A parameter family from the Hall–Littlewood limit

What Lean checks

Lean proves nonmembership in the full real Muirhead semiring for the explicit unique monic triangular eigenfunctions, every rectangular translate and the small-positive-q family for every 3/4≤t<1. All 16 existing modules passed fresh strict Lean 4.35 compilation with warnings treated as errors. The complete owned/private audit checked 1,454 declarations, including 1,254 theorem declarations, 816 private declarations and all 49 required endpoints. Only propext, Classical.choice and Quot.sound are allowed. Independent stock-kernel replay of the revised closure and the combined repository gate remain pending. The README badge remains partial. The current endpoints use the standard monic triangular Macdonald eigenfunctions. Stable Hall orthogonality followed by specialization is being formalized separately. The manuscript's literal truncated finite-variable Hall pairing conflicts with that convention; the degree-three diagnostic must be reconciled before claiming a full literal-source formalization.

This note strengthens the exact interior counterexample in PROOF.md. It uses the same separating functional and resolves no additional conjecture.

Let LL be the degree-six functional defined there by

(L(M600),L(M510),L(M420),L(M411),L(M330),L(M321),L(M222))=(62,28,9,9,2,2,0).\bigl(L(M_{600}),L(M_{510}),L(M_{420}),L(M_{411}), L(M_{330}),L(M_{321}),L(M_{222})\bigr) =(62,28,9,9,2,2,0).

The cone certificate in the main proof is independent of the parameters q,tq,t.

Exact Hall–Littlewood expansions

The standard Hall–Littlewood specialization is Pλ(x;0,t)=Pλ(x;t)P_\lambda(x;0,t)=P_\lambda(x;t). Its symmetrization formula is

Pλ(x1,x2,x3;t)=1vλ(t)∑w∈S3w(x1λ1x2λ2x3λ3∏i<jxi−txjxi−xj),P_\lambda(x_1,x_2,x_3;t) =\frac{1}{v_\lambda(t)}\sum_{w\in S_3} w\left(x_1^{\lambda_1}x_2^{\lambda_2}x_3^{\lambda_3} \prod_{i\lt j}\frac{x_i-tx_j}{x_i-x_j}\right),

where zero parts are included in the multiplicities mr(λ)m_r(\lambda), and

vλ(t)=∏r≥0∏j=1mr(λ)1−tj1−t.v_\lambda(t)=\prod_{r\ge0}\prod_{j=1}^{m_r(\lambda)} \frac{1-t^j}{1-t}.

These are Macdonald, Symmetric Functions and Hall Polynomials, second edition, Chapter III, (2.1), printed page 208, and Chapter VI, (4.14)(ii), printed page 324. The primary book PDF was inspected at these formulas. In particular, v420=1v_{420}=1 and v411=1+tv_{411}=1+t.

Put Δ=(x1−x2)(x1−x3)(x2−x3)\Delta=(x_1-x_2)(x_1-x_3)(x_2-x_3). Multiplying the symmetrization formula by vλ(t)Δv_\lambda(t)\Delta reduces each expansion below to a polynomial identity:

vλ(t)ΔPλ(x;t)=∑w∈S3sgn(w)w(xλ∏i<j(xi−txj)).v_\lambda(t)\Delta P_\lambda(x;t) =\sum_{w\in S_3}\mathop{\mathrm{sgn}}(w) w\left(x^\lambda\prod_{i\lt j}(x_i-tx_j)\right).

Expanding its six terms gives

P420(x;0,t)=m420+(1−t)(m411+m330)+(1−t)(2−t)m321+(1−t)(t2−t+3)m222,P411(x;0,t)=m411+(1−t)m321+(1−t)2m222.\begin{aligned} P_{420}(x;0,t) ={}&m_{420}+(1-t)(m_{411}+m_{330}) +(1-t)(2-t)m_{321}\\ &+(1-t)(t^2-t+3)m_{222},\\ P_{411}(x;0,t) ={}&m_{411}+(1-t)m_{321}+(1-t)^2m_{222}. \end{aligned}

The multiplicities of the relevant monomial orbits are

m420(𝟏)=6,m411(𝟏)=m330(𝟏)=3,m321(𝟏)=6,m222(𝟏)=1.m_{420}(\mathbf1)=6,\qquad m_{411}(\mathbf1)=m_{330}(\mathbf1)=3,\qquad m_{321}(\mathbf1)=6,\qquad m_{222}(\mathbf1)=1.

