Simplicial Complex Extensions

• Simplicial complex extensions are processes that embed a given complex into a larger structured complex while preserving key topological, combinatorial, and algebraic properties.
• They utilize methods like controlled gluing, expansion functors, and universal constructions to ensure properties such as Cohen–Macaulayness, shellability, and vertex-decomposability.
• These extensions are central to diverse applications in topological data analysis, geometric group theory, and representation learning.

A simplicial complex extension is any process or construction that embeds a given simplicial complex as a subcomplex of a larger, often more structured, complex, while controlling the topological, combinatorial, metric, or algebraic properties of the result. Extensions arise in various contexts: gluing operations, expansions, universal constructions, representation-theoretic problems, and learning applications. The study of such extensions provides fundamental tools for combinatorial commutative algebra, topological data analysis, geometric group theory, and higher category theory.

1. Extension via Gluing: Controlled Topological and Algebraic Properties

Farrokhi, Shamsian, and Yazdanpour introduced three general classes of glue-based extensions for simplicial complexes, modeled via independence complexes of hypergraphs (Ghouchan et al., 2021):

• Hybrid (multi-piece) gluing: Given a base complex, independently glue additional pieces (complexes or hypergraphs) along specified vertex-sets. The dimension, purity, (sequential) Cohen–Macaulayness, shellability, and vertex-decomposability of the extended complex are completely characterized in terms of the corresponding properties of the central piece and glued components. The preservation of desirable properties requires combinatorial conditions such as the Proper-Independence-Property (PIP) for the size-truncation and links.
• Corona-type (single-vertex gluing): Each glued piece attaches along a single base vertex. This generalizes classical corona (whiskering) constructions and preserves vertex-decomposability and Cohen–Macaulayness under mild hypotheses.
• Doubling constructions: Attach pieces and their minimal transfer complexes, increasing the dimension of the ambient complex while tightly controlling purity and decomposability.

For all three classes, recursive application of the shedding-vertex criterion reduces proofs to local verifications and joins of properties. These gluing techniques unify many classical ad hoc constructions within a flexible extension framework and enable the systematic creation of large classes of pure, shellable, and Cohen–Macaulay complexes containing prescribed subcomplexes (Ghouchan et al., 2021).

2. Expansion Functors: Algebraic and Combinatorial Invariance

Given a simplicial complex $\Delta$ over vertices $X = \{x_1, ..., x_n\}$ and a vector $\alpha = (s_1, ..., s_n)$, the expansion $\Delta^{\alpha}$ is defined by replacing each vertex $x_i$ with $s_i$ distinct copies and lifting each face of $\Delta$ to all possible selections of corresponding copies (Moradi et al., 2016, Moradi et al., 2017, Rahmati-asghar et al., 2015). This expansion preserves core algebraic and combinatorial properties:

• Vertex decomposability, shellability, Buchsbaumness, $k$-decomposability: Preserved by expansion, with explicit inductive arguments reducing to shedding-vertex and link behavior (Moradi et al., 2016, Rahmati-asghar et al., 2015).
• Cohen–Macaulayness and sequential Cohen–Macaulayness: $\Delta$ is Cohen–Macaulay (resp. sequentially Cohen–Macaulay) iff every expansion $\Delta^{\alpha}$ is so for every field $X = \{x_1, ..., x_n\}$0, via Reisner’s criterion and surjective homology maps (Rahmati-asghar et al., 2015, Moradi et al., 2017).
• Homological invariants: Regularity and depth are controlled; e.g., $X = \{x_1, ..., x_n\}$1 and $X = \{x_1, ..., x_n\}$2 for pure $X = \{x_1, ..., x_n\}$3-complexes with $X = \{x_1, ..., x_n\}$4 (Moradi et al., 2017).
• Stanley–Reisner and facet ideals: The expansion yields monomial ideals with linear quotients inherited from the original complex and provides explicit Betti-splitting decompositions (Rahmati-asghar et al., 2015).

The combinatorial effect is that expansions generalize vertex-doubling and whiskering operations in graph theory to arbitrary dimension, producing large Cohen–Macaulay or shellable complexes with prescribed combinatorics and algebraic invariants.

3. Universality and Homogeneity: The Rado Simplicial Complex

The unique countable Rado simplicial complex $X = \{x_1, ..., x_n\}$5 is characterized by a high-dimensional extension axiom (E): for any finite subset $X = \{x_1, ..., x_n\}$6 and induced subcomplex $X = \{x_1, ..., x_n\}$7, there exists $X = \{x_1, ..., x_n\}$8 such that $X = \{x_1, ..., x_n\}$9 (Farber et al., 2019). Key properties:

• Universality: Every countable complex embeds as an induced subcomplex of $\alpha = (s_1, ..., s_n)$0.
• Homogeneity: Every finite induced subcomplex isomorphism extends to a global automorphism.
• Topological contractibility: The geometric realization $\alpha = (s_1, ..., s_n)$1 is homeomorphic to the infinite-dimensional simplex and is contractible.
• Robustness: Deleting any finite set of simplices yields a complex isomorphic to $\alpha = (s_1, ..., s_n)$2; all links are Rado complexes.

Extensions to $\alpha = (s_1, ..., s_n)$3 exhibit maximal symmetry and “ampleness”: every finite pattern can be extended arbitrarily, and the Rado complex serves as the Fraïssé limit of finite complexes (Farber et al., 2019). This universality is directly analogous to the Rado graph and underpins probabilistic and random constructions in higher dimensions.

