Folding and Mutation of Quivers

• Folding and mutation of quivers are techniques that reconfigure quiver structures via symmetry actions and exchange matrix transformations.
• The methodology employs weighted exchange matrices, group ring formulations, and tropical seed patterns to ensure compatibility in cluster algebras and categorification.
• Key insights include applications to non-crystallographic types and exceptional cluster algebras, advancing both theoretical understanding and practical computation.

Folding and mutation of quivers are intimately related operations that underpin the representation theory, categorification, and cluster algebra structures associated to both crystallographic and non-crystallographic root systems. Folding transforms a quiver by identifying vertices according to a symmetry group or partition, yielding a quotient object (often with weighted or non-integer exchange matrices), while mutation replaces one seed (quiver together with cluster variables) with another via explicitly defined combinatorial and algebraic rules. In the context of weighted quivers, group actions, and their mutations, this framework underlies recent advances in the categorification of non-integer and exceptional types, the structure of cluster algebras of finite and special finite-mutation type, and the generalization of cluster categories to semiring actions and group-ring coefficients (Duffield et al., 2023, Duffield et al., 2022, Kaufman, 2023, Kaufman et al., 22 Jan 2026).

1. Weighted Quivers, Exchange Matrices, and Foldings

A quiver $Q = (Q_0, Q_1)$ consists of a finite set of vertices $Q_0$ and arrows $Q_1$, with the standard requirement of no loops or $2$-cycles. To each quiver (possibly with positive real weights on arrows), one associates an $R$-valued exchange matrix $B \in \text{Mat}_n(R)$, defined by

$$b_{ij} = \sum_{a: i\to j} w(a) - \sum_{a: j\to i} w(a),$$

where $w: Q_1 \to R_{>0}$ assigns weights and $R$ is typically a totally ordered ring such as $\mathbb{Z}[2\cos(\pi/k)]$ or a group ring $\mathbb{Z}[G]$ (Duffield et al., 2022, Kaufman et al., 22 Jan 2026). Foldings of quivers are induced by group actions or partitions of the vertex set into orbits or groups $\{G_a\}$, generating a quotient quiver and folded exchange matrix. Given a group $G\subset S_N$ acting by quiver automorphisms, orbits $O_1,\ldots,O_n$ are formed, and the folded exchange matrix has entries in $\mathbb{Z}[G]$, encoding both multiplicity and group action (Kaufman et al., 22 Jan 2026). In weighted foldings not derived from group actions, combinatorial block decompositions, weight matrices $W$, and rescalings $P$ are required to ensure compatibilities of mutations and exchange matrix summation (Duffield et al., 2023, Duffield et al., 2022).

2. Mutation Rules in Folded and Weighted Settings

Mutation is the fundamental operation on quivers and their exchange matrices. In the standard (Fomin–Zelevinsky) setting over $\mathbb{Z}$, mutation at vertex $k$ replaces $b_{ij}$ by

$$b_{ij}' = \begin{cases} -b_{ij}, & \text{if } k \in \{i,j\}, \ b_{ij} + \frac{|b_{ik}|b_{kj} + b_{ik}|b_{kj}|}{2}, & \text{otherwise}. \end{cases}$$

(Kaufman, 2023, Duffield et al., 2022). In group ring settings, mutation generalizes to ring-valued matrices $B=(b_{ij})\in \text{Mat}_n(\mathbb{Z}[G])$ via a bilinear “circle-product”:

$$b_{ij}' = \begin{cases} -b_{ij}, & i=k \text{ or } j=k,\ b_{ij} + b_{ik}\circ b_{kj}, & \text{otherwise}, \end{cases}$$

where $a\circ b := [a]_+[b]_+ - [a]_-[b]_-$ and $[\cdot]_+$, $[\cdot]_-$ are the positive/negative part in the group ring basis (Kaufman et al., 22 Jan 2026). When folding by groups or weights, group mutations or composite mutations act on entire orbits or blocks, and the resulting folded exchange matrices mutate according to analogous skew-symmetrizable rules respecting the structure of the weights and/or group ring entries (Duffield et al., 2022, Duffield et al., 2023).

When the folding produces nonzero diagonal entries, generalized mutation (via reddening sequences) is employed; this constructs an involutive transformation, for instance by passing to framed quivers and green-to-red sequences (Kaufman et al., 22 Jan 2026). In the case of weighted unfoldings, origami pairs $(B, S)$ (where $S$ unfolds $B$) are constructed so that compatibility of block summation and sign conditions is preserved under all sequences of composite mutations (Duffield et al., 2023, Duffield et al., 2022).

