Dehn-Sommerville Manifolds

• Dehn-Sommerville manifolds are finite abstract simplicial complexes where local face‐enumerative symmetries enforce relationships akin to Poincaré duality.
• They are defined via recursive unit sphere conditions or as homology manifolds, generalizing classical discrete manifold and polytope theory.
• Advanced algebraic and topological methods, including barycentric refinement and toric constructions, extend these symmetries to broader combinatorial settings.

Searching arXiv for recent and foundational papers on Dehn–Sommerville manifolds and related generalizations. Dehn–Sommerville manifolds are finite abstract simplicial complexes defined so that local combinatorial data enforce the same Dehn–Sommerville symmetries that classically arise for simplicial polytopes, spheres, and Eulerian or homology-manifold settings. In recent work, the term has been used in at least two closely related but nonidentical senses: as a recursively defined class of finite abstract simplicial complexes generalizing discrete manifolds (Knill, 20 Aug 2025), and as a label for simplicial complexes that are homology manifolds, possibly with boundary, in the context of generalized Dehn–Sommerville relations (Ceballos et al., 2021). Across these settings, the common theme is that the face-enumerative invariants—especially the $f$-vector, $h$-vector, and their refinements—satisfy linear symmetries or reciprocity formulas that make roughly half of the face data redundant, often reflecting Poincaré duality or its combinatorial analogues (Ayzenberg, 2015).

1. Definitions and competing usages

In the recursively defined finite-complex setting, a Dehn-Sommerville $q$-manifold is a $q$-dimensional finite abstract simplicial complex $G$ such that every unit sphere $S(x)$ is a Dehn-Sommerville $(q-1)$-manifold of Euler characteristic $1-(-1)^q$ (Knill, 20 Aug 2025). Here the unit sphere is $S(x)=\delta U(x)$, where $U(x)=\{y\in G\mid x\subset y\}$ in the finite Alexandrov topology, and the base case is that the empty complex is the $h$0-dimensional Dehn-Sommerville sphere (Knill, 20 Aug 2025). A Dehn-Sommerville $h$1-sphere is a Dehn-Sommerville $h$2-manifold additionally satisfying the Euler-gem formula $h$3 (Knill, 20 Aug 2025).

In a distinct but standard enumerative-topology usage, a Dehn–Sommerville manifold is a finite simplicial complex that is a homology manifold, possibly with boundary, over a field $h$4 (Ceballos et al., 2021). In this framework, the Dehn–Sommerville relations are derived from face-link multiplicities and interior-face enumeration rather than from a recursive unit-sphere axiom (Ceballos et al., 2021).

These usages overlap in motivation but differ in scope. The recursive definition is explicitly presented as a generalization of discrete manifolds and does not require the unit spheres to be spheres; instead, they need only be Dehn-Sommerville manifolds of one lower dimension (Knill, 20 Aug 2025). By contrast, the homology-manifold usage places the theory inside the established Buchsbaum/Cohen–Macaulay framework for simplicial complexes (Ayzenberg, 2015, Ceballos et al., 2021).

2. Face-enumerative symmetries

For a complex $h$5 of maximal dimension $h$6, the $h$7-vector $h$8 counts $h$9-simplices, and the simplex generating function is

$q$0

with $q$1 by convention (Knill, 20 Aug 2025). In the recursive Dehn-Sommerville setting, one equivalent formulation of the symmetry is

$q$2

so the generating polynomial is even or odd around $q$3 (Knill, 20 Aug 2025). The associated $q$4-vector, defined by

$q$5

is palindromic: $q$6 This is presented as one of several equivalent forms of the Dehn-Sommerville symmetries (Knill, 20 Aug 2025).

A parallel formulation appears in the Gauss–Bonnet approach, where Dehn-Sommerville is expressed as the reflection symmetry

$q$7

This symmetry is proved for an inductively defined class $q$8 satisfying $q$9 and $q$0 for all vertices $q$1 (Knill, 2019). The class includes simplicial spheres, including homology spheres, all odd-dimensional discrete manifolds, and any even-dimensional discrete manifold with $q$2 (Knill, 2019).

For homology manifolds, the Dehn–Sommerville relations are encoded in polynomial identities involving the tilded $q$3- and $q$4-polynomials,

$q$5

together with multiplicities

$q$6

The key relation for a homology manifold is

$q$7

from which the generalized linear Dehn–Sommerville relations follow (Ceballos et al., 2021). This formulation subsumes the Euler–Poincaré relation and isolates the role of interior faces (Ceballos et al., 2021).

