[Paper Review] A q-analogue of the four functions theorem

Demetres Christofides|ArXiv.org|Sep 28, 2009

Advanced Mathematical Identities3 references3 citations

TL;DR

This paper establishes a q-analogue of the four functions theorem, proving that for a finite distributive lattice L and non-negative functions α, β, γ, δ satisfying α(x)β(y) ≤ γ(x∨y)δ(x∧y) for all x,y ∈ L, the inequality ∑x∈X α(x)q^{r(x)} ∑x∈Y β(x)q^{r(x)} ≪ ∑x∈X∨Y γ(x)q^{r(x)} ∑x∈X∧Y δ(x)q^{r(x)} holds for all subsets X,Y ⊆ L. The result unifies the classical four functions theorem and Björner's q-analogue of the FKG inequality as special cases.

ABSTRACT

In this article we give a proof of a q-analogue of the celebrated four functions theorem. This analogue was conjectured by Bjorner and includes as special cases both the four functions theorem and also Bjorner's q-analogue of the FKG inequality.

Motivation & Objective

• To resolve a conjecture by Björner regarding a q-analogue of the four functions theorem.
• To extend the four functions theorem and the FKG inequality into a unified q-analytic framework using rank-weighted generating functions.
• To establish a polynomial inequality in q with non-negative coefficients, generalizing known correlation inequalities.
• To provide a new proof of Björner's q-FKG inequality as a consequence of the main result.

Proposed method

• Prove the main result by reducing it to the Boolean lattice P(n) via Birkhoff’s representation theorem.
• Use induction on the size of the symmetric difference of sets to analyze coefficients of q^k in the polynomial inequality.
• Fix F = A ∩ B and G = A ∪ B, then reduce the inequality to a sublattice on G \ F using shifted functions α', β', γ', δ'.
• Apply Theorem 2.2, which establishes a key inequality for complementary sets A and B^c in P(n), by defining f(A) = α(A)β(A^c) and g(A) = γ(A^c)δ(A).
• Use Lemma 2.3 to show that f(P(n))² ≤ g(P(n))², implying the desired coefficient-wise inequality.
• Disprove a stronger conjecture by constructing a counterexample for n=2 where the pointwise inequality fails despite satisfying the original functional condition.

Experimental results

Research questions

• RQ1 Does a q-analogue of the four functions theorem exist that generalizes both the classical four functions theorem and Björner’s q-FKG inequality?
• RQ2 Can the inequality ∑x∈X α(x)q^{r(x)} ∑x∈Y β(x)q^{r(x)} ≪ ∑x∈X∨Y γ(x)q^{r(x)} ∑x∈X∧Y δ(x)q^{r(x)} be proven under the condition α(x)β(y) ≤ γ(x∨y)δ(x∧y) for all x,y in a finite distributive lattice?
• RQ3 Is there a stronger pointwise inequality α(A)β(B) + α(B)β(A) ≤ γ(A)δ(B) + γ(B)δ(A) that would imply the main result, and if so, is it valid?
• RQ4 Can the main result be used to derive new inequalities in extremal combinatorics or probabilistic lattice theory?
• RQ5 What is the role of the rank function r(x) in enabling the q-analogue, and how does it interact with log-supermodular measures?

Key findings

• The q-analogue of the four functions theorem holds: for any finite distributive lattice L and non-negative functions α, β, γ, δ satisfying α(x)β(y) ≤ γ(x∨y)δ(x∧y), the inequality ∑x∈X α(x)q^{r(x)} ∑x∈Y β(x)q^{r(x)} ≪ ∑x∈X∨Y γ(x)q^{r(x)} ∑x∈X∧Y δ(x)q^{r(x)} is valid for all X,Y ⊆ L.
• The result generalizes both the classical four functions theorem (obtained by setting q=1) and Björner’s q-FKG inequality.
• The proof relies on reducing the problem to the Boolean lattice P(n) and analyzing coefficients of q^k via set decompositions based on symmetric differences.
• A stronger conjecture — that α(A)β(B) + α(B)β(A) ≤ γ(A)δ(B) + γ(B)δ(A) — is shown to be false via a counterexample for n=2.
• The counterexample uses specific values on P(2): α(∅)=0, α({1})=0, α({2})=1, α({1,2})=0; β similarly defined, with γ and δ set to 1 on {2} and ∅, respectively, showing the failure of the pointwise inequality.
• The method provides a new proof of Björner’s q-FKG inequality by deriving it as a special case of the main theorem.