---
title: Saturated Newton Polytope (SNP)
url: https://www.emergentmind.com/topics/saturated-newton-polytope-snp
type: topic
---

# Saturated Newton Polytope (SNP)

A saturated Newton polytope (SNP) is the Newton polytope of a polynomial whose lattice points are exactly the exponent vectors of monomials with nonzero coefficient. If
\[
f(z_1,\dots,z_d)=\sum_{\alpha\in\mathbb Z_{\ge 0}^d} c_\alpha z^\alpha,
\]
with support
\[
\operatorname{Supp}(f)=\{\alpha\in\mathbb Z_{\ge 0}^d:c_\alpha\neq 0\},
\]
then its Newton polytope is
\[
\operatorname{Newton}(f)=\operatorname{Conv}(\operatorname{Supp}(f)),
\]
and \(f\) has SNP precisely when
\[
\operatorname{Newton}(f)\cap \mathbb Z^d=\operatorname{Supp}(f).
\]
Equivalently, the Newton polytope is the integer hull of the support: there are no “missing” lattice points inside the convex hull of the exponent set. The condition has been systematically studied in algebraic combinatorics and now appears across symmetric, nonsymmetric, Schubert-theoretic, and cluster-algebraic settings [1703.02583] [2508.00336].

## 1. Definition and geometric meaning

The basic objects are the support and the Newton polytope. For an ordinary polynomial, the support is the finite set of exponent vectors of nonzero monomials, and the Newton polytope is their convex hull. In the Laurent setting used for cluster variables, the same convex-hull construction is applied to exponent vectors in \(\mathbb Z^d\) rather than \(\mathbb Z_{\ge0}^d\) [2507.22528] [2012.07500].

SNP is a saturation condition on this convex geometry. It says that convexification introduces no new integer points beyond the original support. In geometric language, the monomial exponents fill the lattice points of the Newton polytope exactly. The literature repeatedly emphasizes this “no holes” interpretation; in particular, the supersymmetric Schur paper formulates SNP as the statement that the Newton polytope contains no missing integer points and identifies it with the integer hull of the support [2507.22528].

Some emblematic examples already show the strength of the property. The determinant has SNP: its Newton polytope is the Birkhoff polytope, and its only lattice points are the permutation vertices corresponding to the determinant monomials. Squarefree polynomials are also SNP. At the same time, the nonsymmetric Macdonald paper remarks that in higher degree SNP becomes quite restrictive [2508.00336].

## 2. Schur theory, permutahedra, and the classical model

The prototype for SNP is the Schur polynomial. For ordinary Schur polynomials \(s_\lambda\), Rado’s permutahedron theorem and dominance order give SNP: \(\operatorname{Newton}(s_\lambda)\) is the permutahedron \(\Pi(\lambda)\), and \(\Pi(\lambda)\cap\mathbb Z^k=\operatorname{Supp}(s_\lambda)\) [2507.22528]. In the later dual \(k\)-Schur analysis, this same fact is reformulated as: each Schur polynomial is M-convex, or equivalently, it has a saturated Newton polytope and this polytope is a generalized permutahedron [2401.14632].

This Schur-theoretic picture governs a large portion of the subject. Products of Schur polynomials are SNP, and skew Schur polynomials are SNP as a consequence of the SNP property for Stanley symmetric polynomials [1703.02583]. The paper on symmetric Grothendieck polynomials shows that each homogeneous component of a symmetric Grothendieck polynomial has Newton polytope equal to a permutahedron \(\mathcal P_{\mu^{(k)}}\), while the full inhomogeneous polynomial has Newton polytope given by a convex union of such permutahedra and is still SNP [1705.07876].

The Schur model also supplies the dominant comparison principle used repeatedly elsewhere: dominance order controls inclusion of Schur Newton polytopes, and this inclusion is converted into SNP statements for broader families. In the “good symmetric polynomials” framework, the Newton polytope of a polynomial built from a chain of Schur shapes is the convex hull of finitely many Schur permutahedra, and the Schur SNP theorem becomes the input for proving SNP and the integer decomposition property for the larger polynomial [2205.03903].

