🔢 mathematics

Integral closure for (additively idempotent) semirings

This paper explores the divergence of integrality definitions in additively idempotent semirings by establishing a Cayley-Hamilton theorem and computing integral closures to facilitate the normalization of tropical varieties.

Original authors: Netanel Friedenberg, Kalina Mincheva

Published 2026-07-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Netanel Friedenberg, Kalina Mincheva

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are a mathematician trying to organize a chaotic city of numbers. In the old, classic city (called "ring theory"), everyone agrees on the rules for what it means for a building to be "complete" or "integral." If a building satisfies one rule, it satisfies them all. It's like saying if a house has a solid foundation, it automatically has a roof and walls.

But then, the authors of this paper, Netanel Friedenberg and Kalina Mincheva, move to a different, stranger neighborhood called Tropical Geometry. Here, the rules of addition are weird: adding a number to itself doesn't make it bigger; it stays the same. This is called being "additively idempotent." In this neighborhood, the old rules break down. A building might have a solid foundation but still be missing a roof, even though in the old city, that would be impossible.

The paper's main job is to figure out what "completeness" (or integral closure) actually means in this strange new world. They discover that there isn't just one definition of a "complete" building anymore. Instead, there are several different ways to check if a building is finished, and they don't always agree.

The Four Different Checklists

The authors introduce four different "checklists" to see if a number (or a building) is integral over a smaller set of numbers:

The "Monic Polynomial" Checklist (J-integral): This is like checking if a building can be described by a specific, strict blueprint where the main beam is exactly one unit long.
The "Module" Checklist (Quasi-integral): This checks if the building can hold a specific, finite amount of furniture (a "faithful module") without collapsing.
The "Downward" Checklist (D-integral): This is a safety net. It says, "If there's a finished building above this one, and this one is smaller or equal to it, then this one counts as finished too."
The "Valuation" Checklist (Valuative): This is the ultimate judge. It checks if the building fits inside every possible "valuation semiring" (think of these as different types of zoning laws or inspectors) that the original set of numbers obeys.

The big surprise? In the old city, all these checklists gave the same result. In the tropical city, they can disagree, but the authors prove that in many specific, well-behaved neighborhoods, they actually agree.

The Magic Tool: The Tropical Cayley-Hamilton Theorem

To make sense of this mess, the authors invent a powerful tool called the Tropical Cayley-Hamilton Theorem. Imagine you have a grid of numbers (a matrix). In the old world, there's a famous rule that says if you plug this grid into its own special equation, it vanishes. In the tropical world, the grid doesn't vanish; instead, it satisfies a "bend relation."

Think of it like a flexible ruler. If you bend it just right, the two ends touch. The authors prove that any grid of numbers in this tropical world will always "bend" in a way that satisfies its own characteristic equation. This tool helps them prove that in many specific, well-behaved neighborhoods (like those without "zero divisors," which are like numbers that can multiply to nothing, or those that are "cancellatively generated"), all the different checklists do actually agree.

When Do the Rules Agree?

The paper finds that if the neighborhood is "cancellative" (meaning you can cancel out common factors without breaking things) and has no zero divisors, then all four definitions of integral closure become the same thing. It's like finding a district where the strict blueprint, the furniture test, the safety net, and the zoning inspector all say, "Yes, this building is complete."

However, the authors are careful to point out that this doesn't happen everywhere. In some messy, non-cancellative areas (which are very common in tropical geometry), the definitions stay different. They explicitly show an example where a building passes the "Monic Polynomial" test, but its square fails the test. This proves that the "Monic Polynomial" list is not a perfect "closure operation"—meaning if you take all the "finished" buildings and add them together, you might accidentally create a new building that isn't finished according to the same rules.

The Real-World Connection: Normalizing Curves

Why does this matter? The authors connect this abstract math to tropical varieties, which are combinatorial shapes that represent complex algebraic curves. One of the big goals in geometry is "normalization," which is like smoothing out a crumpled piece of paper or fixing a knot in a curve.

The authors suggest that by computing the "valuative integral closure" (the ultimate zoning inspector's list) of a tropical curve's coordinate semiring, we can figure out how to normalize the curve. They compute this for a few specific curves, like a "cuspidal cubic" (a curve with a sharp point). They find that the "finished" version of the tropical curve looks exactly like the tropical version of the "finished" classical curve.

What They Don't Know (Yet)

The paper is very honest about what is still a mystery. They propose a conjecture (a strong guess, not a proven fact) that for a specific type of curve with one singular point, the tropical normalization perfectly matches the classical normalization. They also introduce a method to find "witness pairs"—pairs of polynomials that prove a number is not cancellative (it can't be divided cleanly). They have an algorithm to find these pairs, but they admit that for some complex curves, it's still hard to tell exactly which numbers are cancellative.

In short, the paper maps out the landscape of "completeness" in tropical geometry. It proves that while the old rules don't apply everywhere, there are new, reliable ways to check if a tropical structure is whole, and in many important cases, all the different ways of checking actually lead to the same result. It doesn't solve every puzzle, but it gives us the right tools to start building the solutions.

Technical Summary: Integral Closure for (Additively Idempotent) Semirings

Problem Statement
In classical commutative ring theory, various definitions of integral elements are equivalent: an element is integral if it is a root of a monic polynomial, if it generates a finitely generated faithful module, or if it lies in the intersection of valuation rings containing the base ring. However, when working with additively idempotent semirings (where $a + a = a$ ), these notions diverge. The paper addresses the lack of a unified theory for integral closure in this setting, which is crucial for tropical geometry. Specifically, the authors aim to define, compare, and compute different notions of integrality for semirings, with a focus on the "coordinate semiring" of a tropical variety and its total semiring of fractions. This work is motivated by the need to define and compute the normalization of tropical varieties and to extend the theory of tropical adic spaces.