4. Simplicial Extensions and k-Decomposability: Facet-by-Facet Growth

The problem of extending a pure $\alpha = (s_1, ..., s_n)$4-dimensional complex to a complete $\alpha = (s_1, ..., s_n)$5-skeleton, while preserving decomposability constraints, is a central topic. The notion of $\alpha = (s_1, ..., s_n)$6-decomposability (faces of dimension at most $\alpha = (s_1, ..., s_n)$7 can be recursively shed while preserving purity and lower-dimensional decomposabilities) generalizes shellability (Ghosal et al., 6 Aug 2025):

• Vertex-decomposable ($\alpha = (s_1, ..., s_n)$8) and $\alpha = (s_1, ..., s_n)$9-decomposable complexes can always be extended one facet at a time—vertex-decomposable complexes to $\Delta^{\alpha}$0, and $\Delta^{\alpha}$1-decomposable complexes to $\Delta^{\alpha}$2 with added coning vertices if needed (Ghosal et al., 6 Aug 2025).
• The process involves a two-stage extension: coning (adding $\Delta^{\alpha}$3 vertices to guarantee pure links under addition) and skeleton filling (attaching remaining facets required to complete the skeleton).
• The main theorem for $\Delta^{\alpha}$4-decomposable complexes resolves the $\Delta^{\alpha}$5 case of Simon’s extendability conjecture, showing that the $\Delta^{\alpha}$6-decomposability class is sufficient for controlled facetwise extension (Ghosal et al., 6 Aug 2025).

Open challenges include determining whether the need for added vertices (coning) can be eliminated for $\Delta^{\alpha}$7 and establishing analogous results for higher $\Delta^{\alpha}$8-decomposability or arbitrary shellable complexes.

5. Extensions in Geometry: Moment Curve and Triangulation Extendability

Lee and Nevo show that any geometric complex of dimension $\Delta^{\alpha}$9 with vertices on the moment curve $x_i$0 can always be extended to a triangulation $x_i$1 of the cyclic polytope $x_i$2 on the same vertex set (Lee et al., 18 Nov 2025):

• Existence: For $x_i$3, every non-overlapping geometric simplicial complex (i.e., no face overlaps except in genuine faces) on the moment curve is extendable to a triangulation of the ambient cyclic polytope; for $x_i$4 this extension is generally impossible.
• Constructive process: Extension proceeds by a height-order and greedy boundary-filling, relying on the Stasheff–Tamari poset properties for lower dimensions.
• Algebraic application: For $x_i$5, Oppermann–Thomas’ correspondence equates full triangulations with cluster tilting objects in higher Auslander algebras. Consequently, every maximal rigid object in the $x_i$6 cluster category is cluster-tilting, bridging geometry and homological algebra (Lee et al., 18 Nov 2025).

This demonstrates how geometric extension problems have algebraic significance well beyond their initial combinatorial statement.

6. Truncated and Boolean Representable Extensions

Boolean representability and truncations form a broad extension theory for finite simplicial complexes:

• Truncated Boolean Representable Simplicial Complexes (TBRSC): $x_i$7 is TBRSC if $x_i$8 for some Boolean representable complex $x_i$9, where $s_i$0 truncates all faces of cardinality $s_i$1 (Margolis et al., 2019).
• Extension operator $s_i$2: $s_i$3 is constructed canonically, and $s_i$4 is then a truncation. This forms an adjunction relating arbitrary simplicial complexes to the class admitting matroid-like geometry (closure, flats, ranks, homology, etc.).
• Closure properties: Joins of paving TBRSCs are again paving TBRSCs, and every paving TBRSC admits a decomposition into minimal “block” components corresponding to its $s_i$5-closed sets (Margolis et al., 2019).
• Topological control: For a TBRSC $s_i$6 of dimension $s_i$7, $s_i$8 is homotopy equivalent to a wedge of spheres, and the fundamental group is free.

TBRSCs maximize the class of complexes for which matroid-theoretic and geometric techniques are meaningful, with computable criteria for recognition and extension (Margolis et al., 2019).

7. Further Directions: Metrics, Representation Learning, and Simplicial Structure

• Metric extensions: Any metric on the vertices of a simplicial complex $s_i$9 (satisfying a domination condition by the 1-skeleton word metric) extends to a bona fide metric on the whole realization via a bilinear formula and an $\Delta$0-path metric, correcting earlier shortcomings in constructions used in geometric group theory (Mole, 2013).
• Representation learning: Entire simplicial complexes can be embedded into universal feature spaces by geometric message passing schemes generalizing GNNs to all simplex dimensions, with message propagation aggregating adjacency and coadjacency (coface) data and higher-order combinatorics. The resulting embedding maps preserve complex-to-complex proximity and are robust to indexing and orientation, enabling shape-level inference in geometry applications (Hajij et al., 2021).
• Simplicial category theory: The category of $\Delta$1-complexes in an additive category admits a strict simplicial structure whose décalage is pointwise equivalent to the lax nerve of chain complexes, providing a categorification of triangulated homotopy categories and tracing the axioms of triangulated categories to simplicial origins (Mirmohades, 2014).

These developments extend the concept of simplicial complex extensions well beyond classical algebraic topology and combinatorics, radiating into metric geometry, representation theory, categorical homological algebra, and machine learning.

---

References:

• Gluing and topological/combinatorial properties: (Ghouchan et al., 2021)
• Expansion and algebraic invariants: (Moradi et al., 2016, Moradi et al., 2017, Rahmati-asghar et al., 2015)
• Universality and Rado complex: (Farber et al., 2019)
• k-Decomposability and extendability: (Ghosal et al., 6 Aug 2025)
• Triangulations on the moment curve: (Lee et al., 18 Nov 2025)
• TBRSCs and Boolean representability: (Margolis et al., 2019)
• Metric extension: (Mole, 2013)
• Representation learning on complexes: (Hajij et al., 2021)
• Simplicial structure in category theory: (Mirmohades, 2014)