Combinatorial Invariance Conjecture

• Combinatorial Invariance Conjecture is the claim that Kazhdan–Lusztig polynomials are determined only by the abstract poset structure of Bruhat intervals, independent of the specific Coxeter group.
• It employs techniques like hypercube decompositions, flipclasses, and diamond generators to establish recursions and symmetries in the combinatorial invariants.
• Recent progress in type A, finite, and affine Coxeter groups demonstrates significant breakthroughs while highlighting challenges for extending the conjecture to arbitrary ranks.

The Combinatorial Invariance Conjecture (CIC) asserts that the Kazhdan--Lusztig polynomials associated with Bruhat intervals in Coxeter groups depend only on the abstract poset structure of those intervals and not on the specific group or generating set. This conjecture, originating in the early 1980s due to Lusztig and Dyer, occupies a central position at the intersection of algebraic combinatorics, geometric representation theory, and the study of Hecke algebras. While a complete proof remains elusive in general Coxeter settings, recent advances in type A and related frameworks have significantly clarified the underlying combinatorics, structural symmetries, and connections to geometric and algebraic invariants.

1. Background and Formal Statement

Given a Coxeter group $(W,S)$ with Bruhat order $\le$ and associated length function $\ell\colon W\to\mathbb{N}$, each interval $[u, v]=\{w\in W \mid u\le w\le v\}$ inherits a poset structure and the Bruhat graph $B(W)$ structure. The Kazhdan--Lusztig polynomials $P_{u,v}(q)$ (and the $R$-polynomials $R_{u,v}(q)$) play vital roles in Lie theory, encoding geometric invariants of Schubert varieties (intersection cohomology Poincaré polynomials) and arising recursively from the Hecke algebra. The Combinatorial Invariance Conjecture states:

$$\text{If } [u,v], [u',v'] \text{ are isomorphic as posets, then } P_{u,v}(q) = P_{u',v'}(q).$$

Equivalently, this asserts that $R_{u,v}(q)$ themselves are determined by the isomorphism type of $[u,v]$ (Blundell et al., 2021, Esposito et al., 19 Sep 2025).

2. Poset-Theoretic Recursions: Hypercube Decompositions and Amazing Elements

A principal combinatorial approach to the conjecture, especially in type $A$ (symmetric groups), employs hypercube decompositions and the concept of amazing hypercube decompositions. For an interval $[u,v]\subset S_{n+1}$, an amazing hypercube decomposition $z$ is defined by:

• Diamond-completeness: $[z,v]$ captures all higher-order commutativity encoded in “diamonds” (2-hypercubes);
• Hypercube-cluster: Certain antichain-generated subgraphs at each $p\in[z,v]$ span Boolean hypercubes;
• Amazingness (additional): For each $x\in[u,v]$, $[z,v]\cap[x,v]$ has a unique minimal element, which is itself a hypercube decomposition of $[x,v]$.

When $z$ satisfies that every such hypercube decomposition is an $R$-element (i.e., admits the expected $R$-polynomial decomposition), the key recursion holds:

$$R_{u,v}(q) = \sum_{p\in W_z[u,v]} q^{d(u,p)} R_{p,v}(q),$$

where $d(u,p)$ denotes distance in the Bruhat graph. This formula is purely poset-theoretic and relies only on the combinatorial data of $[u,v]$ and the choice of $z$ (Esposito et al., 2024). The inheritance property—amazingness passes to all subintervals $[x,v]$—ensures a uniform, canonical way to recurse.

3. Symmetry, Recursion Classes, and Double Shortcuts

A critical structural discovery is the role of double shortcuts in classifying recursion classes of amazing decompositions. Given two such decompositions $z, z'$, the multiset

$DS(z,z') = \{ (d(u,p)+d(p,b), b) \}$

(removing technical details) encodes layered shortcut relations. The conjectured symmetry $DS(z,z') = DS(z',z)$ ensures that any two amazing decompositions lie in the same recursion class, so that determining $R$-element status for one propagates to all. This symmetry, if established for all pairs, implies the full combinatorial invariance in type $A$ (Esposito et al., 2024).

