betaFeedback

WorksMathematics

The Dimension-Degree Boundary for Gaussian Moment Vanishing

Abstract

We classify complex polynomials of degree at most three in three independent standard real Gaussian variables whose positive moments all vanish. Up to a complex orthogonal linear change of coordinates and a nonzero scalar, every such polynomial has one-sided circular support or belongs to a single exceptional orbit. For every fixed polynomial QQ, its mixed moments E[QPm]\mathbb E[QP^m] vanish whenever m>3deg⁡Qm>3\deg Q, and the coefficient three is optimal. The proof extracts a Gaussian coordinate in which the polynomial is affine linear, then classifies that family using formal stationary frames and analytic period estimates. Combining this ternary-cubic result with Wilson's two-variable theorem, Liu's quadratic theorem and Long's counterexamples gives the exact dimension-degree boundary for the Gaussian Moments Conjecture: the implication from pure-moment vanishing to eventual mixed-moment vanishing holds universally precisely when there are at most two variables, the degree is at most two, or there are three variables and the degree is at most three. We reproduce the all-order moment identities for the two minimal failures.

What vanishing moments tell us

Let X1,…,XnX_1,\ldots,X_n be independent standard real Gaussian random variables. We allow a polynomial P(X)P(X) to have complex coefficients. Suppose that

E[P(X)m]=0for every integer m≥1.(1)\mathbb{E}[P(X)^m] = 0 \qquad\text{for every integer } m \ge1. \tag*{(1)}

The Gaussian Moments Conjecture asks whether this forces

E[Q(X)P(X)m]=0for all sufficiently large m(2)\mathbb{E}[Q(X)P(X)^m] = 0 \qquad\text{for all sufficiently large } m \tag*{(2)}

for every fixed polynomial QQ. The threshold may depend on both PP and QQ.

The complex coefficients matter. For real PP, the second moment in (1) forces P=0P=0. For complex PP, cancellation can make every positive moment vanish even when PP is far from zero. For example, (X1+iX2)/2(X_1+iX_2)/\sqrt{2} has this property. The question is whether a fixed polynomial multiplier can keep exposing a nonzero moment at arbitrarily large powers.

We call (1) purity and (2), for every QQ, safety. These names distinguish the hypothesis from the conclusion; safety without purity is a separate question. In the original formulation of Derksen, van den Essen and Zhao [1], the assertion that every pure polynomial is safe was a conjecture in every dimension. Long’s examples show why the dimension cannot be ignored [8].

Here the question is more precise. For integers n≥1n\ge1 and d≥0d\ge0, let G(n,d)G(n,d) mean that every pure complex polynomial of total degree at most dd in nn standard Gaussian variables is safe.

Theorem 1.1 (Dimension–degree boundary).

G(n,d)⟺n≤2 or d≤2 or (n,d)=(3,3).G(n,d) \Longleftrightarrow n \le2 \text{ or } d \le2 \text{ or } (n,d)=(3,3).

The entries below describe a universal implication for all polynomials in each dimension–degree range. Individual safe polynomials also exist outside those ranges.

d≤2d \le2d=3d=3d≥4d \ge4
n=1,2n=1,2purity implies safetypurity implies safetypurity implies safety
n=3n=3purity implies safetypurity implies safetyfailures exist
n≥4n \ge4purity implies safetyfailures existfailures exist

Table 3.

The only positive case needed beyond the existing binary and quadratic theorems is a cubic in three variables. Its quantitative form is useful:

Theorem 1.2 (Ternary cubics). Let P∈C[X1,X2,X3]P\in\mathbb{C}[X_1,X_2,X_3] have total degree at most three and satisfy (1). Then, for every nonzero polynomial QQ,

E[QPm]=0whenever m>3deg⁡Q.(3)\mathbb{E}[QP^m]=0 \qquad\text{whenever } m>3\deg Q. \tag*{(3)}

For Q=0Q=0 the conclusion is immediate. No smaller constant can replace three uniformly over this class.

The proof gives a classification as well as a vanishing bound. Put

Z=X1+iX22,W=X1−iX22,T=X3.(4)Z=\frac{X_1+iX_2}{\sqrt{2}}, \qquad W=\frac{X_1-iX_2}{\sqrt{2}}, \qquad T=X_3. \tag*{(4)}

Call a phase one-sided if all its monomials have circular weight deg⁡Z−deg⁡W\deg_Z-\deg_W of one strict sign; TT has weight zero.

Theorem 1.3 (Normal forms for pure ternary cubics). A complex polynomial of degree at most three in three Gaussian variables is pure if and only if, after a complex orthogonal linear change of coordinates and multiplication by a nonzero scalar, it is one-sided or is the exceptional phase

P∗=W+ZT−12Z3.P_{\ast}=W+ZT-\frac{1}{2}Z^{3}.

The zero polynomial is included in the one-sided case: its empty support satisfies either strict-sign condition. Thus the classification consists of one-sided support and one exceptional orbit. One-sided phases have cutoff m>deg⁡Qm>\deg Q; the exceptional orbit accounts for the sharp coefficient three.

