---
title: Schneider's Conjecture in Convex Geometry
url: https://www.emergentmind.com/topics/schneider-s-conjecture
type: topic
---

# Schneider's Conjecture in Convex Geometry

Searching arXiv for the cited convex-geometry paper and closely related Schneider-conjecture context.
Schneider’s conjecture, in the convex-geometric sense attached to Minkowski’s monotonicity problem, is the conjectural description of the support of mixed area measures in terms of “extreme directions.” In the formulation studied in "On Minkowski's monotonicity problem" [2507.20082], it arises from the classical question, dating back to Minkowski, of determining the equality cases in the monotonicity of mixed volumes. The conjecture asserts that for convex bodies \(C_1,\dots,C_{n-1}\subset \mathbb R^n\), the support of the mixed area measure \(S_{C_1,\dots,C_{n-1}}\) is exactly the set of \((C_1,\dots,C_{n-1})\)-extreme directions. The 2025 paper resolves one direction of the conjecture for arbitrary convex bodies in every dimension and proves the full conjecture in \(\mathbb R^3\), equivalently for mixed area measures of the form \(S_{K,\ldots,K,L}\) in arbitrary dimension [2507.20082].

## 1. Mixed volumes, mixed area measures, and the equality problem

Let \(K_1,\dots,K_m\) be convex bodies in \(\mathbb R^n\), and let \(\lambda_1,\dots,\lambda_m\ge 0\). Minkowski’s theorem states that
\[
\operatorname{Vol}_n(\lambda_1 K_1+\cdots+\lambda_m K_m)
=
\sum_{i_1,\dots,i_n=1}^m
V_n(K_{i_1},\dots,K_{i_n})\,\lambda_{i_1}\cdots \lambda_{i_n}.
\]
The coefficients \(V_n(K_1,\dots,K_n)\) are the mixed volumes. They are symmetric, multilinear with respect to Minkowski addition, translation invariant, and nonnegative [2507.20082].

A basic property is Minkowski monotonicity: if \(K\subseteq L\), then for any convex bodies \(C_1,\dots,C_{n-1}\),
\[
V_n(K,C_1,\dots,C_{n-1})\le V_n(L,C_1,\dots,C_{n-1}). \tag{1}
\]
The associated classical problem is to characterize the equality cases in this inequality.

The support-function formalism converts this into a problem on measures on the sphere. For a convex body \(K\), the support function is
\[
h_K(u)=\sup_{x\in K}\langle u,x\rangle,\qquad u\in S^{n-1}.
\]
For convex bodies \(C_1,\dots,C_{n-1}\), the mixed area measure \(S_{C_1,\dots,C_{n-1}}\) is the finite Borel measure on \(S^{n-1}\) characterized by
\[
V_n(K,C_1,\dots,C_{n-1})
=
\frac1n
\int_{S^{n-1}} h_K\, dS_{C_1,\dots,C_{n-1}}. \tag{2}
\]
Its support is the smallest closed subset of \(S^{n-1}\) outside of which the measure vanishes.

This formulation is central because it shows that the equality problem for monotonicity is equivalent to geometrically characterizing \(\operatorname{supp} S_{C_1,\dots,C_{n-1}}\). Schneider’s conjecture is precisely such a characterization [2507.20082].

## 2. Geometric language: faces, normal cones, touching cones

For a convex body \(C\) and \(u\in S^{n-1}\), the exposed face is
\[
F(C,u)=\{x\in C:\langle u,x\rangle=h_C(u)\}.
\]
Its normal cone is
\[
N(C,x)=\{u\in\mathbb R^n:\langle u,x\rangle=h_C(u)\},
\]
which is constant on the relative interior of a face \(F\), hence may be written \(N(C,F)\).

The relevant nonsmooth replacement for a tangent space is the touching cone \(T(C,u)\), defined as the unique face of the normal cone \(N(C,F(C,u))\) that contains \(u\) in its relative interior [2507.20082]. For \(I\subseteq [n-1]\), the notation
\[
C_I:=\sum_{i\in I} C_i
\]
is used.

Schneider’s insight was that the support of the mixed area measure should not be described by the dimensions of exposed faces, but rather by the dimensions of the “tangent directions” \(T(C_I,u)^\perp\). This distinction matters because the naive face-dimension criterion works for polytopes but fails for general convex bodies: strictly convex bodies have only \(0\)-dimensional exposed faces, so exposed-face dimensions alone cannot capture the general situation [2507.20082].

Two notions are then defined.

