Interpolation produces the prescribed flow-down sections
Let \(H\) be a two-generator Coxeter system and let \(\phi:H\to W\) be an injective homomorphism to a finite-generator Coxeter system, carrying reflections to reflections. On \([1,b]_H\), use the induced ambient polynomial labels.
For every model center \(c\le_Hb\), the structure model \(\mathcal O_c\) has flow-down sections: \[ \forall x\le_Hc,\quad \exists\theta_x\text{ global},\qquad \theta_x(x)=P_c(x),\qquad \theta_x(y)=0\ (y\not\le_Hx). \]
Relation to the OpenAI notes
Same model properties, different justification.
OpenAI notes: The induced-model lemma justifies applying the character theorem to the restricted reflection-faithful realization and uses the dihedral character calculation to identify the rank-one model. loc:models.
Lean (primary): The direct construction first proves parity matching in arbitrary lower Bruhat intervals by a simple-reflection involution. In the dihedral application, orbit-spline factors supply the prescribed values and zeros; their product gives flow-down sections. A finite induction on prescribed upper sets proves flabbiness, while generation and quotient maps follow from the structure-sheaf construction. DihedralIntervals:616, DihedralIntervals:622, DihedralIntervals:626, SheafMorphisms:252, SheafMorphisms:288, DihedralIntervals:176, DihedralIntervals:251, DihedralIntervals:372.
Role in the proof: The Blueprint follows the parity, normal-form prescription, spline product and flow-down correction arguments. Injectivity of the reflection-preserving group map provides distinct factor data and pairwise relatively prime incident labels. These model properties feed the later local-ring decomposition.
Proof
Step 1. Turn an opposite-parity external vertex into an outgoing edge
Take \(y\in Y(x,c)\). It satisfies \(y\le_Hc\) and \(y\not\le_Hx\).
If \(\ell_H(y)<\ell_H(x)\), the dihedral Bruhat comparison theorem would give \(y\le_Hx\). Thus \(\ell_H(y)\ge\ell_H(x)\). Their parities differ, so the inequality is strict.
The odd-gap edge theorem now gives \(x\to_Hy\). Since \(y\le_Hc\le_Hb\), this is an edge of the ambient interval whose target is supported in \(\mathcal O_c\).
Step 2. Construct the inverse correspondence
Conversely, take a supported outgoing edge \(e:x\to_Hy\) with \(y\le_Hc\). Its strict length increase implies \(y\not\le_Hx\). Its odd length gap gives opposite parity. Therefore \(y\in Y(x,c)\).
The maps \(y\mapsto(x,y)\) and \(e\mapsto t(e)\) are inverse: the outgoing source is fixed at \(x\), and interval edges are pairs with their reflection-step property. This is outgoingEquiv.
Step 3. Identify the exact label products
For the edge attached to \(y\), both the interpolation root and induced edge root are the canonical positive ambient root of \(\phi(yx^{-1})\). Their polynomial labels are equal.
Reindexing by the bijection of Step 2 gives \[ \prod_{y\in Y(x,c)}\lambda(a_y) =\prod_{e\in\operatorname{Out}_c(x)}\alpha_e =P_c(x). \] This is interpolation_product.
Step 4. Restrict the interpolating function to the model
Choose \(F:H\to A\) from the dihedral interpolation theorem. At a supported vertex \(v\le_Hc\), set \(\theta_x(v)=F(v)\); at an unsupported vertex use its zero stalk.
For a supported edge \(e:v\to_Hw\), its reflection label satisfies \(w=\tau_ev\). The interpolation congruence for \(\tau_e\) says \[ \alpha_e\mid F(v)-F(w). \] This is exactly structure-sheaf compatibility on that edge. Unsupported edge modules are zero, so \(\theta_x\) is a global section.
Step 5. Check the normalization and support
Since \(x\le_Hc\), the value at \(x\) is \(F(x)\). The interpolation theorem and Step 3 give \[ \theta_x(x)=F(x)=P_c(x). \]
If \(v\not\le_Hx\) and \(v\le_Hc\), the interpolating function vanishes at \(v\). If \(v\not\le_Hc\), its model stalk is zero. Thus \(\theta_x\) vanishes at every vertex outside the lower ideal of \(x\).
These are all clauses of HasFlowDown.
References
The local source declarations below give the construction and proof.
Lean formalization
OAI.KLInvariance.Dihedral.external_step: DihedralIntervals.lean:470.
OAI.KLInvariance.Dihedral.step_external: DihedralIntervals.lean:480.
OAI.KLInvariance.Dihedral.externalEdge: DihedralIntervals.lean:491.
OAI.KLInvariance.Dihedral.externalEdge_mem: DihedralIntervals.lean:496.
OAI.KLInvariance.Dihedral.outgoingEquiv: DihedralIntervals.lean:503.
OAI.KLInvariance.Dihedral.interpolation_product: DihedralIntervals.lean:521.
OAI.KLInvariance.Dihedral.hasFlowDown: DihedralIntervals.lean:531.