The Non-Pure Dual Exchange Property in Low Dimensions

Abstract: We investigate monomial ideals satisfying the non-pure dual exchange property, a notion introduced in connection with componentwise polymatroidal ideals. Our contributions are twofold. First, we show that in two variables, every integrally closed monomial ideal satisfies this property; as a consequence, we characterize polymatroidal ideals in two variables. Second, for strongly stable (Borel) ideals in three variables, we establish a practical criterion: it suffices to verify the defining condition only for the Borel generators, and this verification reduces to simple inequalities involving the degrees in the second and third variables.

• The paper demonstrates that integrally closed monomial ideals in two variables always satisfy the non-pure dual exchange property.
• It reduces NDXP verification in three-variable Borel ideals to checking simple degree conditions among minimal generators.
• The results imply that ideals satisfying NDXP possess linear quotients, linking combinatorial structure to homological properties.

The Non-Pure Dual Exchange Property for Monomial Ideals in Low Dimensions

Introduction

This paper addresses the structural properties of monomial ideals in polynomial rings, focusing on the non-pure dual exchange property (NDXP) introduced in the context of componentwise polymatroidal ideals. The authors concentrate on explicit characterizations in two variables and develop practical verification methods for Borel (strongly stable) ideals in three variables. The results illuminate the interplay between integral closure, polymatroidality, and exchange-type conditions that govern the algebraic and combinatorial structure of monomial ideals.

Non-Pure Dual Exchange Property: Formalism and Context

The NDXP is a condition on the minimal generating set $G(I)$ of a monomial ideal $I \subseteq K[x_1, \dots, x_n]$. Given $u, v \in G(I)$ with $\deg(u) \le \deg(v)$, and for each index $i$ with $\deg_{x_i}(v) < \deg_{x_i}(u)$, it requires the existence of an index $j$ such that $\deg_{x_j}(v) > \deg_{x_j}(u)$ and the monomial $x_i (v / x_j) \in I$.

NDXP generalizes the classical exchange property used to characterize polymatroidal ideals. In single-degree (equigenerated) cases, it coincides with the standard exchange property. Prior work established that componentwise polymatroidal ideals—those for which each homogeneous degree is polymatroidal—satisfy NDXP. Additionally, possessing NDXP ensures the ideal has linear quotients, thus being componentwise linear.

Two Variables: Integral Closure, Exchange, and Polymatroidality

For $K[x, y]$, the paper provides a thorough characterization of ideals with the NDXP:

• Every integrally closed monomial ideal in two variables satisfies NDXP; the converse fails, as integrally closedness is strictly stronger ([Remark 2.12]).
• Polymatroidal ideals among equigenerated (single-degree) monomial ideals correspond precisely to integrally closed ideals in $I \subseteq K[x_1, \dots, x_n]$0 ([Theorem 2.14]). This yields an explicit description: polymatroidal ideals are precisely those generated by the monomials $I \subseteq K[x_1, \dots, x_n]$1 for $I \subseteq K[x_1, \dots, x_n]$2, for some $I \subseteq K[x_1, \dots, x_n]$3.

Integral closure is described through the convex hull of the exponent vectors of generators. The paper uses combinatorial properties of the sequence of differences between generator exponents to formulate necessary and sufficient conditions for NDXP, resulting in a testable algebraic/combinatorial criterion ([Proposition 2.2, Corollary 2.3]). The implications extend further:

• Colon ideals and saturations of integrally closed monomial ideals in two variables also satisfy NDXP.
• The product of two integrally closed ideals in $I \subseteq K[x_1, \dots, x_n]$4 remains integrally closed, hence, per the above characterization, polymatroidal ([Remark 2.16]).

These results do not extend naively to higher-variable cases; explicit counterexamples are provided ([Example 2.11, 2.17]).

Three Variables: Borel Ideals and Reduction to Generators

In $I \subseteq K[x_1, \dots, x_n]$5, the equivalence between integral closure and NDXP does not hold. Instead, attention is given to Borel (strongly stable) ideals, which admit canonical minimal generating sets (Borel generators).

The central result ([Theorem 3.7]) establishes that for Borel ideals in three variables, checking NDXP can be reduced to verifying simple degree conditions among Borel generators:

• For $I \subseteq K[x_1, \dots, x_n]$6 in the set of Borel generators $I \subseteq K[x_1, \dots, x_n]$7 with $I \subseteq K[x_1, \dots, x_n]$8, the conditions are:
  • $I \subseteq K[x_1, \dots, x_n]$9
  • If $u, v \in G(I)$0, then $u, v \in G(I)$1

This reduction leads to a practical, finite test for NDXP in Borel ideals, a significant computational advantage. Principal Borel ideals satisfy NDXP automatically; however, not all Borel ideals in three variables do. Several explicit examples illustrate both possibilities ([Examples 3.8, 3.9]).

It is also proven that this reduction is special to three variables; in higher dimensions, verifying conditions only among Borel generators is insufficient ([Remark 3.8]). This demarcates a key structural difference in the behavior of Borel ideals by dimension.

Implications and Outlook

Strong algebraic consequences emerge for ideals with NDXP: linear quotients and componentwise linearity, which bear on free resolutions and Betti numbers. The explicit characterizations for $u, v \in G(I)$2 connect combinatorial and homological properties. The reduction in the three-variable Borel case enables efficient symbolic or computer-assisted verification.

The results expose sharp dichotomies: the equivalence between integral closure, polymatroidality, and NDXP in two variables collapses in higher dimensions. Furthermore, the intricate structure of Borel ideals in three variables allows precise combinatorial control, a feature not preserved in four or more variables.

Future research may extend these combinatorial tests to broader classes of monomial ideals or seek generalizations for additional algebraic or homological properties. The techniques developed may also impact computational approaches in commutative algebra, particularly for syzygy computations or the systematic study of resolutions.

Conclusion

This paper establishes critical connections between the non-pure dual exchange property, integral closure, and polymatroidality in monomial ideals of two and three variables. It yields testable structural criteria and clarifies the algebraic landscape governing such ideals. The results define both the opportunities and limitations of using exchange-type conditions in low dimensions, offering a foundation for further exploration into the combinatorics of monomial ideals and their algebraic invariants.