Theorem Projectivity of image modules on a reflection-stable set
✓ theorem proved

Projectivity of image modules on a reflection-stable set

Fix a finite-generator Coxeter system, a simple reflection \(s\), and a tail word \(\mathbf t\). Let \(U\subseteq W\) be reflection-stable, so \(sU=U\).

If \(L_{\mathbf t}(U)\) is projective over the symmetric coefficient algebra \(A\), then \(L_{s\mathbf t}(U)\) is projective.

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. Split the old image projection

Write \(N=M_{\mathbf t}\), \(N'=M_{s\mathbf t}\), and \(p:N\twoheadrightarrow L_{\mathbf t}(U)\). Projectivity supplies a linear section \[ \sigma:L_{\mathbf t}(U)\to N,\qquad p\sigma=1. \] Set \(e=\sigma p\). Then \(pe=p\), and \(p(n)=0\) implies \(e(n)=\sigma(0)=0\).

In coordinates, \(e\) preserves \(\pi_x(n)\) for every \(x\in U\).

Step 2. Extend the endomorphism through the tensor step

Define \[ \widetilde e:N'=A\otimes_{A^s}N\to N', \qquad a\otimes n\mapsto a\otimes e(n). \] Since \(e\) is \(A\)-linear, it is linear over \(A^s\), so this map respects tensor relations and is \(A\)-linear.

The tail formulas give \[ \mu(\widetilde e(m))=e(\mu(m)),\qquad \nu(\widetilde e(m))={}^\tau e(\nu(m)). \] For example, on \(a\otimes n\), the underlying second equality is \(\tau(a)e(n)=e(\tau(a)n)\). Additivity extends both identities to every \(m\).

Step 3. Prove preservation of the new image coordinates

For \(x\in U\), reflection stability gives \(sx\in U\). Step 1 says that \(e\) preserves the ordinary-tail projection at \(x\) and the reflected-tail projection at \(sx\).

Step 2 therefore shows that the local pair of \(\widetilde e(m)\) at \(x\) equals that of \(m\). The local-pair coordinate theorem gives \(\pi_x^{s\mathbf t}(\widetilde e(m))=\pi_x^{s\mathbf t}(m)\). Collecting over \(x\in U\) proves \[ p_U'\widetilde e=p_U'. \]

Step 4. Prove that the endomorphism kills the projection kernel

Suppose \(p_U'(m)=0\). Set-tail detection gives \(p_U(\mu(m))=0\) and \(p_{sU}(\nu(m))=0\). Since \(sU=U\), Step 1 kills both tail vectors after applying \(e\).

Step 2 then gives \(\mu(\widetilde e(m))=0\) and \(\nu(\widetilde e(m))=0\). The whole-word tail-detection theorem implies \(\widetilde e(m)=0\).

Step 5. Descend to a split image

The new word module \(N'\) is finite free, hence projective. Its image projection \(p_U'\) is surjective. Steps 3–4 give the two hypotheses of the projective-image splitting theorem for \(\widetilde e\).

That theorem descends \(\widetilde e\) to a section of \(p_U'\). The new image is thus a direct summand of the projective word module, proving projectivity.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.TitsSpace.bottTensorEnd: IntegralPairings.lean:707.

OAI.KLInvariance.TitsSpace.bottTensorEnd_first: IntegralPairings.lean:719.

OAI.KLInvariance.TitsSpace.bottTensorEnd_second: IntegralPairings.lean:725.

OAI.KLInvariance.TitsSpace.bottTensorEnd_set_preserve: IntegralPairings.lean:1264.

OAI.KLInvariance.TitsSpace.bottTensorEnd_set_kills: IntegralPairings.lean:1283.

OAI.KLInvariance.TitsSpace.bottStableUpper_projective: IntegralPairings.lean:1299.

Used by