No general positive combinatorial rule for Kronecker coefficients is known

The results of this page are exact but do not amount to a general positive combinatorial rule. The definition (Kronecker coefficients and the internal tensor product) computes $g(\lambda ,\mu ,\nu )$ as the multiplicity of a tensor-product decomposition; the three-character formula (The Kronecker coefficient is a three-character inner product) expresses it as a closed average over $S_n$; the internal product (The internal (Kronecker) product on symmetric functions) and the reduction and row-gap results (The Kronecker reduction lemma of Pak and Panova, Row-gap reduction and conditional k-stability of Kronecker coefficients) reduce and stabilize the coefficient. None of these is a combinatorial interpretation of the numbers $g(\lambda ,\mu ,\nu )$, and no general such rule is known: as recorded by the sources cited below, obtaining a combinatorial interpretation of, or a meaningful positive formula for, all Kronecker coefficients is a well-known open problem in algebraic combinatorics, and the general computation with binary-encoded partition parts is #P-hard. Panova (2025), Abstract and Introduction (pp. 1–2), also records the general combinatorial-interpretation problem as open.

Claim boundary of this page. This page claims no positive combinatorial rule for the Kronecker coefficients, no polynomial-time algorithm for computing them, and no general criterion deciding their positivity. It also does not claim that such a rule is impossible; the sources record a hard open problem, not a proved nonexistence theorem.

Status of the statement. This is a record of the judgment of the cited literature, not a theorem proved on this page; it has no proof section and no items below depend on it as a mathematical supplier.

(a) What the page does supply, and what it does not. The character formula computes $g$ as an exact finite average, but over all $n!$ elements of $S_n$, which is not a method of combinatorial enumeration; the internal product turns the problem into multiplication in the ring of symmetric functions; the reduction lemma either returns $0$ or transports the coefficient to partitions of strictly bounded size; and $k$-stability of row padding says that the padded sequence is eventually constant, without exhibiting the limit as a count of objects. These are structural and computational statements about $g$, and they are consistent with the continued absence of a general positive rule.

(b) The one classical positive case. By Murnaghan and Littlewood, the Littlewood-Richardson coefficients are the Kronecker coefficients of a special stable family, and they do have the Littlewood-Richardson rule; this is the known case in which a positive combinatorial interpretation exists, and it does not extend to the general coefficient. Beyond it, positive rules are known only for restricted families of triples, and the sources record that outside such special cases there is no combinatorial interpretation.

(c) Complexity context. The source notes that computing $g(\lambda ,\mu ,\nu )$ is #P-hard in general. Counting hardness alone does not imply hardness of deciding positivity. The source also explains that a positive combinatorial interpretation analogous to the Littlewood-Richardson rule "would likely show that the decision problem is in NP and the counting problem in #P". The character formula of this page does not contradict this: its displayed sum has $n!$ terms, so it gives no polynomial-time bound in the partition input size.

(d) Individual computations are not a rule. The worked examples of the companion examples page evaluate individual coefficients, and the counterexamples there detect a stability boundary; an individual evaluation or a stability statement is not a general rule, and none of the examples on this pair is presented as one.

No form of the Axiom of Choice is used: this remark records the status of a problem and cites no choice-dependent construction.