A direction \(u\in S^{n-1}\) is \((C_1,\dots,C_{n-1})\)-extreme if
\[
\dim\bigl(T(C_I,u)^\perp\bigr)\ge |I|
\qquad\text{for all } I\subseteq [n-1]. \tag{3}
\]
Lemma 2.6 in the paper gives equivalent formulations: this is equivalent to requiring
\[
\dim\Bigl(\sum_{i\in I} T(C_i,u)^\perp\Bigr)\ge |I|
\qquad\text{for all } I\subseteq [n-1],
\]
and also equivalent to the existence of lines \(L_i\subseteq T(C_i,u)^\perp\), \(i=1,\dots,n-1\), with linearly independent directions [2507.20082].

A stronger notion is that of \((C_1,\dots,C_{n-1})\)-exposed direction:
\[
\dim\bigl(N(C_I,F(C_I,u))^\perp\bigr)\ge |I|
\qquad\text{for all } I\subseteq[n-1]. \tag{4}
\]
Because \(T(C_I,u)\subseteq N(C_I,F(C_I,u))\), every exposed direction is extreme.

These definitions isolate the geometric conditions conjecturally controlling the support of mixed area measures.

## 3. Statement of Schneider’s conjecture

In the notation above, Schneider’s 1985 conjecture is
\[
\operatorname{supp} S_{C_1,\dots,C_{n-1}}
=
\{u\in S^{n-1}: u \text{ is } (C_1,\dots,C_{n-1})\text{-extreme}\}. \tag{SC}
\]
Thus the conjecture identifies the support of the mixed area measure exactly with the set of extreme directions [2507.20082].

For polytopes, the conjecture reduces to a face-dimension criterion. When all \(C_i\) are polytopes,
\[
S_{C_1,\dots,C_{n-1}}(\{u\})
=
V_{n-1}(F(C_1,u),\dots,F(C_{n-1},u)), \tag{5}
\]
and positivity of mixed volumes yields
\[
\operatorname{supp} S_{C_1,\dots,C_{n-1}}
=
\left\{
u\in S^{n-1}:
\dim F(C_I,u)\ge |I|
\ \forall I\subseteq[n-1]
\right\}. \tag{6}
\]
This polyhedral formula gives an important sanity check for the conjecture, but it does not extend directly beyond the polyhedral setting.

A common misconception is to treat Schneider’s conjecture as merely a statement about exposed faces. The 2025 paper emphasizes that this is not the correct general formulation. The conjecture is intrinsically about touching cones and extreme directions, with exposed directions serving only as a stronger auxiliary notion [2507.20082].

## 4. Equivalence with equality in Minkowski monotonicity

The support characterization is equivalent to the equality problem for Minkowski monotonicity. If \(K\subseteq L\), then
\[
h_L-h_K\ge 0
\quad\text{on } S^{n-1},
\]
and therefore
\[
V_n(L,C_1,\dots,C_{n-1})-V_n(K,C_1,\dots,C_{n-1})
=
\frac1n\int (h_L-h_K)\,dS_{C_1,\dots,C_{n-1}}. \tag{7}
\]
Since the integrand is nonnegative, equality
\[
V_n(K,C_1,\dots,C_{n-1})=V_n(L,C_1,\dots,C_{n-1}) \tag{8}
\]
holds if and only if
\[
h_K(u)=h_L(u)
\qquad\text{for all }u\in \operatorname{supp} S_{C_1,\dots,C_{n-1}}. \tag{9}
\]

Geometrically, equality means that \(K\) and \(L\) have the same supporting hyperplanes in every outer normal direction belonging to the support of the mixed area measure. Hence the monotonicity-equality problem and the support-characterization problem are essentially the same problem [2507.20082].

This equivalence explains the significance of Schneider’s conjecture. It is not an isolated measure-theoretic description; it is the geometric mechanism behind all equality cases in one of the basic monotonicity principles of mixed volume theory.

## 5. The 2025 resolution: upper bound in all dimensions, full theorem in dimension \(3\)

The paper "On Minkowski's monotonicity problem" proves one inclusion in Schneider’s conjecture for arbitrary convex bodies in every dimension and proves the full conjecture in \(\mathbb R^3\) [2507.20082].

The first main theorem establishes the upper-bound direction:
\[
S_{C_1,\dots,C_{n-1}}
\Bigl(
\{u\in S^{n-1}: u \text{ is not } (C_1,\dots,C_{n-1})\text{-exposed}\}
\Bigr)=0. \tag{10}
\]
Consequently,
\[
\operatorname{supp} S_{C_1,\dots,C_{n-1}}
\subseteq
\{u: u \text{ is } (C_1,\dots,C_{n-1})\text{-exposed}\}
\subseteq
\{u: u \text{ is } (C_1,\dots,C_{n-1})\text{-extreme}\}. \tag{11}
\]
In particular,
\[
\operatorname{supp} S_{C_1,\dots,C_{n-1}}
\subseteq
\{\text{extreme directions}\}. \tag{12}
\]
This is one half of Schneider’s conjecture in full generality.

