Arbitrary Space Decompositions

• Arbitrary Space Decompositions are a framework for encoding how mathematical structures compose and decompose by relaxing classical constraints.
• They apply to diverse areas such as binomial posets, Hall algebras, and combinatorial Hopf algebras, offering concrete methods for constructing algebraic structures.
• Their flexible formalism bridges combinatorics, homotopy theory, and higher category theory by leveraging simplicial objects and pullback conditions to model complex decompositions.

Arbitrary space decompositions, in the sense of decomposition spaces and their generalizations, encompass a unifying framework in homotopical and combinatorial mathematics for encoding how structures can be composed and decomposed. A decomposition space is a simplicial object satisfying the weakened “decomposition” (or 2-Segal) condition, and supports the direct construction of incidence (co)algebras, Möbius inversion, and a range of algebraic structures fundamental to combinatorics, category theory, and their applications. The flexibility of this framework enables the treatment of spaces (discrete, groupoid-valued, or homotopical) far beyond the classical settings of finite posets or categories, and underlies modern approaches to combinatorial Hopf algebras, algebraic topology, and higher category theory.

1. Definition and Properties

A decomposition space is a simplicial object (often a simplicial groupoid or set) that relaxes the Segal condition—traditionally encoding composition in categories—so as to instead model decomposition. This is formalized as follows: For a simplicial groupoid $X: \Delta^{\mathrm{op}} \to \mathrm{Grpd}$, the space $X$ is a decomposition space if it sends every active-inert pushout square in the simplex category $\Delta$ to a pullback square in $\mathrm{Grpd}$. The active-inert terminology differentiates ‘compositional’ versus ‘insertion’ (distance-preserving) simplicial maps. Unlike Segal spaces, which encode associative composition (every $n$-simplex is determined by a string of $1$-simplices), decomposition spaces allow structures whose decomposition is well-defined even if composition is not.

Every Segal space is a decomposition space, but not conversely. The formalism requires only functoriality under the simplex category and the pullback condition described above, conferring a rich and flexible calculus for encoding combinatorial, algebraic, and topological data.

2. Examples and Applications

Decomposition spaces are realized in an extensive collection of classical and contemporary contexts:

• Binomial posets and incidence algebras: Viewing the Boolean lattice of subsets as a decomposition space yields the binomial incidence algebra—binomial coefficients arise naturally from counting decompositions of sets into disjoint subsets.
• Hall algebras and the Waldhausen S-construction: The S-construction on an abelian category provides a decomposition space whose associated incidence coalgebra is the Hall algebra, integral to representation theory and quantum groups. The section coefficients generalize the $q$-binomial coefficients in the context of vector spaces over finite fields.
• Hopf algebras of combinatorial structures: Examples include Schmitt’s chromatic Hopf algebra of graphs, which encodes the induced subgraph decompositions; the Faà di Bruno bialgebra arising from set partitions and composition of power series; and bialgebras from trees and operads (e.g., the Butcher–Connes–Kreimer Hopf algebra).
• Directed restriction species and poset/graph structures: Decomposition spaces formalize and extend the combinatorial structures arising from directed graphs and posets, including their generalization to restriction species and their "directed" versions.

The framework encompasses new algebraic possibilities; for instance, at the “objective” (groupoid) level, power series products (Cauchy, shuffle, Dirichlet, etc.) arise from various decomposition space structures on species.

3. Mathematical Structures: Algebras and Products

A decomposition space $X$ always yields an incidence coalgebra on $\mathrm{Grpd}_{/X_1}$, with comultiplication defined by the span: $$X_1 \xleftarrow{d_1} X_2 \xrightarrow{(d_2,d_0)} X_1 \times X_1$$ On cardinalizing, this encodes the familiar comultiplication $\Delta(f) = \sum_{ab=f} a \otimes b$ for categories, and more generally encodes the “decomposition” of $f$ into pieces $a$ and $b$.