## 3. Polyhedral and discrete-convex mechanisms

Several distinct structural mechanisms now underlie SNP proofs. One route is discrete convexity. The non-symmetric Macdonald theorem proves that supports of non-symmetric Macdonald polynomials are M-convex; by definition of M-convexity, this already forces saturation, so SNP becomes a corollary. The same paper identifies the resulting Newton polytopes as generalized permutahedra [2508.00336]. The dual \(k\)-Schur paper uses the same equivalence in the homogeneous setting: M-convex support is equivalent to SNP together with generalized-permutahedron Newton polytope [2401.14632].

A second route is polymatroidality. The double Schubert paper proves that the support of every double Schubert polynomial is a discrete polymatroid, obtained from multidegrees of Cohen–Macaulay prime ideals after a standardization procedure for non-standard multigradings. Since discrete polymatroids are exactly the lattice points of polymatroid base polytopes, SNP follows immediately [2109.10299].

A third route is explicit polyhedral integrality. The supersymmetric Schur theorem encodes the support by hook inequalities, constructs a polyhedron \(H\), proves that the defining constraint matrix is totally unimodular, and invokes the Hoffman–Kruskal criterion to show integrality of \(H\). This yields
\[
H=\operatorname{Conv}(H\cap\mathbb Z^{k+\ell})=\operatorname{Newton}(S_\lambda),
\]
so every lattice point of the Newton polytope is realized by a tableau content vector. The paper states that this is, to the authors’ knowledge, the first explicit use of total unimodularity and the Hoffman–Kruskal criterion in an SNP proof [2507.22528].

A fourth route is polyhedral feasibility via representation-theoretic inequalities. For special Kronecker products \(s_\lambda * s_\mu\), the support can be characterized by a polytope \(\mathcal P(\lambda,\mu;\mathbf a)\) built from Horn inequalities for positivity of Littlewood–Richardson coefficients. In the proved cases, nonemptiness of this polytope implies the existence of an integer point, which is enough to deduce SNP [2311.10276].

## 4. Established families and major theorems

The current literature contains a substantial collection of SNP theorems, together with some partial results.

| Family | Structural description | SNP outcome |
|---|---|---|
| Schur polynomials | Permutahedra via dominance order | SNP [2507.22528] |
| Symmetric Grothendieck polynomials | Homogeneous components have permutahedral Newton polytopes | SNP for components and whole polynomial [1705.07876] |
| Double Schubert polynomials | Support is a discrete polymatroid | SNP [2109.10299] |
| Non-symmetric Macdonald polynomials | Support is M-convex | SNP [2508.00336] |
| Supersymmetric Schur polynomials | Hook-inequality polyhedron with totally unimodular matrix | SNP [2507.22528] |
| Dual \(k\)-Schur polynomials | Same support as \(s_\lambda\) | Same saturated Newton polytope as Schur [2401.14632] |
| Affine Stanley and cylindric skew Schur polynomials | Derived from dual \(k\)-Schur positivity | M-convex, hence SNP [2401.14632] |
| Cluster variables in types \(A\) and \(D\) | Snake-graph matching polytopes in stated coefficient regimes | Saturated in the proved regimes [2012.07500] |
| Kronecker products \(s_\lambda * s_\mu\) | Horn-inequality polyhedra in special cases | SNP in several special cases [2311.10276] |

Beyond these headline results, the class of “good symmetric polynomials” gives a unified SNP theorem covering symmetric Grothendieck polynomials, inflated symmetric Grothendieck polynomials, Stembridge’s symmetric polynomials associated with totally nonnegative matrices, cycle index polynomials, Reutenauer’s symmetric polynomials, Schur \(P\)- and \(Q\)-polynomials, Stanley’s symmetric polynomials, chromatic symmetric polynomials for several special graph classes, and dual Grothendieck polynomials [2205.03903].

For cluster algebras, the type \(A\) and \(D\) saturation theorem covers all cluster variable Newton polytopes with boundary frozen variables or principal coefficients, and also the no-frozen case in type \(A\). In type \(A\) with boundary frozen variables or principal coefficients, the Newton polytopes are even empty, meaning every lattice point is a vertex [2012.07500].

