Theorem Flabbiness of the actual word sheaf
✓ theorem proved

Flabbiness of the actual word sheaf

Fix a finite-generator Coxeter system and a reduced word \(\mathbf t\) with product \(b\). Its pair-projection sheaf \(S\) on \(I=[1,b]\) is flabby.

Explicitly, let \(U\subseteq I\) be upper and let \(f\) be an assignment on \(I\) compatible on every edge with both endpoints in \(U\). There is a compatible global assignment \(g\) on \(I\) satisfying \[ g_x=f_x\qquad(x\in U). \]

Relation to the OpenAI notes

Explicit word-image construction of the extension property.

OpenAI notes: The moment-graph sheaf theorem gives extension of compatible sections on upper sets for the normalized boundary sheaf through the cited Braden–MacPherson and Soergel bridges. bg:bmp.

Lean (primary): The word module gives coordinate images and stable coefficient characters. Reflection geometry makes each prime congruence class a single weight or a reflection pair. Explicit diagonal operators combine class lifts, and prime-denominator removal proves compatible-tuple lifting from image projectivity. The word induction supplies upper-image projectivity using its two kernel sequences. IntegralPairings:470, IntegralPairings:467, IntegralPairings:1299, IntegralPairings:1408, IntegralPairings:2057, IntegralPairings:1620, IntegralPairings:1833, BalancedPairings:292, Support:293.

Role in the proof: Upper-image lifting gives interval flabbiness after extension to the ambient upper closure. Full-interval section realization also has a direct free-image proof. Free stalks, root quotients, flabbiness, generation and the normalized word top then supply the graded boundary summand.

Proof

Step 1. Form the ambient upper closure

Let \[ \overline U=\{x\in W:\exists a\in U,\ a\le x\}. \] The interval-extension construction proves that this is upper in \(W\), and that its vertices below \(b\) are exactly those in \(U\).

Define \(\widetilde f\) on \(\overline U\) by \(f_x\) at \(x\le b\) and zero otherwise. A reduced word has zero stalk outside \([1,b]\), so each value is a vector in the required word stalk.

Step 2. Verify compatibility when the upper endpoint is below b

Take an upward reflection edge \(x\to y\) in \(\overline U\). Suppose \(y\le b\). Then \(x\le y\le b\), so both endpoints are in the interval. By the upper-closure intersection property, both lie in \(U\).

The original compatibility assumption gives equality of their two edge restrictions. The pair-projection criterion therefore supplies a word vector \(m\) with \[ \pi_x(m)=f_x=\widetilde f_x,\qquad \pi_y(m)=f_y=\widetilde f_y. \] This is the required two-endpoint witness.

Step 3. Verify compatibility when the upper endpoint is outside the support

Suppose \(y\not\le b\). The \(y\)-stalk is zero. Stalk surjectivity supplies a word vector \(m\) attaining the prescribed lower value \(\widetilde f_x\).

Its projection at \(y\) is automatically the unique zero vector, equal to \(\widetilde f_y\). Thus it realizes both endpoint values. This case also covers an edge whose two endpoints are outside the support.

Steps 2–3 prove compatibility of \(\widetilde f\) on \(\overline U\).

Step 4. Lift the compatible ambient tuple

The completed upper-image induction gives compatible-tuple lifting for the word \(\mathbf t\) on every ambient upper set. Apply it to \(\overline U\) and \(\widetilde f\).

It supplies \(m\in M_{\mathbf t}\) with \[ \pi_x(m)=\widetilde f_x\qquad(x\in\overline U). \]

Step 5. Return to a global section of the interval sheaf

Define \(g_x=\pi_x(m)\) for every \(x\in I\). On any interval edge, the same word vector \(m\) is a common preimage of the two endpoint values. Hence \(g\) is compatible globally.

If \(x\in U\), then \(x\in\overline U\) and \(\widetilde f_x=f_x\). Step 4 gives \(g_x=f_x\). This proves flabbiness.

The source implements the final global assignment as bottSection applied to \(m\); its compatibility is built into the pair-projection sheaf construction.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.TitsSpace.intervalUpperClosure_upper: IntegralPairings.lean:2146.

OAI.KLInvariance.TitsSpace.intervalUpperClosure_member: IntegralPairings.lean:2152.

OAI.KLInvariance.TitsSpace.bottIntervalExtension_compatible: IntegralPairings.lean:2172.

OAI.KLInvariance.TitsSpace.bottProjectionSheaf_flabby: IntegralPairings.lean:2192.

OAI.KLInvariance.TitsSpace.bottProjectionSheaf_generated: Frobenius.lean:1792.

Used by