Theorem The edge factor retains one plane exponent after local reduction
✓ theorem proved

The edge factor retains one plane exponent after local reduction

Fix a finite-generator real Coxeter system \(W\) and polynomial coefficient algebra \(A\) from a basis of its extended realization. Let \(\lambda(c)\) denote the polynomial of an embedded root \(c\). Let \(x\in W\), and let \(a,b\) be distinct positive roots. Put \[ P=\operatorname{span}\{a,b\},\qquad R_P=A_{\mathfrak p_P},\qquad E_{P,a}=R_P/(\eta_P(\lambda(a))). \] Here \(\eta_P:A\to R_P\) is localization and \(q_{P,a}:R_P\to E_{P,a}\) is the quotient map. Write \[ r_{P,a}(c)=q_{P,a}(\eta_P(\lambda(c))) \] for localized root reduction, and put \(p=r_{P,a}(b)\).

The image of the polynomial lift satisfies \[ q_{P,a}(\eta_P(F_{x,a}))\sim p^{\,k_P(x,a)} \quad\text{in }E_{P,a}, \] where \(\sim\) means equality up to multiplication by a unit. The weight is interpreted as its equal natural-number value in the exponent.

Relation to the OpenAI notes

Associated products give the local scalar-image calculation.

OpenAI notes: In a localized edge quotient viewed as a discrete valuation ring, the notes compare valuations of later-label products. The current plane contributes its exponent to the correction factor, and all other plane factors are units. edge:bound.

Lean (primary): The factor proof directly establishes association with the power of a chosen reduced plane root, checking the same-plane and other-plane factors and the absent-plane case. LocalProducts then works over arbitrary compatible commutative scalar rings: a cardinality equation and associated factors give equality of scalar-image submodules. PlaneLocalization:1560, PlaneLocalization:1569, PlaneLocalization:1579, PlaneLocalization:1509, PlaneLocalization:1523.

Role in the proof: The Blueprint explains the unit and power calculations explicitly. The later-edge count remains the stated hypothesis of the product-image lemma until the dihedral transport argument supplies it.

Proof

Step 1. Show that a different plane has its representative outside P

Take \(Q\in\mathcal P(x,a)\) with \(Q\ne P\). Its representative root satisfies \[ b_Q\ne a,\qquad Q=\operatorname{span}\{a,b_Q\}. \] If \(b_Q\) belonged to \(P\), this span would be contained in \(P\).

The roots \(a,b_Q\) are independent, so their span has dimension two. The plane \(P\) also has dimension two because \(a\ne b\). Inclusion and equal dimensions would give \(Q=P\), a contradiction. Thus \(b_Q\notin P\).

Step 2. Prove the other-plane factors are units

The plane-prime unit criterion makes \(\eta_P(\lambda(b_Q))\) a unit in \(R_P\) whenever \(Q\ne P\).

A ring homomorphism sends a unit and its inverse to mutually inverse elements. Hence its image \(r_{P,a}(b_Q)\) is a unit in \(E_{P,a}\), and so is any nonnegative power of that image. Therefore \[ r_{P,a}(b_Q)^{k_Q}\sim1\qquad(Q\ne P). \]

Step 3. Identify the factor from P itself

If \(P\) belongs to the supported set, its representative \(b_P\) lies in \(P\) and differs from \(a\). Write \[ b_P=sa+tb. \] The coefficient \(t\) is nonzero, since otherwise \(b_P\) would lie on the line of \(a\), contradicting independence of distinct positive roots.

Reduction kills \(a\), so \[ r_{P,a}(b_P)=t\,r_{P,a}(b)=tp. \] The nonzero real scalar \(t\) has inverse \(t^{-1}\), whose image is its inverse in the quotient ring. Thus the localized root-reduction theorem gives \[ r_{P,a}(b_P)\sim p,\qquad r_{P,a}(b_P)^{k_P}\sim p^{k_P}. \]

Step 4. Multiply the factor associations

The localization and quotient maps preserve the defining product: \[ q_{P,a}(\eta_P(F_{x,a})) =\prod_{Q\in\mathcal P(x,a)}r_{P,a}(b_Q)^{k_Q}. \] Every factor from \(Q\ne P\) is associated to \(1\) by Step 2. If \(P\) is present, its unique factor is associated to \(p^{k_P}\) by Step 3. Multiplying the witnessing units gives the required association of the full product.

Step 5. Treat an absent plane

If \(P\notin\mathcal P(x,a)\), then \(k_P=0\). Every factor is from another plane, so Step 2 makes the full product a unit. Its target power is \(p^0=1\).

This proves the association in both support cases. It records explicitly the contribution of every factor, including the unit factors that remain after localization.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.TitsSpace.planeRoot_outside: PlaneLocalization.lean:1560.

OAI.KLInvariance.TitsSpace.reduced_planeRoot_other_isUnit: PlaneLocalization.lean:1569.

OAI.KLInvariance.TitsSpace.edgeFactor_local_associated: PlaneLocalization.lean:1579.

Used by