for a reflection rr; its label is yx−1yx^{-1}.✦ The root of an edge x→yx \to y of the Bruhat graph of (W,S)(W,S) is the positive root αyx−1\alpha_{yx^{-1}} of its label✦, and the root of an edge of the Bruhat graph of (WP,{sα,sβ})(W_P,\{s_\alpha,s_\beta\}) is its root as an edge of the Bruhat graph of (W,S)(W,S).✦

The Claim

Theorem 1. In a Coxeter system with two simple reflections, every odd integer aa and every integer ii satisfy wawi=wa−i.w_a w_i = w_{a-i}.(7)✦ Let (α,β)(\alpha,\beta) be a plane angular basis of a subspace PP of ℝI\mathbb{R}^I, and let x→yx \to y and x′→y′x' \to y' be edges of the Bruhat graph of (WP,{sα,sβ})(W_P,\{s_\alpha,s_\beta\}) whose labels are yx−1=wayx^{-1} = w_a and y′x′−1=wby'x'^{-1} = w_b for odd integers aa and bb. If ℓP\ell_P is unbounded on WPW_P, then μ(αyx−1)<μ(αy′x′−1)⇔a≺b.\mu(\alpha_{yx^{-1}}) < \mu(\alpha_{y'x'^{-1}}) \iff a \prec b .(8)✦ If WPW_P has an element uu of greatest length, then, with cmod⁡nc \bmod n the least nonnegative residue of cc modulo nn, μ(αyx−1)<μ(αy′x′−1)⇔amod⁡2ℓP(u)<bmod⁡2ℓP(u).\mu(\alpha_{yx^{-1}}) < \mu(\alpha_{y'x'^{-1}}) \iff a \bmod 2\ell_P(u) < b \bmod 2\ell_P(u) .(9)✦

In the paper, the result reads: if WPW_P is a maximal plane subgroup with canonical generators sαs_\alpha and sβs_\beta, positive roots α\alpha and β\beta, and κ=−B(α,β)/2≥0\kappa = -B(\alpha,\beta)/2 \geq 0, then for WPW_P of type I2(m)I_2(m) one has κ=cos⁡(π/m)\kappa = \cos(\pi/m) and the angular order from α\alpha to β\beta lists the indices aa of the reflections waw_a as 1≺3≺⋯≺2m−11 \prec 3 \prec \cdots \prec 2m-1 modulo 2m2m, for WPW_P infinite one has κ≥1\kappa \geq 1 and the order is 1≺3≺5≺⋯≺−5≺−3≺−11 \prec 3 \prec 5 \prec \cdots \prec -5 \prec -3 \prec -1, and the geometric construction gives this order or its reverse.

What Is Checked

The claim is stated over the signed words of a Coxeter system with two simple reflections, the signed order, and, for a plane angular basis (α,β)(\alpha,\beta) of a subspace PP, the Coxeter system on WPW_P, the edges of its Bruhat graph and the geometric slopes of their roots, which rest on the Tits form, the Tits action, the positive real roots and the geometric functional.

The library proves that the Tits form takes the values B(αi,αj)=gijB(\alpha_i,\alpha_j) = g_{ij} on simple roots✦, and that for any vectors αi\alpha_i with these values B(αi,αi)=2B(\alpha_i,\alpha_i) = 2✦ and the geometric simple reflections satisfy the Coxeter relations (σiσj)mij=1(\sigma_i\sigma_j)^{m_{ij}} = 1 for finite mijm_{ij}, so that si↦σis_i \mapsto \sigma_i extends to a homomorphism on WW.✦ It proves that some linear functional ff makes f(γ)/ht(γ)f(\gamma)/\mathrm{ht}(\gamma) injective on Φ+\Phi^+✦, and that the label of an edge of the Bruhat graph of (W,S)(W,S) is a reflection.✦ For ss and tt in a group with s2=t2=1s^2 = t^2 = 1 and s≠ts \neq t, the image of the pair lift is ⟨s,t⟩\langle s,t\rangle✦; for a plane angular basis (α,β)(\alpha,\beta), the reflections sαs_\alpha and sβs_\beta are distinct✦ and generate WPW_P.