For qq not a root of unity, it is a basic open problem to decide when the tensor product L(M)⊗L(N)L(M)\otimes L(N) of two simple finite-dimensional modules over the quantum affine algebra Uq(sl^k)U_q(\widehat{\mathfrak{sl}}_k) is again simple. Bittmann and Li (2022) combined Lapid–Mínguez's criterion for parabolic induction of pp-adic GLNGL_N with quantum affine Schur–Weyl duality and an explicit Grothendieck-ring isomorphism to prove such a criterion for every rank k≥1k\geq 1 in two cases: both factors snake modules, or one factor a fundamental module at an extremal node and the other arbitrary. In both cases irreducibility is equivalent to two explicit matching conditions LCk(mM,mN)\mathrm{LC}_k(m_M,m_N) and LCk(mN,mM)\mathrm{LC}_k(m_N,m_M) on the associated multisegments. They conjecture the same criterion for the asymmetric case of an arbitrary snake module tensored with an arbitrary simple module (Conjecture 5.10), with the identification of the matching conditions with socle/cosocle conditions for arbitrary ladders (Conjecture 5.6) as the expected key step. The conjectured criterion, and with it any effective irreducibility criterion for general tensor products in this category, remains open.

Background

Let q∈C×q\in\mathbb{C}^{\times} be not a root of unity and let Uq(sl^k)U_q(\widehat{\mathfrak{sl}}_k) be the quantum affine algebra of type Ak−1A_{k-1}, the quantized enveloping algebra of the loop algebra of slk\mathfrak{sl}_k. The category Ck\mathscr{C}_k of finite-dimensional type-1 representations — the standard class of finite-dimensional modules, containing all simple modules, on which the loop parameter enters only through integral powers — is a non-semisimple braided tensor category whose simple objects L(M)L(M) are classified, by work of Chari and Pressley, by dominant monomials MM in commuting variables Yi,rY_{i,r} (1≤i≤k−11\leq i\leq k-1, r∈Zr\in\mathbb{Z}). Its Grothendieck ring is encoded by the qq-character map χq\chi_q of Frenkel–Reshetikhin, an injective ring homomorphism from K0(Ck)K_0(\mathscr{C}_k) to a Laurent polynomial ring in the Yi,r±1Y_{i,r}^{\pm 1}; the algorithm of Frenkel–Mukhin computes qq-characters of many simple modules iteratively. A structural question open since the 1990s is: for which pairs of simple modules is L(M)⊗L(N)L(M)\otimes L(N) simple? This also matters for the cluster-algebra structure of the Grothendieck ring discovered by Hernandez–Leclerc (a cluster algebra being a commutative ring equipped with a distinguished set of generators, its cluster variables, produced by an iterative mutation process): a simple module is called real when its tensor square is simple, and compatibility of cluster variables is expected to correspond to simplicity of the corresponding tensor product.

A parallel combinatorial theory exists for representations of GLN(F)GL_N(F) over a non-archimedean field FF. Irreducibles are classified by Zelevinsky multisegments, finite formal sums of segments Δ=[a,b]\Delta=[a,b] with a≤ba\leq b; Z(m)Z(m) denotes the irreducible attached to mm by the Zelevinsky classification, and the parabolic induction of two irreducibles — the representation of GLN+M(F)GL_{N+M}(F) obtained by inducing Z(m)⊗Z(n)Z(m)\otimes Z(n) up from the Levi factor GLN(F)×GLM(F)GL_N(F)\times GL_M(F), where Z(m)Z(m) and Z(n)Z(n) are representations of GLN(F)GL_N(F) and GLM(F)GL_M(F) — is written Z(m)×Z(n)Z(m)\times Z(n). A multisegment is a ladder if both endpoints of its segments increase along the sum; Δ≺Δ′\Delta\prec\Delta' means b(Δ)+1≤b(Δ′)≤e(Δ)+1≤e(Δ′)b(\Delta)+1\leq b(\Delta')\leq e(\Delta)+1\leq e(\Delta'). Lapid–Mínguez (2016) gave a purely combinatorial criterion for irreducibility of Z(m)×Z(n)Z(m)\times Z(n) with mm a ladder and nn arbitrary, in terms of matching functions between the sets Xm,n={(i,j):Δi≺Δj′}X_{m,n}=\{(i,j):\Delta_i\prec\Delta'_j\} and Ym,n={(i,j):Δi←≺Δj′}Y_{m,n}=\{(i,j):\overleftarrow{\Delta_i}\prec\Delta'_j\} of linked segment pairs, where Δ←=[a−1,b−1]\overleftarrow{\Delta}=[a-1,b-1] denotes the segment Δ=[a,b]\Delta=[a,b] shifted down by one (the condition LC(m,n)\mathrm{LC}(m,n)); their 2018 paper connected these conditions to the stronger notion of □\square-irreducibility (a representation is □\square-irreducible when its parabolic induction with any irreducible representation is again irreducible).

