Tail maps on coordinate-image modules
Fix a finite-generator Coxeter system with symmetric coefficient algebra \(A\). Prepend a simple reflection \(s\) to a word \(\mathbf t\), and write \(\tau=\chi_s\). Let \[ \mu:M_{s\mathbf t}\to M_{\mathbf t},\qquad \nu:M_{s\mathbf t}\to{}^\tau M_{\mathbf t} \] be the ordinary and reflected tail maps.
For \(U\subseteq W\), write \(p_U'\) for the new word projection to \(L_{s\mathbf t}(U)\) and \(p_U\) for the tail projection to \(L_{\mathbf t}(U)\). The image tail maps are \[ F_U:L_{s\mathbf t}(U)\to L_{\mathbf t}(U), \qquad F_U(p_U'(m))=p_U(\mu(m)), \] and \[ G_U:L_{s\mathbf t}(U)\to{}^\tau L_{\mathbf t}(sU), \qquad G_U(p_U'(m))=p_{sU}(\nu(m)). \] The second formula denotes an element in the twisted target. The source names these maps bottSetFirst and bottSetSecond.
For each \(x\in U\), the local-pair criterion identifies vanishing of the new stalk projection with vanishing of the ordinary tail at \(x\) and the reflected tail at \(sx\). Collecting these conditions over \(U\) gives vanishing of the ordinary tail on \(U\) and the reflected tail on \(sU\). Each displayed composite therefore kills \(\ker p_U'\), and descends through this surjection.
The unit-tensor insertion
The source's bottSeed is \[ n\longmapsto 1\otimes n\in M_{s\mathbf t}. \] Both tail maps return the underlying vector \(n\) on this tensor. Consequently \[ F_U(p_U'(1\otimes n))=p_U(n),\qquad G_U(p_U'(1\otimes n))=p_{sU}(n) \] in their respective ordinary and twisted targets.
These formulas give concrete preimages when studying the first image map and its restrictions to kernels.
Restriction and the first image map
For \(U\subseteq K\), restriction commutes with the first map: \[ F_U r'_{K,U}=r_{K,U}F_K. \] To verify this, lift an image vector to \(m\in M_{s\mathbf t}\). Both sides have coordinates \(\pi_x(\mu(m))\) for \(x\in U\).
The set-tail detection theorem also proves that \((F_U,G_U)\) is injective. Thus the two descended tail coordinates determine each new image vector.
References
The local source declarations below give the construction and proof.
Lean formalization
OAI.KLInvariance.TitsSpace.bottSetFirst: IntegralPairings.lean:673.
OAI.KLInvariance.TitsSpace.bottSetSecond: IntegralPairings.lean:679.
OAI.KLInvariance.TitsSpace.bottSetFirst_project: IntegralPairings.lean:687.
OAI.KLInvariance.TitsSpace.bottSetSecond_project: IntegralPairings.lean:691.
OAI.KLInvariance.TitsSpace.bottSeed: IntegralPairings.lean:913.
OAI.KLInvariance.TitsSpace.bottSeed_first: IntegralPairings.lean:916.
OAI.KLInvariance.TitsSpace.bottSeed_second: IntegralPairings.lean:919.
OAI.KLInvariance.TitsSpace.bottSetFirst_seed: IntegralPairings.lean:922.
OAI.KLInvariance.TitsSpace.bottSetSecond_seed: IntegralPairings.lean:928.
OAI.KLInvariance.TitsSpace.bottSetFirst_restrict: IntegralPairings.lean:948.