The ascent and descent recursions for word-stalk characters
Fix a finite-generator Coxeter system and a finite basis of its extended realization. For every word \(\mathbf t\), simple reflection \(s\), and weight \(x\), the formal character polynomials satisfy \[ C_x^{s\mathbf t}(q)= \begin{cases} C_x^{\mathbf t}(q)+qC_{sx}^{\mathbf t}(q),&\ell(x)<\ell(sx),\\ C_{sx}^{\mathbf t}(q)+qC_x^{\mathbf t}(q),&\ell(sx)<\ell(x). \end{cases} \]
Relation to the OpenAI notes
Different intermediate objects.
OpenAI notes: The normalization discussion identifies indecomposable boundary-sheaf characters with KL polynomials by the cited character theorem. bg:normalization.
Lean (primary): The word-stalk character is proved to equal a positive Hecke-word coefficient. Its R-convolution is then proved before model decomposition and cancellation. The separate boundary-character identification is exposed as StalkCharacterObligation. IntegralPairings:289, CharacterReciprocity:1134, SheafMorphisms:219, BottSamelson:1924, BottSamelson:1945, IntegralPairings:104, IntegralPairings:120, CharacterReciprocity:915.
Role in the proof: The word formula and R-convolution describe the tensor word. A common decomposition into boundary models then supplies the character identities used in the cancellation argument.
Proof
Step 1. Fix the common coefficient factor
All stalks in the formula use the same coefficient algebra and its fixed \(N\) degree-one variables. For real \(0<t<1\), put \[ D(t)=(1-t)^{-N}. \] This factor is nonzero. The word-stalk Hilbert theorem describes each stalk's convergent dimension sum by its character evaluated at \(t\), multiplied by \(D(t)\).
Step 2. Apply the degreewise ascent identity
Suppose \(\ell(x)<\ell(sx)\). The exact-sequence theorem gives, for every \(n\), \[ f_{s\mathbf t,x}(n) =f_{\mathbf t,x}(n)+f_{\mathbf t,sx}(n-1). \] These three sequences have polynomial numerators \(C_x^{s\mathbf t}\), \(C_x^{\mathbf t}\), and \(C_{sx}^{\mathbf t}\), with the common factor from Step 1.
The Hilbert-numerator shift theorem therefore gives \(C_x^{s\mathbf t}=C_x^{\mathbf t}+qC_{sx}^{\mathbf t}\).
Step 3. Apply the degreewise descent identity
Suppose \(\ell(sx)<\ell(x)\). The dimension identity is now \[ f_{s\mathbf t,x}(n) =f_{\mathbf t,sx}(n)+f_{\mathbf t,x}(n-1). \] Apply the same numerator theorem, with the two tail roles exchanged. It gives \(C_x^{s\mathbf t}=C_{sx}^{\mathbf t}+qC_x^{\mathbf t}\).
These are equalities in \(\mathbb Z[q]\): the numerator theorem first proves equality at every \(t\in(0,1)\), then recovers the polynomial coefficients. Simple multiplication changes length by one, so Steps 2–3 cover every \(x\).
Step 4. Compute two equal letters
In the two-element Coxeter group \(1<s\), the empty-word values and the recurrence give \[ C_1^s(q)=C_s^s(q)=1. \] Prepending \(s\) again gives \[ C_1^{ss}(q)=1+q,\qquad C_s^{ss}(q)=1+q. \] Thus each stalk of the repeated word has two homogeneous generators, in degrees zero and one. This agrees with the explicit basis of its reflected pair lattice.
References
The local source declarations below give the construction and proof.
Lean formalization
OAI.KLInvariance.TitsSpace.bottStalkCharacter_ascent: IntegralPairings.lean:104.
OAI.KLInvariance.TitsSpace.bottStalkCharacter_descent: IntegralPairings.lean:120.