The two worlds are bridged by quantum affine Schur–Weyl duality (Jimbo; Chari–Pressley): a monoidal functor from modules over the affine Hecke algebra H^N(q2)\widehat{\mathcal{H}}_N(q^2), which by type theory for pp-adic groups (Bushnell–Kutzko, Heiermann) realizes Bernstein blocks of GLN(F)GL_N(F), to the level-NN subcategory of Ck\mathscr{C}_k. The functor is an equivalence of categories only when N<kN<k, and it sends a segment [a,b][a,b] to L(Yb−a+1,−a−b)L(Y_{b-a+1,-a-b}); ladders correspond exactly to snake modules, the simple modules introduced by Mukhin–Young whose highest weights form a staircase chain of the Yi,rY_{i,r}. Hence Lapid–Mínguez's criterion transfers to tensor products of a snake module with an arbitrary simple module only in the large-kk regime.

Bittmann–Li (2022) treat all k≥1k\geq 1. They construct a category Ck\mathcal{C}_k by quotienting and localizing the pp-adic-side category (segments of length >k>k are sent to zero, segments of length kk to the trivial representation), through which the Schur–Weyl functor factors as an exact monoidal functor F~:Ck→Ck\widetilde{\mathcal{F}}:\mathcal{C}_k\to\mathscr{C}_k, and they prove that F~\widetilde{\mathcal{F}} induces an isomorphism of Grothendieck rings. Writing Δ≺kΔ′\Delta\prec_k\Delta' for Δ≺Δ′\Delta\prec\Delta' with e(Δ′)−b(Δ)<ke(\Delta')-b(\Delta)<k, they define modified matching conditions LCk(m,n)\mathrm{LC}_k(m,n) using the sets Xm,n(k)X^{(k)}_{m,n}, Ym,n(k)Y^{(k)}_{m,n} built from ≺k\prec_k instead of ≺\prec, and analogously RCk\mathrm{RC}_k. Their main theorem states that for two snake modules, L(M)⊗L(M′)L(M)\otimes L(M') is simple if and only if both LCk(mM,mM′)\mathrm{LC}_k(m_M,m_{M'}) and LCk(mM′,mM)\mathrm{LC}_k(m_{M'},m_M) hold; the same is proved when one factor is a fundamental module L(Y1,r)L(Y_{1,r}) or L(Yk−1,r)L(Y_{k-1,r}) at an extremal node of the Dynkin diagram and the other is an arbitrary simple module. Via the Hernandez–Leclerc correspondence and the parametrization of the dual canonical basis of the quotient cluster algebra C[Gr(k,n,∼)]\mathbb{C}[\mathrm{Gr}(k,n,\sim)] by rectangular semistandard Young tableaux (Chang–Duan–Fraser–Li), this also yields a tableau-compatibility criterion generalizing the classical result of Leclerc–Zelevinsky that two Plücker coordinates — the standard generators of the Grassmannian coordinate ring, indexed by kk-element column sets — are compatible if and only if their column sets are weakly separated.

