🔢 mathematics

Positivity and non-positivity results for the sixth-order $Q$ -curvature of conformal metrics in $\mathbb{R}^n$

This paper establishes that for conformally Euclidean metrics on $\mathbb{R}^n$ , the positivity of the sixth-order $Q$ -curvature ( $Q^{(6)}$ ) is guaranteed when the $2m$ -th order $Q$ -curvature is nonnegative in dimensions up to $4m-6$ , but fails in sufficiently high dimensions where counterexamples can be constructed.

Original authors: Jérôme Vétois, Samuel Zeitler

Published 2026-07-21

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jérôme Vétois, Samuel Zeitler

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 an architect designing a building, but instead of using bricks and mortar, you are weaving the very fabric of space itself. In the world of mathematics, specifically a field called differential geometry, scientists study "manifolds"—shapes that can curve, twist, and stretch, much like a rubber sheet. A key part of understanding these shapes is measuring their "curvature." You might know the curvature of a sphere (it's round) or a saddle (it curves up one way and down the other). But mathematicians have invented a whole family of more complex measurements called "Q-curvatures." Think of these as different layers of a cake: the first layer tells you about the basic shape (like the scalar curvature), the second layer tells you about how that shape is bending in more intricate ways, and so on.

The big question this paper tackles is about the relationship between these layers. If you know that the top, most complex layer of curvature is positive (meaning the space is "bending" in a certain good way everywhere), does that guarantee that the lower, simpler layers are also positive? For a long time, mathematicians believed the answer was a resounding "yes." They thought that if the highest layer was well-behaved, the layers underneath it had to be too. This idea was so strong that it became a guiding rule for solving many other difficult problems in geometry. But, as this paper shows, nature is sometimes more mischievous than our rules predict.

The authors, Jérôme Vétois and Samuel Zeitler, decided to test this rule in a very specific setting: flat space (like our everyday 3D world, but extended to higher dimensions) that has been stretched or squashed in a smooth, uniform way (a "conformal" metric). They focused on a specific scenario involving the sixth-order Q-curvature (a high-level measurement) and asked: if this high-level measurement is positive everywhere, must the third-order measurement (a lower layer) also be positive?

Their findings are a tale of two worlds, depending entirely on the number of dimensions the space has.

The Good News: Small Dimensions
First, they looked at spaces with a "moderate" number of dimensions. Specifically, if the dimension of the space is between $2m$ and $4m-6$ (where $m$ is a number related to the complexity of the curvature being measured, and in this case, $m \ge 4$ ), the old rule holds true. In these dimensions, if the high-level curvature is positive, the lower-level curvature is indeed positive. The authors proved this by turning the problem into a giant algebraic puzzle. They showed that the mathematical formula describing the lower curvature is made up of parts that are all positive, like a recipe where every ingredient adds a sweet flavor. In these specific dimensions, the "sweetness" of the high curvature forces the lower curvature to be positive as well.

The Bad News: Large Dimensions
However, the story takes a sharp turn when the space gets "too big." The authors discovered that if the number of dimensions is large enough (specifically, larger than a certain threshold $N_m$ which depends on $m$ ), the rule breaks down completely.

To prove this, they didn't just guess; they built a mathematical counter-example. Imagine you are stretching a rubber sheet. The authors constructed a specific shape where the high-level curvature is positive everywhere—like a perfectly smooth, gently rising hill. But, surprisingly, at one specific point on this hill, the lower-level curvature dips down and becomes negative. It's as if you have a hill that looks perfectly smooth from a distance, but if you zoom in on one tiny spot, you find a hidden valley.

They found that this "negative dip" happens when the dimension $n$ is greater than a specific number $N_m$ . For example, if $m=4$ , this switch happens when the dimension is greater than about 29.1. If $m=5$ , it happens around 39.7. As the complexity $m$ increases, this threshold grows, but it always exists. In these high-dimensional spaces, the positivity of the top layer does not guarantee the positivity of the layer below it.

Why This Matters
This result is a bit of a shock to the mathematical community because it overturns a conjecture that many experts believed to be true for all dimensions. The authors showed that while the "positivity rule" works for a wide range of dimensions, it is not a universal law. It fails when the space gets sufficiently large.

The paper also notes that these weird, high-dimensional shapes can be mapped onto a sphere (using a technique called stereographic projection), meaning these counter-examples aren't just abstract math; they exist on shapes that are topologically similar to the surface of a ball.

In summary, the paper proves two main things:

For dimensions $n$ between $2m$ and $4m-6$ (with $m \ge 4$ ): If the high-order curvature is positive, the lower-order curvature is strictly positive. This is a proven fact.
For dimensions $n$ larger than a specific threshold $N_m$ : It is possible to construct a shape where the high-order curvature is positive everywhere, but the lower-order curvature is negative at some points. This disproves the idea that the rule holds for all dimensions.

The authors didn't just suggest this might happen; they provided a rigorous mathematical proof and constructed explicit examples to show exactly how and where the rule fails. They showed that in the vast landscape of high-dimensional geometry, the simple intuition that "good top layers mean good bottom layers" is not always true.

Technical Summary: Positivity and Non-Positivity Results for the Sixth-Order Q-Curvature of Conformal Metrics in $\mathbb{R}^n$

Problem Statement
The paper investigates the positivity properties of lower-order $Q$ -curvatures, specifically $Q^{(2k)}_g$ , for conformally Euclidean metrics $g = \phi^2 |dx|^2$ on $\mathbb{R}^n$ . The central question is whether the non-negativity of a higher-order $Q$ -curvature, $Q^{(2m)}_g \geq 0$ (with $Q^{(2m)}_g \not\equiv 0$ ), implies the positivity of lower-order curvatures $Q^{(2k)}_g$ for $k \in \{1, \dots, m-1\}$ .