Notation and reading guide

The coordinate and parameter conventions used throughout are:

SymbolMeaning
n,dn,dNumber of real Gaussian variables and upper bound on deg⁡P\deg P
Z,W,TZ,W,TCircular pair with E[ZaWb]=δaba!\mathbb{E}[Z^{a}W^{b}]=\delta_{ab}a!, and an independent standard Gaussian
t,rt,rFormal moment parameter; r=1/tr=1/t near infinity, with a finite root cover when stated
ΦQ\Phi_QMarked series ∑m≥0E[QPm]tm/m!\sum_{m\geq0}\mathbb{E}[QP^{m}]t^{m}/m!
q(x)q(x)Gaussian exponent after a coordinate change, including its linear and constant terms

Table 1.

All polynomial degrees are total degrees. An expectation is the algebraic Gaussian moment functional, extended to formal series coefficient by coefficient. Complex coordinate changes preserve this functional when they preserve its bilinear covariance form. Analytic periods enter only after the comparison hypotheses are checked.

For the boundary statement and its examples, read § 1–§ 7 and § 12. The full ternary-cubic proof occupies § 8–§ 11; the proof map below locates its two main reductions.

Proof map

Write P=F3+F2+F1+F0P=F_{3}+F_{2}+F_{1}+F_{0} in homogeneous pieces. The logical route is:

StepStatement usedProof location
Find a Gaussian directionPurity gives a nonzero vv with vTv=0v^{T}v=0, ∇F3(v)=0\nabla F_{3}(v)=0, F2(v)=0F_{2}(v)=0§ 11
Make one coordinate affine linearThe frame test gives P=Wh(Z,T)+G(Z,T)P=Wh(Z,T)+G(Z,T)Lemma 7.1
Classify the affine familyOne-sided support or the exceptional orbit§ 9
Transfer the mixed boundWeight cancellation or the exact shear source identity§ 6

Table 2.

Formal tools in § 8 and analytic comparison in § 10 support the two classification steps. § 12 assembles the conclusions.

The leading cubic has nine geometric types. The abbreviated table below groups types with the same exclusion argument; the full table appears in § 11. In every singular row, the task is to prove that purity supplies a direction passing the frame test. § 6 proves sharpness with a three-term example. The difficult step is reducing an arbitrary pure ternary cubic to a family with a direct moment calculation; § 7 states the two reductions.

Leading cubicArgument when the frame test fails
SmoothGradient criterion, Corollary 11.3
Nodal; transverse line–conic; nonconcurrent linesOrdinary-node analysis, Corollary 11.16
Cuspidal; tangent line–conicProposition 11.21
Three concurrent distinct linesProposition 11.22
Double line plus a line; triple lineQuadratic elimination, Proposition 11.25

Table 4.

Attribution and scope

We prove the ternary-cubic normal-form classification and its sharp mixed cutoff, and deduce the dimension–degree boundary. Wilson’s Theorem 1 supplies the unrestricted two-variable result and the classification by dimension alone [13]. The quadratic theorem is due to Liu [7 Theorems 2.1 and 2.2]; we include a short proof of the form used here. Long’s cubic four-variable and quartic three-variable counterexamples are adopted from [8 Proposition 4.1 and Theorem 5.1], with their all-order moment identities reproduced below.

For formal exponential decomposition and local Fourier–Laplace background, see Sabbah [11]. The polynomial stationary-frame argument below allows nonconvergent Laurent coefficients over a valued field. The analytic period comparisons are applied to rational parameter families, whose local Puiseux expansions converge.

A few Gaussian identities

All our generating functions are formal power series. We never need to assume that the integral of etP(X)e^{tP(X)} converges. Define

ΦQ(t)=∑m≥0E[QPm]tmm!(5)\Phi_Q(t)=\sum_{m\geq0}\mathbb{E}[QP^m]\frac{t^m}{m!} \tag*{(5)}

Then PP is pure exactly when Φ1=1\Phi_1=1, and is safe exactly when every ΦQ\Phi_Q is a polynomial. Thus a degree bound on ΦQ\Phi_Q gives an explicit vanishing threshold.

In the circular coordinates of (4), ZZ, WW are polynomials in the independent real Gaussians X1,X2X_1,X_2, but are not independent random variables: W=Z‾W=\overline{Z}. Their moments are

E[ZaWb]=δaba!.(6)\mathbb{E}[Z^aW^b]=\delta_{ab}a!. \tag*{(6)}

Rotating (X1,X2)(X_1,X_2) proves the vanishing when a≠ba\ne b; when a=ba=b, ZW=(X12+X22)/2ZW=(X_1^2+X_2^2)/2 has moments a!a!. An additional independent standard Gaussian TT has E[T2j]=(2j−1)!!\mathbb{E}[T^{2j}]=(2j-1)!! and zero odd moments.

Lemma 2.1 (One sign of weight). Give Z,W,TZ,W,T weights 1,−1,01,-1,0, respectively. If every monomial of PP has strictly positive weight, or every monomial has strictly negative weight, then PP is pure and

E[QPm]=0(m>deg⁡Q).\mathbb{E}[QP^m]=0 \qquad(m>\deg Q).