What the paper leaves open is the asymmetric case: one factor an arbitrary snake module, the other an arbitrary simple module (Conjecture 5.10 in the arXiv v3 numbering; Conjecture 6.13 in the v2 numbering). The expected key ingredient, that LCk\mathrm{LC}_k and RCk\mathrm{RC}_k identify with the socle and cosocle conditions LIk\mathrm{LI}_k and RIk\mathrm{RI}_k (socle and cosocle meaning maximal semisimple submodule and maximal semisimple quotient; LIk(π,σ)\mathrm{LI}_k(\pi,\sigma) says Z(m+n)=soc(π×σ)Z(m+n)=\mathrm{soc}(\pi\times\sigma) in Ck\mathcal{C}_k) for arbitrary ladders and not just singleton segments [a,a][a,a], is likewise conjectural (Conjecture 5.6, v2 Conjecture 6.8). The paper additionally records two further open statements: that F~\widetilde{\mathcal{F}} is an equivalence of categories (Conjecture 3.7, v2 Conjecture 3.6), and that the tableau criterion should lift from the quotient ring C[Gr(k,n,∼)]\mathbb{C}[\mathrm{Gr}(k,n,\sim)] to the full homogeneous coordinate ring C[Gr(k,n)]\mathbb{C}[\mathrm{Gr}(k,n)] (Remark 7.9, v2 Remark 6.14); these are adjacent problems with their own objectives. The fully general problem — an effective criterion for L(M)⊗L(N)L(M)\otimes L(N) with both factors arbitrary — remains open in the field at large.

Problem Statement

Prove or refute the conjectured irreducibility criterion of Bittmann–Li for tensor products with one snake factor, valid for every rank k≥1k\geq 1:

(1) Conjecture 5.10 (arXiv v3 numbering; Conjecture 6.13 in v2). For every k≥1k\geq 1, every simple snake module L(M)L(M) of Uq(sl^k)U_q(\widehat{\mathfrak{sl}}_k) — that is, every simple module whose multisegment mMm_M, read off the dominant monomial via [a,b]↦Yb−a+1,−a−b[a,b]\mapsto Y_{b-a+1,-a-b}, is a ladder — and every arbitrary simple module L(N)L(N): the tensor product L(M)⊗L(N)L(M)\otimes L(N) is irreducible if and only if the two matching conditions LCk(mM,mN)\mathrm{LC}_k(m_M,m_N) and LCk(mN,mM)\mathrm{LC}_k(m_N,m_M) both hold. Here the sets of kk-linked segment pairs are Xm,n(k)={(i,j):Δi≺kΔj′}X^{(k)}_{m,n}=\{(i,j):\Delta_i\prec_k\Delta'_j\} and Ym,n(k)={(i,j):Δi←≺kΔj′}Y^{(k)}_{m,n}=\{(i,j):\overleftarrow{\Delta_i}\prec_k\Delta'_j\}, where, for segments Δ=[b(Δ),e(Δ)]\Delta=[b(\Delta),e(\Delta)], Δ≺Δ′\Delta\prec\Delta' means b(Δ)+1≤b(Δ′)≤e(Δ)+1≤e(Δ′)b(\Delta)+1\leq b(\Delta')\leq e(\Delta)+1\leq e(\Delta'), Δ≺kΔ′\Delta\prec_k\Delta' means Δ≺Δ′\Delta\prec\Delta' and e(Δ′)−b(Δ)<ke(\Delta')-b(\Delta)<k, and Δ←=[b(Δ)−1,e(Δ)−1]\overleftarrow{\Delta}=[b(\Delta)-1,e(\Delta)-1] is the shifted segment; each condition LCk(m,n)\mathrm{LC}_k(m,n) requires an injective map f ⁣:Xm,n(k)→Ym,n(k)f\colon X^{(k)}_{m,n}\to Y^{(k)}_{m,n}, f(i,j)=(i′,j′)f(i,j)=(i',j'), such that for every (i,j)∈Xm,n(k)(i,j)\in X^{(k)}_{m,n} either i=i′i=i' and Δj′′≺kΔj′\Delta'_{j'}\prec_k\Delta'_j, or j=j′j=j' and Δi≺kΔi′\Delta_i\prec_k\Delta_{i'}.

