Definition Root spans and two-dimensional root planes
✓ definition formalized

Root spans and two-dimensional root planes

Fix a Coxeter system with finitely many simple generators, its original real root space \(V_0=\mathbb R^I\), and its normalized positive roots \(\Phi^+\). Every root is nonzero and has Tits squared norm \(2\).

For \(a,b\in\Phi^+\), define \[ P(a,b)=\operatorname{span}_{\mathbb R}\{a,b\}. \] If \(a=b\), this is the root line \(\mathbb R a\). If \(a\ne b\), it is a two-dimensional root plane.

Independence of distinct positive roots

Suppose \(a=cb\). Since \(b\) has a positive coordinate and \(a\) has nonnegative coordinates, \(c\ge0\). Nonvanishing of \(a\) gives \(c>0\). The norm equation gives \[ 2=B(a,a)=c^2B(b,b)=2c^2, \] so \(c=1\) and \(a=b\). Distinct positive roots are therefore linearly independent, which proves the dimension assertion.

Changing the spanning pair

If \(c,d\) are distinct positive roots in \(P(a,b)\), their span is contained in \(P(a,b)\). Both spans have dimension two, so \[ P(c,d)=P(a,b). \] In particular, for \(a\ne b\), \[ P(a,c)=P(a,b) \quad\Longleftrightarrow\quad c\ne a\text{ and }c\in P(a,b). \]

Two rank-two examples

For type \(A_2\), let \(a,b\) be the simple roots with \[ B(a,a)=B(b,b)=2,\qquad B(a,b)=-1. \] The positive roots are \(a,b,a+b\), and any two distinct ones span the same plane. The restricted form has matrix \(\begin{pmatrix}2&-1\\ -1&2\end{pmatrix}\).

For infinite dihedral type, the simple-root matrix is \(\begin{pmatrix}2&-2\\ -2&2\end{pmatrix}\). The two simple roots still span a two-dimensional vector space. Its restricted form has radical \(\mathbb R(a+b)\). Thus the vector-space dimension and the degeneracy of the restricted bilinear form are separate data.

The induced dihedral construction uses the positive roots contained in this actual span. The extended realization supplies a nondegenerate ambient form when the argument needs to recover a plane from a product of reflections.

References

The local source declarations below define the objects and prove the stated properties.

Lean formalization

OAI.KLInvariance.TitsSpace.rootPlane: ReflectionSchedules.lean:335.

OAI.KLInvariance.TitsSpace.positiveRoots_independent: ReflectionSchedules.lean:640.

OAI.KLInvariance.TitsSpace.rootPlane_finrank: ReflectionSchedules.lean:648.

OAI.KLInvariance.TitsSpace.rootPlane_eq_iff: ReflectionSchedules.lean:669.

Used by