Consequently their values at 𝟏\mathbf1 are

A(t)=27−28t+8t2−t3,B(t)=10−8t+t2.A(t)=27-28t+8t^2-t^3,\qquad B(t)=10-8t+t^2.

They are positive for 0<t<10\lt t<1: both are decreasing on [0,1][0,1], with A(1)=6A(1)=6 and B(1)=3B(1)=3.

Strict separation on an interval

The displayed expansions and the definition of LL give

L(P420(x;0,t)P420(𝟏;0,t))=3(4t2−23t+37)A(t),L\left(\frac{P_{420}(x;0,t)}{P_{420}(\mathbf1;0,t)}\right) =\frac{3(4t^2-23t+37)}{A(t)},

and

L(P411(x;0,t)P411(𝟏;0,t))=3(13−4t)B(t).L\left(\frac{P_{411}(x;0,t)}{P_{411}(\mathbf1;0,t)}\right) =\frac{3(13-4t)}{B(t)}.

Their difference factors exactly as

L(P420(x;0,t)P420(𝟏;0,t)−P411(x;0,t)P411(𝟏;0,t))=3(1−t)(10t2−35t+19)A(t)B(t).L\left(\frac{P_{420}(x;0,t)}{P_{420}(\mathbf1;0,t)} -\frac{P_{411}(x;0,t)}{P_{411}(\mathbf1;0,t)}\right) =\frac{3(1-t)(10t^2-35t+19)}{A(t)B(t)}.

For 3/4≤t<13/4\le t<1, the quadratic in the numerator is negative. Its derivative is 20t−35<020t-35<0 on this interval, and its value at t=3/4t=3/4 is −13/8-13/8. Therefore the functional is strictly negative throughout this interval.

The point q=0q=0 is a boundary point of the conjecture's parameter domain. We use it only to obtain admissible interior counterexamples by continuity. For each fixed 0<t<10\lt t<1, the Macdonald coefficients extend continuously to q=0q=0, with the Hall–Littlewood values above. One can see the regularity directly from the defining Hall inner product: in each fixed degree its power-sum Gram matrix has positive diagonal entries zρ∏i(1−qρi)/(1−tρi)z_\rho\prod_i(1-q^{\rho_i})/(1-t^{\rho_i}) at q=0q=0. Gram–Schmidt therefore has positive norm denominators there and produces coefficients continuous in qq. The normalizing values at 𝟏\mathbf1 are positive there. Hence, for every t∈[3/4,1)t\in[3/4,1), there is ε(t)>0\varepsilon(t)>0 such that

L(P420(x;q,t)P420(𝟏;q,t)−P411(x;q,t)P411(𝟏;q,t))<0(0<q<ε(t)).L\left(\frac{P_{420}(x;q,t)}{P_{420}(\mathbf1;q,t)} -\frac{P_{411}(x;q,t)}{P_{411}(\mathbf1;q,t)}\right)<0 \qquad(0\lt q\lt \varepsilon(t)).

Thus the failure occurs in a continuous parameter family inside the stated domain. The exact rational point q=1/10,t=9/10q=1/10,t=9/10 in the main proof supplies an explicit interior example without relying on an unspecified continuity radius.

Independent calculation

The expansions were obtained independently from the six-term symmetrization formula, without the branching implementation used to discover the candidate. Exact polynomial division by vλΔv_\lambda\Delta, evaluation at 𝟏\mathbf1, and factorization of the functional difference reproduce all formulas above. The cone values were also recomputed independently in independent_cone_check.py; all 38 reduced products and all seven pure monomials have nonnegative functional value.