Methodology and Definitions
The authors establish a framework by defining four primary notions of integrality for an extension of additively idempotent semirings $A \subseteq R$ and an element $x \in R$ :

N-integral: $x$ satisfies an equation $x^n = \sum_{i=0}^{n-1} a_i x^i$ with $a_i \in A$ .
J-integral: $x$ satisfies an inequality $x^n \le \sum_{i=0}^{n-1} a_i x^i$ with $a_i \in A$ .
D-integral: $x$ is bounded above by a J-integral element (the downward closure of J-integral elements).
Quasi-integral (and variants): $x$ admits a faithful (or annihilator-free) $A\langle x \rangle$ -module $M \subseteq R$ that is finitely generated as a downward-closed $A$ -module.
Valuative integral closure ( $A^V$ ): The intersection of all valuation sub-semirings $R_P$ containing $A$ , where $P$ ranges over prime congruences.

The authors denote the sets of such elements as $A^N, A^J, A^D, A^Q, A^{aQ}, A^{wQ}$ , and $A^V$ . A key structural tool developed is the Tropical Cayley-Hamilton Theorem (Theorem 5.2), which states that any square matrix over an additively idempotent semiring satisfies the "bend relations" of its characteristic polynomial. This theorem is proven using a combinatorial adaptation of Straubing's proof of the classical Cayley-Hamilton theorem, utilizing path decompositions and involutions on partial permutations.

Key Contributions and Results

Relations between Integrality Notions: The paper establishes a hierarchy of inclusions: $A^N \subseteq A^J \subseteq A^D \subseteq A^Q \subseteq A^{aQ} \subseteq A^{wQ}$ . Crucially, the authors prove that $A^J$ is not necessarily a semiring (Example 3.14) and that the operation $A \mapsto A^J$ is not a closure operation. In contrast, they prove that $A^D$ is always a semiring and that the operation $A \mapsto A^D$ is a closure operation (Proposition 4.5, Proposition 4.8).
Equivalence Theorems: The authors identify conditions under which these distinct notions coincide:
- If $R$ is cancellative, then all notions agree and equal the downward closure of $A$ in $R$ (Theorem 6.20).
- If $R$ is cancellatively generated, $\sqrt{\Delta_R}$ is cancellative, and the downward-closed $A$ -module generated by any $x \in R$ is finite, then $A^J = A^D = A^Q = A^{aQ} = A^{wQ} = A^V$ (Theorem 6.21).
- For tropicalizations of ideals $I \subseteq K[M]$ where the variety has no component in the toric boundary, the valuative integral closure $A^V$ coincides with the quasi-integral closure $A^Q$ (Corollary 7.4).
Cancellative Elements and Fractions: A significant portion of the work is dedicated to identifying cancellative elements in semirings of the form $S[M]/\text{Bend}(J)$ . The authors prove that in $S[M]$ , the only cancellative elements are monomials (Proposition 8.4). They investigate when elements in the quotient by bend congruences remain cancellative, leading to Conjecture 8.10, which posits that for a principal prime ideal $I$ not containing monomials or binomials, an element in $S[M]/\text{Bend}(\text{trop } I)$ is cancellative if and only if it is the image of a term.
Computational Tools: The paper provides an algorithm (Construction 9.7 and Theorem 9.12) to construct "witness pairs" $(g, h)$ such that $g \neq h$ but $fg = fh$, thereby certifying that $f$ is not cancellative. This allows for the computational determination of the total semiring of fractions $\text{Frac}(A)$ .

Applications and Examples
The authors apply their theoretical results to compute integral closures in specific geometric contexts:

Tropical Curves: They compute the integral closure of coordinate semirings for tropical curves, such as the cuspidal cubic ( $x^2 + x^3 + y^2$ ) and the nodal cubic. In these cases, the integral closure corresponds to the coordinate ring of the normalization of the classical curve (e.g., adjoining $y/x$ ).
Toric Varieties: They show that for toric semirings $S[M]$ , the integral closure in their total semiring of fractions is the semiring itself (i.e., they are normal).
Normalization: The paper demonstrates that computing the normalization of a tropical variety can be approached algebraically by computing the integral closure of its coordinate semiring in its total semiring of fractions.

Significance and Claims
The paper claims to provide the first systematic comparison of multiple definitions of integrality in the context of additively idempotent semirings. By proving the Tropical Cayley-Hamilton theorem, the authors establish a bridge between matrix theory and integrality, allowing them to show that under broad geometric conditions (specifically involving cancellative generation and properties of the radical of the diagonal), the various definitions of integral closure collapse into a single, well-behaved notion ( $A^D$ ).

The authors modestly frame their work as a foundational step toward defining tropical adic spaces and understanding the normalization of tropical varieties. They explicitly state that their conjectures (such as Conjecture 8.10 and Conjecture 1.6 regarding the lifting of integral closures from tropical to classical varieties) remain open and require further investigation, particularly regarding the identification of cancellative elements in specific quotients. The work does not claim to solve all problems in tropical normalization but provides the necessary algebraic machinery and specific computational examples to advance the field.

The Four Different Checklists

The Magic Tool: The Tropical Cayley-Hamilton Theorem

When Do the Rules Agree?

The Real-World Connection: Normalizing Curves

What They Don't Know (Yet)

More like this