(2) The conjectured key ingredient, Conjecture 5.6: for any ladder multisegment m∈Multkm\in\mathrm{Mult}_k and any multisegment n∈Multkn\in\mathrm{Mult}_k, the condition LCk(m,n)\mathrm{LC}_k(m,n) is equivalent to the socle condition LIk(Z(m),Z(n))\mathrm{LI}_k(Z(m),Z(n)), and RCk(m,n)\mathrm{RC}_k(m,n) to the cosocle condition RIk(Z(m),Z(n))\mathrm{RI}_k(Z(m),Z(n)) — currently proved only for singleton segments m=[a,a]m=[a,a] (Proposition 5.5).

A complete answer establishes (1) for all k≥1k\geq 1 and all snake/arbitrary pairs, either directly or by first establishing (2) and deducing (1); the definitions of Ck\mathcal{C}_k, ≺k\prec_k, the sets X(k),Y(k)X^{(k)},Y^{(k)}, matching functions, and ladders are those of the source paper. A single explicit counterexample — a specific k≥1k\geq 1, an explicit snake dominant monomial MM and dominant monomial NN, with the composition factors of L(M)⊗L(N)L(M)\otimes L(N) computed, exhibiting a mismatch between irreducibility and the two LCk\mathrm{LC}_k conditions — refutes the conjecture and equally resolves the question. The surrounding general problem (an effective criterion for two arbitrary simple factors) and the paper's separate open statements — the equivalence of categories F~\widetilde{\mathcal{F}} (Conjecture 3.7) and the lift of the tableau criterion to C[Gr(k,n)]\mathbb{C}[\mathrm{Gr}(k,n)] (Remark 7.9) — are context and are not required for completion.