Proof. In the positive case, every monomial of PmP^m has weight at least mm. A monomial of QQ has weight at least −deg⁡Q-\deg Q. For m>deg⁡Qm>\deg Q their product therefore has nonzero weight, and (6) kills it. The negative case follows by exchanging ZZ and WW.

Complex orthogonal changes of polynomial coordinates preserve the Gaussian moment functional. This follows from its source identity E[euTX]=euTu/2\mathbb{E}[e^{u^{T}X}]=e^{u^{T}u/2}. We use this algebraic invariance; we do not interpret a complex coordinate change as a real change of integration contour.

Lemma 2.2 (Linear circular contraction). Let a,ba,b be constants and let F(t,z)F(t,z) be a formal power series in tt whose coefficients are polynomials in zz. Then

E[F(t,Z)et(a+bZ)W]=11−btF(t,at1−bt).(7)\mathbb{E}\left[F(t,Z)e^{t(a+bZ)W}\right]=\frac{1}{1-bt}F\left(t,\frac{at}{1-bt}\right). \tag*{(7)}

Proof. It suffices, coefficient by coefficient in tt, to take F=ZjF=Z^{j}. Using (6), the left side is

∑k≥jtk[zk]zj(a+bz)k=∑r≥0(j+rj)ajbrtj+r=(at)j(1−bt)j+1.\sum_{k\geq j}t^{k}[z^{k}]z^{j}(a+bz)^{k}=\sum_{r\geq0}\binom{j+r}{j}a^{j}b^{r}t^{j+r}=\frac{(at)^{j}}{(1-bt)^{j+1}}.

Every fixed coefficient uses only finitely many terms, which also justifies the statement for tt-dependent FF.

Quadratics in any number of variables

We include a short proof of the positive quadratic result. It explains why degree two is safe in every dimension without introducing the finite-moment thresholds that are unnecessary for the boundary. The classification and sharper surrounding results are due to Liu [7 Theorem 2.1].

References

  1. [1]H. Derksen, A. van den Essen and W. Zhao, The Gaussian Moments Conjecture and the Jacobian Conjecture, Israel Journal of Mathematics 219 (2017), no. 2, 917–928; arXiv:1506.05192.arxiv.org/abs/1506.05192
  2. [2]NIST Digital Library of Mathematical Functions, 10.25.1 and 10.31.2, modified Bessel equation and logarithmic power series.
  3. [3]D. Drusvyatskiy and A. S. Lewis, Semi-algebraic functions have small subdifferentials, Mathematical Programming, Series B 140 (2013), 5–29, Theorems 2.25 and 2.32; doi:10.1007/s10107-012-0624-x.DOI
  4. [4]R. M. Hardt, P. Lambrechts, V. Turchin and I. Volić, Real homotopy theory of semi-algebraic sets, arXiv:0806.0476v3, Sections 2.1 and 3, especially Theorem 3.5.
  5. [5]L. Simon, Introduction to Geometric Measure Theory, author’s 2017/2018 notes, Chapter 6, Section 3, Theorem 3.15, printed page 175; author’s primary text.
  6. [6]M. Hien and C. Roucairol, Integral representations for solutions of exponential Gauß–Manin systems, Bulletin de la Société Mathématique de France 136 (2008), no. 4, 505–532, doi:10.24033/bsmf.2564.DOI
  7. [7]Z. Liu, Sharp finite Gaussian moments for quadratic polynomials, version 3, public preprint (2026), doi:10.5281/zenodo.21688251.
  8. [8]C. D. Long, Small Counterexamples to the Gaussian Moments Conjecture, public preprint (2026), arXiv:2607.18186v1.arxiv.org/abs/2607.18186
  9. [9]F. Pham, La descente des cols par les onglets de Lefschetz, avec vues sur Gauss–Manin, Astérisque 130 (1985), 11–47, Numdam primary text.
  10. [10]C. Roucairol, Formal structure of direct image of holonomic 𝒟-modules of exponential type, Manuscripta Mathematica 124 (2007), no. 3, 299–318; arXiv:math/0604134v2.arxiv.org/abs/math/0604134
  11. [11]C. Sabbah, An explicit stationary phase formula for the local formal Fourier–Laplace transform, in Singularities I, Contemporary Mathematics 474, American Mathematical Society, 2008, 309–330; arXiv:0706.3570v2.arxiv.org/abs/0706.3570
  12. [12]B. Mannaa and T. Coquand, Dynamic Newton–Puiseux Theorem, public preprint (2013), Corollary 1.10, arXiv:1304.6770v2.DOI
  13. [13]M. Wilson, A face-isolation proof of the two-variable Gaussian Moments Conjecture, public preprint (2026), arXiv:2607.23887v1.arxiv.org/abs/2607.23887

Paper details

vlnc:W_ee6fpw5f
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 5

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

OpenAI ChatGPT, OpenAI Codex

View statement
Looking for related papers…
Contents