Previous work by Gursky and Malchiodi [20] established this positivity for $m=2, k=1$ on closed manifolds. Subsequent extensions by Li and Xu [32] and Li, Wei, and Xu [30] confirmed this result for $m \geq 2$ and $k \in \{1, \min(m-1, 2)\}$ on $\mathbb{R}^n$ under specific decay conditions. These authors conjectured that the result holds for all $1 \leq k \leq m-1$ and $n \geq 2m$ . This paper challenges that conjecture by focusing on the case $m \geq 4$ and $k=3$ (the sixth-order curvature, $Q^{(6)}_g$ ).

Methodology
The authors employ an integral representation approach derived from the conformal factor $u$ of the metric. Under the assumption that $Q^{(2m)}_g$ satisfies a specific decay or non-negativity condition (Assumption A), the conformal factor $u$ can be expressed as a convolution involving the measure $d\mu^{(2m)}_g$ defined by the $Q$ -curvature.

Integral Formula Derivation: The core technical contribution of Section 2 is the derivation of an explicit integral formula for $Q^{(6)}_g$ . By applying the Laplacian operator $\Delta$ three times to the conformal factor (specifically to $u^t$ where $t = \frac{n-6}{n-2m}$ ), the authors express $Q^{(6)}_g(x)$ as an integral over $\mathbb{R}^{6n}$ involving a cubic polynomial $P_{n,m}(Z)$ . The variables $Z$ in this polynomial are geometric quantities derived from the distances between points $y_i$ and the evaluation point $x$ (specifically terms like $Z_{i,j} = \frac{|y_i-y_j|^2}{|x-y_i|^2|x-y_j|^2}$ ).
Polynomial Analysis: The sign of $Q^{(6)}_g$ $Q_{g}^{(6)}$ is reduced to the sign of the polynomial $P_{n,m}$ $P_{n, m}$ integrated against a non-negative measure.
- For the positivity result, the authors analyze the coefficients of $P_{n,m}$ for specific ranges of $n$ relative to $m$ . They demonstrate that for certain dimensions, the polynomial can be bounded below by non-negative terms using algebraic inequalities and Fubini's theorem.
- For the non-positivity result, the authors construct a specific family of conformally Euclidean metrics where the measure $d\mu^{(2m)}_g$ concentrates at two distinct points ( $\pm e_1$ ). This reduces the integral to a polynomial in a parameter $r$ (the ratio of masses at the two points). By analyzing the roots of this resulting polynomial, they identify conditions where the integral becomes negative.

Key Results

Theorem 1.1 (Positivity): For $m \geq 4$ , if the dimension $n$ satisfies $2m \leq n \leq 4m-6$ , and the metric satisfies Assumption (A) with $Q^{(2m)}_g \geq 0$ , then $Q^{(6)}_g > 0$ everywhere in $\mathbb{R}^n$ .
- The proof involves dividing the range of $n$ into three cases ( $5m-3 < n < \frac{8m-6}{3}$ , $\frac{8m-6}{3} < n < 3m-3$ , and $3m-3 < n < 4m-6$ ) and proving the positivity of the associated polynomial coefficients or combinations thereof in each case.
Theorem 1.2 (Non-Positivity): For $m \geq 4$ , there exists a threshold $N_m$ (the largest real root of a specific degree-5 polynomial $\gamma_m(n)$ ) such that for all $n > N_m$ , there exist conformally Euclidean metrics where $Q^{(2m)}_g$ is strictly positive everywhere, yet $Q^{(6)}_g$ is negative at some point (specifically at the origin).
- The threshold $N_m$ scales asymptotically as $N_m \sim \Lambda m$ where $\Lambda \approx 10.55$ . For small $m$ , numerical values are provided (e.g., $N_4 \approx 29.1$ ).
- The counterexamples are constructed using metrics with "peaks" of different heights at two points, extending via stereographic projection to metrics conformal to the standard sphere $S^n$ .

Significance and Claims
The paper claims to resolve the conjecture regarding the general positivity of lower-order $Q$ -curvatures. While previous results suggested a universal positivity for $k \in \{1, \dots, m-1\}$ , this work demonstrates that the situation is "surprisingly different" for $k=3$ when the dimension $n$ is sufficiently large relative to $m$ .

Contrast with Prior Work: The results stand in contrast to the findings of Gursky-Malchiodi [20] and Li-Xu [32]/Li-Wei-Xu [30], showing that their positivity results do not extend to all dimensions for $k=3$ .
Dimensional Dependence: The paper establishes a sharp dimensional dichotomy: positivity holds for "low" dimensions ( $n \leq 4m-6$ ) but fails for "high" dimensions ( $n > N_m$ ).
Methodological Contribution: The derivation of the integral formula for $Q^{(6)}_g$ (Proposition 2.1) is presented as a necessary tool for these proofs, reducing the geometric problem to the analysis of a cubic polynomial in multiple non-negative variables.

The authors do not claim to have solved the problem for all $k$ or all $m$ , nor do they propose new applications beyond the theoretical understanding of $Q$ -curvature positivity. The work is framed as a necessary correction to the conjecture that lower-order $Q$ -curvatures are always positive given a non-negative higher-order curvature.

More like this