Category of Chains: Categorical Constructions

• Category of Chains is a term describing categorical structures where chain-like diagrams or objects, such as chain complexes and linear orders, serve as the primary focus across various mathematical fields.
• The construction in symplectic topology, where chain algebras control wrapped Fukaya categories, and in module theory, with filtered modules yielding Krull–Schmidt classifications, illustrates its methodological diversity.
• In order theory and anyon chain stabilization, chain categories capture ordering properties and fusion symmetries, enabling deep interconnections between algebraic logic, topology, and operator algebras.

In contemporary mathematical literature, “Category of Chains” is not a single universally fixed construction. The expression is used for several categorical frameworks in which a chain-like structure becomes the primary object of study: singular chains on topological stacks, chain algebras controlling Fukaya categories and Lie-group representations, categories of chain bundles and bundle-derived diagrams, categories of linear orders under embeddings, additive categories of filtered modules, algebraic-logical categories of MTL-chains, and groupoids of stabilized anyon-chain realizations (Coyne et al., 2015, Laflamme et al., 2014, Campanini, 16 Apr 2025, Castiglioni et al., 2017, Bunner et al., 20 May 2026).

1. Homotopy-theoretic and chain-algebra meanings

One prominent usage treats a “category of chains” as a functorial assignment from geometric objects to chain complexes. For topological stacks, the singular simplicial set functor is extended to

$$\mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{sSet},$$

and composition with normalized chains gives

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$

The construction is defined by $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$, generalizes the classical singular functor on spaces, respects weak equivalences of Serre stacks, and sends morphisms that are both Serre and Reedy fibrations to Kan fibrations of simplicial sets (Coyne et al., 2015). In this sense, the category of chains is the target category $\mathrm{Ch}_R$, together with the homotopy-theoretically well-behaved functor $C_*(-;R)$.

A second usage appears in symplectic topology. For a closed, smooth, oriented manifold $Q$, the “category of chains” is the $A_\infty/$dg algebra of chains on the based loop space $\Omega_qQ$ together with its category of perfect modules. In this setting,

$$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$

and the endomorphism algebra of a cotangent fibre satisfies

$CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$

The paper’s main theorem states that a cotangent fibre split-generates the wrapped Fukaya category, so the chain algebra of $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$0 controls the entire split-closed derived wrapped Fukaya category of $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$1 (Abouzaid, 2010).

For simply connected Lie groups, the phrase refers to the dg-algebra $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$2 of smooth singular chains and the module category $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$3. The central structural result is a natural monoidal equivalence

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$4

where $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$5 is the dg-Lie algebra universal for the Cartan relations. This extends the classical correspondence between representations of $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$6 and representations of $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$7 to a chain-level setting (1908.10460).

A further categorical use occurs in virtual-chain theory, where $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$8 is explicitly the category of $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$9-graded cochain complexes over a commutative ring $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$0, equipped with tensor product and cubical enrichment. In that framework, “virtual chains” are encoded as a symmetric monoidal natural transformation

$$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$1

from a stratum cochain functor on Kuranishi presentations to the constant chain-complex functor (Abouzaid, 2022). Closely related homotopical work studies the simplicial coalgebra of chains $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$2, constructing three model structures on reduced simplicial sets and on connected simplicial cocommutative $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$3-coalgebras; when $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$4 is algebraically closed, the chains functor is homotopically full and faithful in each of the three corresponding Quillen adjunctions (Raptis et al., 2022).

2. Chain bundles and bundle-derived chain categories

Another major meaning is genuinely object-theoretic: a category whose objects are themselves chains. In the construction called the category of chain bundles, a chain bundle in a category $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$5 with zero is a sequence of objects together with the entire homsets between successive vertices, identities, and all possible composites. The resulting category $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$6 has chain bundles as objects and functorial chain-bundle maps as morphisms. If $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$7 has subobjects, then $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$8 is again a category with subobjects; if $$\mathrm{Sing}(\mathcal X)=\mathrm{Diag}(N(\mathcal X_A))$$9 has factorization, then a full-morphism-map subcategory $\mathrm{Ch}_R$0 admits factorization, and selecting at most one morphism from each adjacent homset produces a category of chains inside $\mathrm{Ch}_R$1. In an abelian category, imposing $\mathrm{Ch}_R$2 recovers the usual category of chain complexes (Romeo et al., 2020).

