Theorem The head-kernel maps raise degree by one
✓ theorem proved

The head-kernel maps raise degree by one

Fix a finite-generator Coxeter system and a finite basis of its extended realization, giving the degree-one symmetric coefficient grading. Let \(\mathbf t\) be a word, \(s\) a simple reflection and \(x\in W\). Use the stalks \(X=M_{\mathbf t}^x\), \(Y=M_{\mathbf t}^{sx}\), \(Z=M_{s\mathbf t}^x\), their head kernels \(K_0,K_1\), and the maps \(\kappa_0,\kappa_1\).

In the ascent case \(\ell(x)<\ell(sx)\), \[ \kappa_0(Y_n)\subseteq(K_0)_{n+1}. \] In the descent case \(\ell(sx)<\ell(x)\), \[ \kappa_1(X_n)\subseteq(K_1)_{n+1}. \] In each case the inverse lowers degree by one. Consequently, \[ Y_{n-1}\simeq(K_0)_n\quad\text{in the ascent case}, \qquad X_{n-1}\simeq(K_1)_n\quad\text{in the descent case}, \] as real vector spaces.

Relation to the OpenAI notes

Degree-one kernels and shift conventions.

OpenAI notes: The normalization proof temporarily assigns degree two to linear forms and uses \(M(k)^j=M^{j+k}\) for its shifted self-dual object. bg:normalization.

Lean (primary): The word construction keeps linear forms in degree one. Each kernel equivalence sends an old degree-\(n\) vector to degree \(n+1\). The local module-data convention is \(M(1)_n=M_{n-1}\). GradedModules:1244, BottSamelson:1618, BottSamelson:2084, BottSamelson:1744.

Role in the proof: The old degree-\((n-1)\) piece contributes to the new degree-\(n\) kernel. Reindexing its Hilbert sum gives the factor \(q\) in the word-character recurrence.

Proof

Step 1. Read the ascent map in evaluation coordinates

Write \(\tau=\chi_s\), \(\alpha=\alpha_s\). The pair-coordinate map takes \((p,q)\) to the values \(p_\eta\) on the zero-bit branch and \(\tau(q_\eta)\) on the one-bit branch.

Thus, for \(y\in Y\), \[ (\kappa_0(y))_{(0,\eta)}=0,\qquad (\kappa_0(y))_{(1,\eta)}=\tau(\alpha y_\eta). \] Only tail indices of the required weight occur in each branch.

Step 2. Check its degree

If \(y\in Y_n\), every stalk coordinate \(y_\eta\) lies in \(A_n\). The root \(\alpha\) has degree one, so \(\alpha y_\eta\in A_{n+1}\). The coefficient action \(\tau\) preserves degrees, hence \(\tau(\alpha y_\eta)\in A_{n+1}\).

The zero coordinates also belong to \(A_{n+1}\). The coordinatewise stalk grading therefore places \(\kappa_0(y)\) in \(Z_{n+1}\), and it lies in \(K_0\) by construction. This proves the ascent assertion.

Step 3. Check the descent map

For \(z\in X\), the pair \((\alpha z,0)\) has new coordinates \[ (\kappa_1(z))_{(0,\eta)}=\alpha z_\eta,\qquad (\kappa_1(z))_{(1,\eta)}=0. \] If \(z\in X_n\), the same root-degree calculation puts every coordinate in \(A_{n+1}\). Thus \(\kappa_1(z)\in(K_1)_{n+1}\).

Step 4. Recover the inverse degree

Each \(\kappa_i\) is a real-linear equivalence between directly graded real vector spaces. The homogeneous component theorem applied to its degree-one map gives \[ \operatorname{pr}_n(\kappa_i(y)) =\kappa_i(\operatorname{pr}_{n-1}(y)). \] If \(\kappa_i(y)\) is homogeneous of degree \(n\), the left side equals \(\kappa_i(y)\). Injectivity then gives \(y=\operatorname{pr}_{n-1}(y)\), so the inverse image has degree \(n-1\).

Therefore the forward and inverse maps restrict to the displayed degreewise isomorphisms. In the source shift convention \(M(1)_n=M_{n-1}\), the kernel has the old stalk shifted upward by one.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.TitsSpace.bottHeadKernelEquiv_eval: BottSamelson.lean:1589.

OAI.KLInvariance.TitsSpace.bottHeadLeft_kernel_homogeneous: BottSamelson.lean:1601.

OAI.KLInvariance.TitsSpace.bottHeadKernelEquiv_homogeneous: BottSamelson.lean:1618.

OAI.KLInvariance.TitsSpace.bottHeadRightKernelEquiv_eval: BottSamelson.lean:2056.

OAI.KLInvariance.TitsSpace.bottHeadRight_kernel_homogeneous: BottSamelson.lean:2071.

OAI.KLInvariance.TitsSpace.bottHeadRightKernelEquiv_homogeneous: BottSamelson.lean:2084.

Used by