3. Categorification and Semiring Actions

Weighted foldings and mutations admit a categorification via module, derived, and cluster categories associated to the unfolded and folded quivers. Given a folding $F: Q^S \to Q^B$, the module category $\operatorname{mod}K Q^S$, its bounded derived category $D^b(\operatorname{mod}K Q^S)$, and the 2-Calabi–Yau cluster category $C_F = D^b/\tau^{-1}\circ[1]$ are equipped with additional symmetry: a semiring $R_+$ (generated by Chebyshev elements in the appropriate ring $\mathbb{Z}[2\cos(\pi/k)]$ or $\mathbb{Z}[\theta]$) acts via exact (or triangulated) endofunctors (Duffield et al., 2023, Duffield et al., 2022). Auslander–Reiten (AR) quiver rows correspond to Chebyshev polynomial weights, and the semiring action shifts or labels rows accordingly.

For each indecomposable module $M\in \operatorname{mod}K Q^S$, the folded/projected dimension vector $d_F(\dimvect M)$ lies in the root system of the folded type. Under AR-translation, the dimension vector rotates by the root system angle, and rows of the AR quiver have constant length, coinciding with weights determined by the folding (Duffield et al., 2023). In cluster categories, $R_+$-tilting objects are rigid, maximal under the semiring action, and always have a number of indecomposable summands equal to the number of folded orbits. Changing a single summand in an $R_+$-tilting object corresponds precisely to mutation at the associated orbit or block.

4. Tropical Seed Patterns, $c$- and $g$-Vectors, and Block Compatibility

Tropicalizations of seeds—namely, $c$-vectors and $g$-vectors—play a critical role in tracking mutations and bases in cluster algebras. In the weighted, folded setting, the extended matrices $\widetilde{B}=(B | I)$ undergo block-structured mutations compatible with the underlying folding or group action (Duffield et al., 2022). Under weighted unfoldings, each block of the unfolded $C$-matrix is the regular representation of a single semiring element. The projection (Chebyshev projection $d_F$) recovers the folded $C$-matrix, ensuring that tropical mutation compatibility is preserved (the “commuting cube” property).

Folded $c$-vectors are sign-coherent and bijective with the (real) roots of the corresponding non-crystallographic root system. All $g$-vectors lie in the dual root cone. These results generalize Gabriel's theorem for simply-laced quivers to the folded, weighted, and group ring settings (Duffield et al., 2022).

5. Special and Standard Foldings, Examples, and New Phenomena

Foldings can be classified as standard (folded matrix skew-symmetrizable) or special (non-skew-symmetrizable). Special foldings, such as those realized in the exceptional quivers $X_6$ and $X_7$, produce new finite mutation type cluster algebras not arising from skew-symmetrizable matrices. They can be constructed via group partitions compatible with the involution or surface covering constructions, resulting in cluster complexes isomorphic to generalized permutohedra or Coxeter permutohedra of Dynkin type (Kaufman, 2023). A typical example is folding $\widehat{E}_6^{(2,2)}$ along an involution to obtain $X_7$, with explicit checks that FC1 and FC2 (no in-group arrows, stability under group mutation) hold, but the folded matrix is not skew-symmetrizable.

In the context of surface cluster algebras, folded algebras corresponding to triangulations with self-folded triangles are constructed via two-fold covers, and group mutations upstairs realize special flip rules and new variables for self-folded arcs. The resulting cluster algebras inherit a new (folded) mutation structure and modular group action isomorphic to $W(X_n) \rtimes \mathrm{Aut}(X_n)$, where $W(X_n)$ is the Weyl group (Kaufman, 2023).

6. Generalization to Group Rings and Non-Cycle-Free Mutations

The formalism of matrices over group rings $\mathbb{Z}[G]$ enables further unification of folding and mutation, encompassing both classical folding and weighted unfoldings. In this framework, folding a quiver with symmetry under $G$ yields an exchange matrix whose entries record both the number of arrows and the group elements mediating them. The mutation rule generalizes accordingly, and when orbits are not cycle-free, generalized mutation sequences (including reddening or green-to-red) are used to define and compute the mutated matrix (Kaufman et al., 22 Jan 2026). This approach yields explicit descriptions of cluster structure on moduli spaces (with symmetry), cyclic Grassmannian cluster algebras, and the integration with Donaldson–Thomas transformations.

Detailed worked examples (e.g., folding two disjoint $3$-cycles under $\mathbb{Z}/2$-action) illustrate the explicit computation of group ring exchange matrices and their mutations, with the mutation on the folded matrix mirroring simultaneous mutations on the covering quiver.

7. Theoretical and Structural Implications

Weighted folding and mutation provide a systematic mechanism for passing from simply-laced to non-simply-laced cluster and representation-theoretic objects and for constructing categorifications of non-integer or non-crystallographic types ($I_2(2n)$, $I_2(2n+1)$, $H_3$, $H_4$). Semiring actions, Chebyshev polynomials, and block-compatibility phenomena introduce new structures not present in simply-laced or integer coefficient settings. Special foldings and group ring methodologies extend the reach of cluster theory to finite-mutation types outside the classical classification and connect cluster combinatorics to more general Coxeter–Weyl objects. Open problems include the complete classification of finite mutation type group ring matrices and deeper connections to quantum cluster structures over noncommutative bases (Duffield et al., 2023, Duffield et al., 2022, Kaufman, 2023, Kaufman et al., 22 Jan 2026).