This range of examples suggests that SNP is compatible with several different combinatorial models—tableaux, fillings, matchings, Bruhat ideals, multidegrees, and discrete convex sets—rather than belonging to a single polyhedral template.

## 5. Interactions with discrete geometry, representation theory, and computation

SNP is repeatedly linked to stronger integrality properties. The “good symmetric polynomials” theorem proves not only SNP but also the integer decomposition property (IDP) for the associated Newton polytopes. The paper isolates a combinatorial condition on Schur expansions that forces both properties simultaneously [2205.03903]. A related three-dimensional result proves IDP for 2-partition maximal symmetric polytopes in a hyperplane of \(\mathbb R^3\) by showing that certain sums of Schur polynomials have SNP for all dilations of the relevant combinatorial parameter [2501.04191].

Representation-theoretic geometry also enters directly. For non-symmetric Macdonald polynomials, the Newton polytope is the convex hull of a lower Bruhat ideal, and the same polytope is identified with the moment polytope of an affine Schubert variety in the affine Grassmannian of type \(\mathrm{GL}\). The M-convexity/SNP theorem therefore implies that all such moment polytopes are generalized permutahedra [2508.00336].

In Schubert theory, SNP has algorithmic consequences. The computational-complexity paper explains that for a combinatorially positive SNP family, nonvanishing can be reduced to membership in the Newton polytope, placing the nonvanishing problem in \(NP\cap coNP\) once a halfspace description is available. For Schubert polynomials, the Schubitope description together with total unimodularity yields a polynomial-time algorithm for nonvanishing [1810.10361].

The subject also now interacts closely with Lorentzian and M-convex theories. The non-symmetric Macdonald and dual \(k\)-Schur papers explicitly interpret SNP through M-convex support and generalized permutahedra, while the supersymmetric Schur paper proposes Lorentzianity as a further structural direction [2508.00336] [2401.14632] [2507.22528].

## 6. Limitations, counterexamples, and open directions

SNP is strong and fragile. The foundational survey records that it is not preserved under powers, that Schur-positivity does not imply SNP, that \(e\)-positivity does not imply SNP, that the involution \(\omega\) does not preserve SNP, and that \(p_\lambda\) is almost never SNP except for \(\lambda=(1^k)\) [1703.02583]. These examples rule out the common misconception that positivity in a standard symmetric-function basis is close to saturation.

Several important general problems remain unresolved. For Kronecker products \(s_\lambda * s_\mu\), SNP is proved only in special cases. The 2023 paper explicitly notes skepticism that SNP holds for all Kronecker products, while also emphasizing that no counterexample is known [2311.10276]. For cluster algebras from more general surfaces, saturation can fail: explicit counterexamples are given for boundary-coefficient cluster variables outside types \(A\) and \(D\), leading to the conjecture that all cluster variable Newton polytopes are saturated for all seeds and arcs only in the polygon and once-punctured polygon cases [2012.07500].

Open problems also concern the relation between SNP and stronger convexity properties. One paper states that it is not known whether there exists a symmetric polynomial that has SNP but whose Newton polytope does not have IDP [2205.03903]. The supersymmetric Schur work conjectures that supersymmetric Schur polynomials are Lorentzian and announces further work on the integer decomposition property and matroid-theoretic aspects of the associated polytopes [2507.22528]. The dual \(k\)-Schur paper conjectures M-convexity for \(k\)-Schur polynomials and Lorentzianity for normalized \(k\)-Schur and dual \(k\)-Schur polynomials [2401.14632].

Historically, the subject has moved from a survey-and-conjecture stage to a theorem-driven one. The survey of Monical–Tokcan–Yong collected many SNP instances and conjectures [1703.02583]; later work settled major cases such as double Schubert polynomials [2109.10299], non-symmetric Macdonald polynomials [2508.00336], and supersymmetric Schur polynomials [2507.22528]. The remaining landscape suggests that SNP is neither automatic nor exceptional: it is a precise integrality phenomenon whose proof typically requires a detailed combinatorial or polyhedral model of support.

Source: https://www.emergentmind.com/topics/saturated-newton-polytope-snp