A specialized version using double hypercubes applies to co-elementary intervals, where Barkley and Gaetz demonstrate $DH(z,z') = DH(z',z)$—exploiting reflection orders and Dyer-path arguments—to establish the conjecture for these cases (Esposito et al., 2024).

4. Flipclasses, Path Decompositions, and Partial Results

The flipclass paradigm provides another, finer combinatorial decomposition for intervals in (finite) Weyl groups. Here, all $h$-step paths from $u$ to $v$ are decomposed into flipclasses—minimal subsets stable under “dihedral flips” (exploiting the dihedral substructure of the underlying group). Each flipclass is classified up to combinatorial isomorphism of the underlying time-support graph.

The main technical achievement is determining, for all $h\leq 6$ (in type $A$), that the number of increasing paths in each flipclass, and hence the coefficients $[q^h]R_{u,v}(q)$, are combinatorial invariants, independently of the choice of reflection ordering. Explicit formulas rely only on structural invariants of the time-support graph (Esposito et al., 2024, Esposito et al., 19 Sep 2025). This establishes CIC for intervals of length up to 8 (type $A$ up to length 10).

5. Finite and Affine Coxeter Groups: Elementary and Co-elementary Intervals

Progress in the symmetric group and its generalizations includes:

• Proof of CIC for all elementary intervals in $S_n$—i.e., intervals isomorphic to simple intervals with linearly independent root sets—using strong hypercube decompositions and explicit $H$-polynomial formulas (Barkley et al., 2023).
• In the affine group $\widetilde{A}_2$, the full conjecture is verified by inductively leveraging Z-invariants and m-join sets under poset isomorphisms, handling the combinatorial complexity that emerges beyond the finite ADE types (Burrull et al., 2021).
• For lower intervals and certain parabolic analogues, the conjecture and its parabolic versions are now known to be equivalent, reducible to the behavior on maximal quotients and special matchings (Sentinelli, 2024, Marietti, 2018, Barkley et al., 2024).

6. Poset Invariants, Diamond Generators, and Low-Degree Coefficients

Recent advances demonstrate that certain coefficients of $P_{u,v}(q)$—notably the linear $q$ coefficient—are universal poset invariants for all Coxeter groups. Precisely, the $d$-invariant extracted from $R_{u,v}(q)$ matches a purely poset-theoretic minimum, the size of a diamond-generating set in the Hasse diagram of $[u,v]$. Consequently, full combinatorial invariance now holds for all intervals of length at most $6$ in arbitrary Coxeter groups (Barkley et al., 12 Jan 2026).

7. Outlook, Open Problems, and Generalizations

Despite substantial progress in type $A$, for intervals up to modest length, and in specific families (elementary, co-elementary, short-edge, and lower intervals), the full CIC remains open for arbitrary Coxeter groups. While structural insights—such as the role of hypercube decompositions, flipclasses, and diamond closure—have clarified the combinatorial underpinnings, obstacles remain in extending these approaches to arbitrary ranks and types, especially as the combinatorial complexity of Bruhat intervals increases.

Outstanding open questions include:

• A general combinatorial or geometric description of higher-degree coefficients of $P_{u,v}(q)$;
• Direct combinatorial proofs for symmetry and recursion properties beyond the verified cases;
• Full classification of intervals admitting strong hypercube decompositions, and potential extension of combinatorial invariance to all intervals via such decompositions or flipclass techniques.

Collectively, the current state-of-the-art situates the Combinatorial Invariance Conjecture as a keystone conjecture, guiding a convergence of combinatorial, algebraic, and geometric techniques and prompting the exploration of further structural symmetries in Coxeter and Hecke-theoretic settings (Esposito et al., 2024, Esposito et al., 19 Sep 2025, Barkley et al., 12 Jan 2026, Barkley et al., 2023, Burrull et al., 2021, Esposito et al., 2024, Blundell et al., 2021).