3. Local structure, links, and errors from the links

The recursive definition makes the local structure of unit spheres primary. A discrete $q$8-manifold is a simplicial complex where all unit spheres are $q$9-spheres; Dehn-Sommerville manifolds enlarge this class by requiring only that all unit spheres be Dehn-Sommerville $G$0-manifolds (Knill, 20 Aug 2025). This produces a class described as “much larger than traditional manifolds,” and examples include suspensions of non-sphere manifolds (Knill, 20 Aug 2025).

A central warning in this framework is that most Dehn-Sommerville manifolds do not satisfy Poincaré duality (Knill, 20 Aug 2025). This sharply distinguishes the recursive class from closed orientable homology manifolds, where Poincaré duality is typically the source of $G$1-vector symmetry (Ayzenberg, 2015, Limonchenko et al., 16 Jun 2025).

For general pure simplicial complexes, the failure of Dehn–Sommerville symmetry can be quantified explicitly by local link defects. The generalized formula is

$G$2

where

$G$3

This expresses any asymmetry in terms of “errors coming from the links” (Sawaske et al., 2020). The same philosophy appears in the 2021 revisiting paper, where the error term

$G$4

governs the general Dehn–Sommerville relation

$G$5

for arbitrary simplicial complexes (Ceballos et al., 2021).

This local-error perspective suggests a broad conceptual unification: exact Dehn–Sommerville symmetry corresponds to sphere-like or Eulerian behavior of links, while departures from symmetry are entirely localized in the combinatorics of those links (Sawaske et al., 2020, Ceballos et al., 2021).

4. Topological realizations and duality explanations

In several settings, Dehn–Sommerville symmetries are explained by topological duality rather than by direct combinatorial manipulation. For simplicial posets whose geometric realization is a closed orientable homology manifold, a manifold with boundary $G$6 can be constructed so that

$G$7

and

$G$8

Because this is a Poincaré duality algebra, one gets

$G$9

giving a topological basis for the generalized Dehn–Sommerville symmetry of $S(x)$0-numbers (Ayzenberg, 2015).

A related topological explanation appears for Bier spheres. Given a simplicial complex $S(x)$1, a canonical $S(x)$2-dimensional toric manifold $S(x)$3 is associated to a canonical $S(x)$4-dimensional complete regular fan $S(x)$5, whose underlying simplicial complex is isomorphic to the Bier sphere $S(x)$6 (Limonchenko et al., 16 Jun 2025). The Dehn-Sommerville relations for Bier spheres take the form

$S(x)$7

and the proof proceeds by Danilov’s theorem,

$S(x)$8

combined with Poincaré duality on the closed, orientable, smooth manifold $S(x)$9,

$(q-1)$0

Thus, the symmetry of the $(q-1)$1-vector becomes a direct consequence of manifold duality (Limonchenko et al., 16 Jun 2025).

These constructions show that, in manifold-like settings, Dehn–Sommerville relations often encode the same formal dualities as cohomological Poincaré duality. A plausible implication is that the term “Dehn-Sommerville manifold” functions partly as a bridge concept between combinatorial face enumeration and duality phenomena in topology.

5. Algebraic, valuation-theoretic, and refinement properties

The recursive theory associates Dehn-Sommerville symmetry with vanishing of specific linear valuations $(q-1)$2 and eigenvector functionals $(q-1)$3 of the barycentric refinement operator (Knill, 20 Aug 2025). For $(q-1)$4,

$(q-1)$5

The curvature lemma states that the curvature at a vertex for $(q-1)$6 is the Dehn-Sommerville valuation for its unit sphere (Knill, 20 Aug 2025).

Barycentric refinement is also central in the Gauss–Bonnet formulation. If $(q-1)$7 denotes the barycentric refinement operator on $(q-1)$8-vectors, then $(q-1)$9, all eigenvalues are $1-(-1)^q$0, and the Perron–Frobenius eigenvector has Dehn-Sommerville symmetry in the barycentric limit (Knill, 2019). Even eigenvectors of $1-(-1)^q$1 yield Dehn-Sommerville functionals that vanish on the class $1-(-1)^q$2; for $1-(-1)^q$3-manifolds, one such relation is

$1-(-1)^q$4

(Knill, 2019).

The recursive class is also stable under several operations. It is stated to be invariant under edge refinement, barycentric refinement, and Cartesian products (Knill, 20 Aug 2025). The join operation defines a monoid structure on Dehn-Sommerville spheres and on odd-dimensional Dehn-Sommerville manifolds, with the void as identity (Knill, 20 Aug 2025). Connected sums are likewise said to preserve the Dehn-Sommerville property (Knill, 20 Aug 2025).