References

  1. Léa Bittmann and Jian-Rong Li, On the simplicity of the tensor product of two simple modules of quantum affine algebras, preprint (2022, arXiv:2203.17268v3, revised 21 July 2025), DOI: 10.48550/arXiv.2203.17268, https://arxiv.org/abs/2203.17268
  2. Erez Lapid and Alberto Mínguez, On parabolic induction on inner forms of the general linear group over a non-archimedean local field, Selecta Mathematica 22 (2016) 2347–2400, DOI: 10.1007/s00029-016-0281-7, https://doi.org/10.1007/s00029-016-0281-7
  3. Erez Lapid and Alberto Mínguez, Geometric conditions for □-irreducibility of certain representations of the general linear group over a non-archimedean local field, Advances in Mathematics 339 (2018) 113–190, DOI: 10.1016/j.aim.2018.09.027, https://doi.org/10.1016/j.aim.2018.09.027
  4. Maxim Gurevich, Quantum invariants for decomposition problems in type A rings of representations, Journal of Combinatorial Theory Series A 180 (2021) 105431, DOI: 10.1016/j.jcta.2021.105431, https://doi.org/10.1016/j.jcta.2021.105431
  5. E. Mukhin and C. A. S. Young, Path description of type B q-characters, Advances in Mathematics 231 (2012) 1119–1150, DOI: 10.1016/j.aim.2012.06.012, https://doi.org/10.1016/j.aim.2012.06.012
  6. Wen Chang, Bing Duan, Chris Fraser, and Jian-Rong Li, Quantum affine algebras and Grassmannians, Mathematische Zeitschrift 296 (2020) 1539–1583, DOI: 10.1007/s00209-020-02496-7, https://doi.org/10.1007/s00209-020-02496-7
  7. Ryo Fujita, Affine highest weight categories and quantum affine Schur-Weyl duality of Dynkin quiver types, Representation Theory 26 (2022) 211–263, DOI: 10.1090/ert/601, https://doi.org/10.1090/ert/601
  8. Seok-Jin Kang, Masaki Kashiwara, and Myungho Kim, Symmetric quiver Hecke algebras and R-matrices of quantum affine algebras, Inventiones Mathematicae 211 (2018) 591–685, DOI: 10.1007/s00222-017-0754-0, https://doi.org/10.1007/s00222-017-0754-0
  9. Edward Frenkel and Evgeny Mukhin, Combinatorics of q-characters of finite-dimensional representations of quantum affine algebras, Communications in Mathematical Physics 216 (2001) 23–57, DOI: 10.1007/s002200000323, https://doi.org/10.1007/s002200000323
  10. Bernard Leclerc and Andrei Zelevinsky, Quasicommuting families of quantum Plücker coordinates, in Kirillov's seminar on representation theory, AMS Translations Series 2 181 (1998) 85–108, DOI: 10.1090/trans2/181/03, https://doi.org/10.1090/trans2/181/03
  11. Vyjayanthi Chari and Andrew Pressley, Quantum affine algebras and affine Hecke algebras, Pacific Journal of Mathematics 174 (1996) 295–326, DOI: 10.2140/pjm.1996.174.295, https://doi.org/10.2140/pjm.1996.174.295
  12. David Hernandez and Bernard Leclerc, Cluster algebras and quantum affine algebras, Duke Mathematical Journal 154 (2010) 265–341, DOI: 10.1215/00127094-2010-040, https://doi.org/10.1215/00127094-2010-040
  13. Nick Early and Jian-Rong Li, Tropical geometry, quantum affine algebras, and scattering amplitudes, preprint (2023, arXiv:2303.05618), DOI: 10.48550/arXiv.2303.05618, https://arxiv.org/abs/2303.05618
  14. Kang Lu, Schur-Weyl duality for quantum toroidal superalgebras, preprint (2021, arXiv:2109.09005), DOI: 10.48550/arXiv.2109.09005, https://arxiv.org/abs/2109.09005
  15. Nick Early and Jian-Rong Li, Classification of prime modules of quantum affine algebras corresponding to 2-column tableaux, preprint (2024, arXiv:2406.16879), DOI: 10.48550/arXiv.2406.16879, https://arxiv.org/abs/2406.16879
  16. Matheus Brito and Vyjayanthi Chari, Alternating snake modules and a determinantal formula, preprint (2024, arXiv:2412.03750), DOI: 10.48550/arXiv.2412.03750, https://arxiv.org/abs/2412.03750
  17. Matheus Brito and Adriano Moura, Imaginary modules arising from tensor products of snake modules, preprint (2025, arXiv:2505.21159), DOI: 10.48550/arXiv.2505.21159, https://arxiv.org/abs/2505.21159
  18. Matheus Brito, Adriano Moura, and Clayton Silva, Totally ordered pseudo q-factorization graphs and prime factorization, preprint (2024, arXiv:2410.01519), DOI: 10.48550/arXiv.2410.01519, https://arxiv.org/abs/2410.01519
  19. Jingmin Guo, Jian-Rong Li, and Keyu Wang, Langlands branching rule for type B snake modules, preprint (2025, arXiv:2507.06570), DOI: 10.48550/arXiv.2507.06570, https://arxiv.org/abs/2507.06570
  20. Léa Bittmann, On a determinant formula for some real regular representations, preprint (2023, arXiv:2301.00784), DOI: 10.48550/arXiv.2301.00784, https://arxiv.org/abs/2301.00784
  21. Henrik Juergens, On the structure of tensor products of the finite-dimensional representations of quantum affine sln\mathfrak{sl}_n, preprint (2025, arXiv:2501.14801, master's thesis survey), DOI: 10.48550/arXiv.2501.14801, https://arxiv.org/abs/2501.14801