Theorem Signed parameter reciprocity and the inverse-parameter bar coefficient
✓ theorem proved

Signed parameter reciprocity and the inverse-parameter bar coefficient

Let \(q\) be a unit in a commutative ring. The actual R-values satisfy \[ r_q(x,y)=(-q)^{\ell(y)}(-q^{-1})^{\ell(x)}r_{q^{-1}}(x,y). \] Consequently \[ r_{q^{-1}}(x,y)=q^{\ell(x)}b_y(q)(x). \] For comparable endpoints the first equality can also be written \(r_q(x,y)=(-1)^{\ell(y)-\ell(x)}q^{\ell(y)-\ell(x)}r_{q^{-1}}(x,y)\).

Relation to the OpenAI notes

Algebraic construction behind the polynomial conventions.

OpenAI notes: The conventions specify the ordinary R descent recurrence. The reflection-order path theorem introduces the normalized family through path counting and states its square-parameter relation to ordinary R. sec:introduction, bg:paths.

Lean (primary): The regular operators are built from Coxeter length. Left/right commutation and the cyclic identity vector prove reduced-word independence. Inverse operators define R-values; their recurrence proves polynomiality, and a second matching-recurrence argument proves signed parameter reciprocity. Hecke:705, Hecke:1000, OrdinaryR:31, Dihedral:440.

Role in the proof: The KB develops these algebraic constructions before the geometric path formula. The recurrence and inverse-parameter coefficient identities used by the character argument follow from the constructed operators.

Proof

Step 1. Define the comparison family

Set \[ D_q(x,y)=(-q)^{\ell(y)}(-q^{-1})^{\ell(x)}r_{q^{-1}}(x,y). \] On the diagonal, the two scalar powers multiply to one, so \(D_q(x,x)=1\). Outside Bruhat order, the last factor is zero.

We prove that \(D_q\) satisfies the same recurrence as \(r_q\).

Step 2. Compute simultaneous descents

Let \(sy<y\), and suppose \(sx<x\). Put \(m=\ell(sx)\), \(n=\ell(sy)\). The inverse-parameter recurrence is \(r_{q^{-1}}(x,y)=r_{q^{-1}}(sx,sy)\). Therefore \[ D_q(x,y) =(-q)^{n+1}(-q^{-1})^{m+1}r_{q^{-1}}(sx,sy) =D_q(sx,sy), \] because the additional product \((-q)(-q^{-1})\) is one.

Step 3. Compute the ascent branch

Suppose \(sx>x\), put \(v=sy\), and let \(A=(-q)^{\ell(v)}(-q^{-1})^{\ell(x)}\). The inverse-parameter recurrence gives \[ \begin{aligned} D_q(x,y) &=(-q)A\bigl((q^{-1}-1)r_{q^{-1}}(x,v) +q^{-1}r_{q^{-1}}(sx,v)\bigr)\\ &=(q-1)A r_{q^{-1}}(x,v)-A r_{q^{-1}}(sx,v). \end{aligned} \] The first term is \((q-1)D_q(x,v)\). Since \(\ell(sx)=\ell(x)+1\), \[ qD_q(sx,v)=q(-q^{-1})A r_{q^{-1}}(sx,v) =-A r_{q^{-1}}(sx,v). \] This proves the required recurrence in the second case.

Step 4. Identify the two families by induction

Induct strongly on \(\ell(y)\), for all \(x\). At \(y=1\), the diagonal is one and every other bottom label gives zero, for both families.

For \(y\ne1\), choose a left descent. Both recurrences have identical coefficients and use only the shorter upper endpoint \(sy\). Induction identifies their values at \((x,sy)\) and \((sx,sy)\), hence at \((x,y)\). Thus \(r_q=D_q\).

Step 5. Extract the inverse-parameter coefficient

Apply Step 4 with parameter \(q^{-1}\) and expand the definition of \(r_q\): \[ \begin{aligned} r_{q^{-1}}(x,y) &=(-q^{-1})^{\ell(y)}(-q)^{\ell(x)}r_q(x,y)\\ &=(-q^{-1})^{\ell(y)}(-q)^{\ell(x)} (-1)^{\ell(x)}(-q)^{\ell(y)}b_y(q)(x)\\ &=q^{\ell(x)}b_y(q)(x). \end{aligned} \] The top powers cancel because \((-q^{-1})(-q)=1\); the bottom factors satisfy \((-q)(-1)=q\). This is the coefficient identity used in positive-word convolution.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.HeckeQ.dualR: Dihedral.lean:387.

OAI.KLInvariance.HeckeQ.dualR_diagonal: Dihedral.lean:390.

OAI.KLInvariance.HeckeQ.dualR_zero: Dihedral.lean:395.

OAI.KLInvariance.HeckeQ.dualR_recursion: Dihedral.lean:399.

OAI.KLInvariance.HeckeQ.rValue_eq_dualR: Dihedral.lean:440.

OAI.KLInvariance.HeckeQ.rValue_inverse_eq_barBasis: Dihedral.lean:467.

Used by