This algebraic and refinement-theoretic behavior indicates that the class is designed to be closed under standard combinatorial constructions, even when classical manifold axioms would fail. That closure is one reason the recursive notion is substantially broader than the homology-manifold notion.

6. Higher characteristics, chromatic bounds, and odd-dimensional flatness

One of the strongest global properties asserted for recursive Dehn-Sommerville manifolds concerns higher characteristics. For any complex $1-(-1)^q$5, the higher characteristics are

$1-(-1)^q$6

For any Dehn-Sommerville $1-(-1)^q$7-manifold,

$1-(-1)^q$8

(Knill, 20 Aug 2025). In odd dimensions,

$1-(-1)^q$9

and therefore all higher characteristics vanish (Knill, 20 Aug 2025). The paper describes odd-dimensional Dehn-Sommerville manifolds as flat and states that they form a monoid under join (Knill, 20 Aug 2025).

Another structural statement is the chromatic bound

$S(x)=\delta U(x)$0

for a Dehn-Sommerville $S(x)=\delta U(x)$1-manifold, while barycentric refinements have chromatic number $S(x)=\delta U(x)$2 (Knill, 20 Aug 2025). In addition, level sets are stable: if $S(x)=\delta U(x)$3, then the set of simplices $S(x)=\delta U(x)$4 such that $S(x)=\delta U(x)$5 covers all values is a Dehn-Sommerville $S(x)=\delta U(x)$6-manifold if not empty (Knill, 20 Aug 2025).

These statements have no direct analogue in the classical polytope formulation of Dehn–Sommerville relations. They emphasize that the recursive class is intended as a genuine manifold-like category with its own internal calculus of constructions, invariants, and induction principles.

7. Extensions beyond classical manifold theory

Dehn–Sommerville theory now extends well beyond simplicial spheres and ordinary manifolds. One direction is purely combinatorial: in oriented matroids, decompositions of topes with respect to symmetric cycles yield simplicial complexes $S(x)=\delta U(x)$7 whose $S(x)=\delta U(x)$8-vectors satisfy Dehn-Sommerville type relations when $S(x)=\delta U(x)$9 (Matveev, 2017). In that setting, the relevant relations include

$U(x)=\{y\in G\mid x\subset y\}$0

and further recursive identities for $U(x)=\{y\in G\mid x\subset y\}$1 (Matveev, 2017). This does not define a Dehn-Sommerville manifold in the simplicial-manifold sense, but it shows that Dehn-Sommerville-type symmetry can arise from purely matroidal decomposition data.

Another direction is enumerative and toric. Weighted lattice-point sums in lattice polytopes are encoded by a generating function

$U(x)=\{y\in G\mid x\subset y\}$2

which satisfies the functional equation

$U(x)=\{y\in G\mid x\subset y\}$3

For $U(x)=\{y\in G\mid x\subset y\}$4 and $U(x)=\{y\in G\mid x\subset y\}$5, this recovers the classical Dehn–Sommerville relations; for $U(x)=\{y\in G\mid x\subset y\}$6, it simultaneously encodes Ehrhart–Macdonald reciprocity (Beck et al., 2018). The same framework extends to non-simple polytopes through $U(x)=\{y\in G\mid x\subset y\}$7-polynomial weights and connects the Dehn–Sommerville phenomenon with intersection cohomology and toric varieties (Beck et al., 2018).

The poset-theoretic direction develops linear relations between the numbers $U(x)=\{y\in G\mid x\subset y\}$8 of Boolean intervals in a distributive lattice $U(x)=\{y\in G\mid x\subset y\}$9 and clique counts $h$00 in an associated graph $h$01, rederiving the Dehn–Sommerville equations for simplicial polytopes through Stanley–Reisner theory and free resolutions (Hegedüs, 2010). This again shows that Dehn–Sommerville symmetry can be realized in algebraic-combinatorial categories that are not manifolds in any ordinary topological sense.

Taken together, these developments suggest that “Dehn-Sommerville manifold” sits inside a broader landscape of Dehn-Sommerville objects: homology manifolds, semi-Eulerian and reciprocal complexes, balanced and completely balanced complexes, toric constructions from Bier spheres, recursive simplicial complexes defined by unit spheres, and oriented-matroidal simplicial complexes (Ceballos et al., 2021, Sawaske et al., 2020, Limonchenko et al., 16 Jun 2025, Matveev, 2017). The unifying principle is not a single manifold axiom, but the persistence of face-enumerative dualities—often expressible as palindromicity, reciprocity, or vanishing curvature-like valuations—under conditions that range from Poincaré duality to purely local link constraints.