A closely related formulation defines a chain bundle as a sequence

$\mathrm{Ch}_R$3

where each $\mathrm{Ch}_R$4 is a subset of $\mathrm{Ch}_R$5. A morphism $\mathrm{Ch}_R$6 is a sequence $\mathrm{Ch}_R$7 satisfying

$\mathrm{Ch}_R$8

Within this framework, the paper studies subchain bundles, factorization of chain-bundle maps, products, coproducts, kernels, cokernels, universal arrows, and the specialization to chains with a single chosen morphism $\mathrm{Ch}_R$9 at each level (Romeo et al., 2021).

The same pattern is then transported to categories of bundles in the sense of Nambooripad’s categories with subobjects. The 2026 paper constructs categories of chains of bundles, chains of fibres, and chains of long exact sequences arising from fibrations, and verifies the “category with subobjects” axiom in each workable case. For chains of bundles, a morphism

$C_*(-;R)$0

between

$C_*(-;R)$1

and

$C_*(-;R)$2

is required to satisfy

$C_*(-;R)$3

Subobjects are subchains defined by termwise subbundles and compatible inclusion equations. The paper also treats fibrewise chains, long exact sequences associated with fibrations, and jet-bundle chains, but it explicitly notes that chains of fibres over different base points need not admit a canonical subobject relation (Romeo et al., 27 May 2026).

3. Chains as linear orders and bi-embedding classes

In order theory, a chain is a linear order. The relevant category is the category $C_*(-;R)$4 of chains with embeddings, or equivalently the quasiordered category under the preorder $C_*(-;R)$5 meaning that there exists an embedding $C_*(-;R)$6. Two chains are equimorphic, written $C_*(-;R)$7, when each embeds in the other, and the principal invariant is

$C_*(-;R)$8

the set of isomorphism types of chains equimorphic to $C_*(-;R)$9; its cardinality is written $Q$0 in the paper (Laflamme et al., 2014).

The central theorem is a sharp dichotomy: for every chain $Q$1, $Q$2 is either $Q$3 or at least $Q$4. More refined statements describe when the small-cardinality case occurs. If $Q$5 is a finite sum of ordinals and reverse ordinals, then $Q$6. For scattered chains below continuum, the exact structure is controlled by finite sums of surordinals and reverse surordinals, and the paper proves a full characterization for all $Q$7: $Q$8 has $Q$9 iff it decomposes as a dense sum $A_\infty/$0 with each $A_\infty/$1 scattered, $A_\infty/$2 for all but finitely many $A_\infty/$3, $A_\infty/$4, and every self-embedding of $A_\infty/$5 preserves each $A_\infty/$6 setwise (Laflamme et al., 2014).

This usage is conceptually distinct from chain complexes or singular chains. Here, “category of chains” means a category whose objects are linear orders, and whose morphisms are order-embeddings. The paper’s categorical perspective emphasizes the quotient of $A_\infty/$7 by equimorphism, viewed as the poset of bi-embedding classes, and studies how far equimorphism is from isomorphism inside a given class (Laflamme et al., 2014).

4. Filtered modules and the additive category $A_\infty/$8

In module theory, the phrase denotes an additive category of modules equipped with a fixed finite chain of submodules. For a fixed $A_\infty/$9, the category $\Omega_qQ$0 has objects

$\Omega_qQ$1

and morphisms are $\Omega_qQ$2-module homomorphisms preserving the filtration at every level. The category is additive, with direct sums defined componentwise. Its distinguished full subcategory $\Omega_qQ$3 consists of those objects whose successive factors

$\Omega_qQ$4

are nonzero uniserial right $\Omega_qQ$5-modules (Campanini, 16 Apr 2025).

For an object $\Omega_qQ$6, the endomorphism ring

$\Omega_qQ$7

is the ring of chain-preserving endomorphisms of the underlying module. The filtration induces a canonical surjective ring homomorphism

$\Omega_qQ$8

whose kernel is a nilpotent ideal $\Omega_qQ$9 with $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$0. In the uniserial case, the paper defines for each $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$1 two completely prime ideals,

$$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$2

and proves that every proper right ideal and every proper left ideal of $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$3 is contained in one of these $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$4 ideals. Consequently, $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$5 has at most $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$6 maximal two-sided ideals, and

$$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$7

for suitable division rings $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$8 (Campanini, 16 Apr 2025).

The same framework yields a Krull–Schmidt-type classification for finite direct sums in $$D^\pi \mathcal W_b(T^*Q)\simeq \operatorname{Tw}^\pi\!\big(C_{-*}(\Omega_qQ;k)\big)\simeq \operatorname{Perf}\!\big(C_{-*}(\Omega_qQ;k)\big),$$9. Two finite direct sums are isomorphic in $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$0 exactly when, for each $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$1 and each $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$2, the corresponding multisets of $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$3-th monogeny and epigeny classes agree after suitable permutations. This extends the classical classification of finite direct sums of uniserial modules by monogeny and epigeny classes (Campanini, 16 Apr 2025).

