1 Root pairings
1.1 Definitions
Recall that a perfect pairing is a quadruple \((R,M,N,\mathcal{L})\) where \(R\) is a commutative ring, \(M\) and \(N\) are \(R\)-modules, and \(\mathcal{L}:M\times N\to R\) is a bilinear map such that the induced maps \(M\to \operatorname {Hom}_R(N,R)\) and \(N\to \operatorname {Hom}_R(M,R)\) are isomorphisms of \(R\)-modules. If \(I\) is a set, denote by \(\operatorname {Perm}(I)\) the group of permutations of \(I\).
A root pairing is a tuple \((R,M,N,\mathcal{L},I,\alpha ,\alpha ^\vee ,s)\) where:
\((R,M,N,\mathcal{L})\) is a perfect pairing,
\(I\) is a set,
The maps
\begin{align*} \alpha :& \; I\to M, \quad i\mapsto \alpha _i, \\ \alpha ^\vee :& \; I\to N, \quad i\mapsto \alpha _i^\vee , \\ s:& \; I\to \operatorname {Perm}(I), \quad i\mapsto s_i, \end{align*}are such that
\(\alpha \) and \(\alpha ^\vee \) are injective,
for all \(i\in I\), \(\mathcal{L}(\alpha _i,\alpha _i^\vee )=2\),
for all \(i,j\in I\),
\begin{align*} \alpha _{s_i(j)} & =\alpha _j-\mathcal{L}(\alpha _j,\alpha _i^\vee )\alpha _i,\\ \alpha _{s_i(j)}^\vee & =\alpha _j^\vee -\mathcal{L}(\alpha _i,\alpha _j^\vee )\alpha _i^\vee . \end{align*}
Fix a root pairing \(\Phi =(R,M,N,\mathcal{L},I,\alpha ,\alpha ^\vee ,s)\). We say that \(\Phi \) is:
crystallographic if \(\mathcal{L}(\alpha _i,\alpha _j^\vee )\in \mathbb {Z}\) for all \(i,j\in I\),
reduced if, for all \(i,j\in I\), the roots \(\alpha _i\) and \(\alpha _j\) are linearly dependent over \(R\) if and only if \(\alpha _i=\pm \alpha _j\),
finite if \(I\) is finite.