Incidence algebras are obtained by duality, with the convolution product explicitly described in terms of decompositions. Classical products—Cauchy (ordinary species), shuffle (L-species), Dirichlet (arithmetic species), and external products for $q$-species—are realized in this way. In each case, the resulting algebraic structure, when cardinalized, reproduces classical power series or combinatorial products, now justified at the higher categorical/groupoid level.

Section coefficients (the structure constants of incidence coalgebras) are given by explicit formulas involving automorphism groups: $$c_{a,b}^f = \frac{|\operatorname{Aut}(f)| \cdot |\{\varphi \mid \ell_{a,b}(\varphi)\simeq f\}| }{|\operatorname{Aut}(a)||\operatorname{Aut}(b)|}$$

4. Connections to Other Mathematical Theories

Decomposition spaces are deeply connected to operad theory, restriction species, and higher categorical structures:

• Operads and polynomial functors: Decomposition spaces generalize the combinatorics of colored operads and polynomial monads. For example, the decomposition space of $P$-trees encodes the combinatorics of trees decorated by a polynomial functor $P$, which is fundamental in renormalization and universal algebra.
• Hall algebras and higher algebraic structures: The S-construction in $K$-theory and related categorical tools manifest as decomposition spaces, providing a unified perspective for Hall-type algebras and their applications in representation theory.
• Directed restriction species: Schmitt’s restriction species framework is realized within decomposition spaces, particularly their directed variants, linking to structures on posets and digraphs.

These connections position decomposition spaces as a bridge among combinatorics, category theory, homotopy theory, and representation theory, with their formalism concretely relating algebraic and combinatorial data.

5. Computational and Möbius Theoretic Aspects

Decomposition spaces naturally support Möbius inversion when suitable finiteness and completeness conditions are present. The Möbius function, generalizing the classical incidence algebra Möbius function, is expressed as an alternating sum over nondegenerate simplices: $\mu = \sum_{n=0}^\infty (-1)^n \Phi_n$ where $\Phi_n$ counts nondegenerate $n$-simplices with a given long edge. The convolution identity

$\zeta * \mu = \varepsilon = \mu * \zeta$

is universal across all decomposition spaces, echoing the Möbius inversion in poset and category theory.

At the groupoid level, some cancellations in the Möbius function can be realized as actual bijections (the "objective method"). However, upon taking cardinality, further algebraic cancellations may appear that do not have an underlying combinatorial bijection, illuminating a distinction between combinatorial and algebraic phenomena.

6. Implications, Generalizations, and Future Directions

Decomposition spaces provide a universal, flexible framework for incidence (co)algebra theory, greatly generalizing the scope beyond posets and categories. They allow the direct realization of combinatorial (co)algebra structures in contexts replete with symmetry and quotienting, where composition is not always defined or unique.

The category-theoretic approach, focusing on groupoids and functors, advances objective combinatorics and leads to conceptual insights into classical structures—such as symmetric and quasi-symmetric functions, binomial analogues, and their presence in $q$-deformations and Hall algebras.

Expected future research includes the paper of decomposition spaces for broader families (including infinite, higher, and $\infty$-categories), further investigation of the algebraic significance of cancellations at cardinality, and the systematic exploration of their connections to polytopes, species, and comodule bialgebras across mathematics and mathematical physics.

---

Structure	In Decomposition Spaces
Coalgebra	Incidence coalgebra $(\mathrm{Grpd}_{/X_1}, \Delta, \varepsilon)$
Algebra	Incidence algebra (functions dual to above, with convolution product)
Möbius inversion	Universal formula using non-degenerate simplices
Products (Species)	Cauchy, shuffle, Dirichlet, and other combinatorial products
Examples	Binomial posets, Hall algebras, Schmitt’s Hopf algebra, Faà di Bruno bialgebra
Connections	Operad theory, restriction species, higher category theory

---

Decomposition spaces and their incidence (co)algebras represent a universal framework for encoding the mechanisms of decomposition in combinatorial, topological, and algebraic objects, with a reach extending across classical combinatorics, higher category theory, representation theory, and beyond. Their conceptual and structural generality provides a foundation for the development and analysis of diverse algebraic and enumerative structures, and for the ongoing integration of combinatorial and categorical approaches in mathematics.