5. MTL-chains, labeled forests, and generalized chain products

In algebraic logic, a chain is a totally ordered MTL-algebra or semihoop, and the “category of chains” appears both as a category of finite archimedean MTL-chains and as a source of decomposition theory. One paper introduces the category $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$4 of finite archimedean MTL-chains and a skeleton $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$5, then uses $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$6 as the label category for finite labeled forests. From a finite MTL-algebra $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$7, one forms a labeled forest $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$8, where $CW^*(L_q,L_q)\simeq C_{-*}(\Omega_qQ;k).$9 is the poset of join-irreducible idempotents and the label at $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$00 is the archimedean quotient

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$01

A contravariant functor $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$02 sends finite MTL-algebras to labeled forests, and a contravariant functor $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$03 reconstructs a forest product. On the representable subcategory, these functors induce a dual equivalence between representable finite MTL-algebras and finite labeled forests (Castiglioni et al., 2017).

The same paper gives a sheaf-theoretic interpretation of these products. For a forest $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$04 and chains $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$05, the forest product is an MTL-algebra; it is an MTL-chain exactly when $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$06 is totally ordered. Over the Alexandrov topology $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$07, the associated presheaf is a sheaf of MTL-chains, and for finite labeled forests the algebra can be reconstructed recursively from ordinal sums along tree edges and direct products across connected components (Castiglioni et al., 2017).

A second paper places these constructions in the semi-abelian category $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$08 of semihoops. It proves that every finite locally unital MTL-chain is an ordinal sum of archimedean MTL-chains. The proof uses the finite chain of local units

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$09

and iteratively splits the chain via quotients and filters determined by idempotents. The paper then introduces generalized chain products: a family of finite chains $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$10 indexed by a finite MTL-chain $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$11, together with monotone maps

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$12

satisfying zeros-as-absorbing-elements, joint associativity, joint commutativity, and a global unit condition. Any such generalized chain product defines an MTL-algebra $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$13 fitting into a split short exact sequence

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$14

and ordinal sums of finite locally unital MTL-chains arise as particular cases (Castiglioni et al., 2018).

6. Stabilized anyon chains and symmetry groupoids

In operator-algebraic and topological-order settings, “category of chains” refers to stabilized anyon chains and the morphisms between their symmetry realizations. For an indecomposable unitary multi-fusion category $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$15 and a strong tensor generator $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$16, the anyon chain has quasi-local algebra

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$17

and its stabilization is obtained by tensoring local algebras with

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$18

The principal structural theorem states that if $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$19 and $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$20 are Morita equivalent indecomposable unitary multi-fusion categories, then

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$21

via a bounded spread isomorphism. In particular, any stabilized anyon chain is bounded spread isomorphic to the “free boson chain” $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$22 (Bunner et al., 20 May 2026).

This stabilization yields a universal realization theorem for fusion-category symmetries. The paper proves that for a fusion spin chain $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$23,

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$24

and deduces that every unitary fusion category $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$25 can be realized as a symmetry category on the tensor-product quasi-local algebra $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$26. It also proves a stable uniqueness theorem: if $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$27 and $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$28 are anyon-chain models realizing the same symmetry category $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$29, then there exists a bounded spread isomorphism

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$30

with $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$31 (Bunner et al., 20 May 2026).

The paper makes this into an explicit categorical statement. For fixed $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$32, one considers the groupoid whose objects are stabilized symmetry inclusions

$$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$33

and whose morphisms are symmetry-compatible locality-preserving unitaries, equivalently bounded spread isomorphisms preserving the distinguished boundary subalgebras. The resulting category is a groupoid with a single isomorphism class per unitary fusion category $$C_* = N\circ \mathrm{Sing}\colon \mathrm{TopStacks}\to \mathrm{Ch}_R.$$34. A parallel classification theorem for Levin–Wen boundary algebras states that two fusion spin chains are stably equivalent iff their Drinfeld centers are equivalent as unitary braided fusion categories (Bunner et al., 20 May 2026).

The expression “Category of Chains” therefore functions as a family of technically specific categorical constructions rather than a single doctrine. Depending on context, it may designate a functor to chain complexes, a category of chain-shaped diagrams, a category of linear orders, a filtered-module category, an algebraic-logical category of MTL-chains, or a groupoid of stabilized lattice realizations. The common feature is categorical control of a chain structure, but the mathematical content is determined entirely by the ambient theory in which that structure is defined.