The proof proceeds through a stronger statement: for \(K,L\in\mathcal K^n\) and \(k=0,\dots,n-2\),
\[
S_{K[k+1],L[n-k-2]}
\bigl(
\{u\in S^{n-1}:\dim N(K,F(K,u))\ge n-k\}
\bigr)=0. \tag{13}
\]
This is then transferred to general tuples by applying it to Minkowski sums \(C_I\) [2507.20082].

The second main theorem proves the converse inclusion for a large class of mixed area measures, namely those of the form \(S_{K,\ldots,K,L}\). The paper states that it proves the full conjecture in dimension \(3\), equivalently for mixed area measures of the form \(S_{K,\ldots,K,L}\) in arbitrary dimension. This yields a complete characterization in \(\mathbb R^3\) [2507.20082].

The following table summarizes the status established in the paper.

| Setting | Result proved | Formulation |
|---|---|---|
| Arbitrary convex bodies in \(\mathbb R^n\) | One direction | \(\operatorname{supp} S_{C_1,\dots,C_{n-1}}\subseteq\{\text{extreme directions}\}\) |
| \(\mathbb R^3\) | Full conjecture | \(\operatorname{supp} S_{C_1,C_2}=\{\text{extreme directions}\}\) |
| Measures \(S_{K,\ldots,K,L}\) in arbitrary dimension | Full conjecture | Equivalent formulation of the dimension-\(3\) result |

This substantially advances the geometric understanding of equality in Minkowski’s monotonicity inequality.

## 6. Consequences, analogies, and geometric significance

Among the implications emphasized in the paper is a mixed counterpart of the classical fact, due to Monge, Hartman–Nirenberg, and Pogorelov, that a surface with vanishing Gaussian curvature is a ruled surface [2507.20082]. The paper presents this as an implication of the new support characterization results.

This connection clarifies the geometric depth of Schneider’s conjecture. The conjecture is not only about where a measure lives on the sphere; it encodes degenerate geometric behavior of convex bodies and their mixed interactions. The appearance of a ruled-surface analogue suggests that mixed area measures detect a form of directional flatness analogous to vanishing Gaussian curvature.

A plausible implication is that the support of mixed area measures functions as a refined invariant of the directional geometry of multiple bodies simultaneously, rather than of a single body in isolation. The paper’s use of touching cones instead of exposed faces strongly supports this interpretation.

The work also sharpens the conceptual distinction between polyhedral and nonpolyhedral convexity. In the polyhedral regime, positivity of mixed volumes reduces the problem to dimension conditions on exposed faces. For arbitrary bodies, the correct invariant is the touching-cone geometry. This shift is one of the central structural lessons of the modern formulation [2507.20082].

## 7. Relation to prior status and remaining directions

Before the 2025 advance, Schneider’s conjectural characterization had been verified only for special classes of convex bodies. The paper explicitly states that Schneider’s conjecture had “been verified to date only for special classes of convex bodies” and that the new results resolve one direction for arbitrary convex bodies and the full conjecture in dimension \(3\) [2507.20082].

The current status may therefore be summarized as follows. The general upper-bound inclusion is now established in all dimensions; the full characterization is known in dimension \(3\); and the conjecture remains open in full generality in higher dimensions for arbitrary tuples \((C_1,\dots,C_{n-1})\).

Another possible misconception is to regard the dimension-\(3\) theorem as merely low-dimensional bookkeeping. The paper’s equivalence between the \(\mathbb R^3\) result and the general class \(S_{K,\ldots,K,L}\) shows that the three-dimensional case controls an important family of mixed area measures in arbitrary dimension [2507.20082].

This suggests two principal research directions. One is to extend the lower-bound direction beyond the class \(S_{K,\ldots,K,L}\) to fully arbitrary tuples in higher dimensions. The other is to deepen the relation between mixed area measure support and differential-geometric rigidity phenomena, in the spirit of the ruled-surface implication highlighted by the paper. These are inferences from the structure of the results rather than claims stated as completed theorems.

In contemporary convex geometry, Schneider’s conjecture occupies a precise position at the interface of mixed volumes, support measures, and geometric rigidity. Its modern formulation through touching cones has clarified the correct general invariant, and the results of 2025 place the equality theory of Minkowski monotonicity on substantially firmer ground [2507.20082].

Source: https://www.emergentmind.com/topics/schneider-s-conjecture