Joint Cone Property: Coordinated Structures

• Joint cone property is a conceptual umbrella for coordinated cone structures that simultaneously govern families of constraints, operators, or geometric features.
• In matrix analysis and nonlinear SOCP, shared invariant cones and reduced critical-cone geometry enable continuity, convergence, and regularity results.
• Applications span Einstein and Alexandrov geometries, birational settings, and cone-compactness, highlighting distinct compactness and rigidity phenomena.

Searching arXiv for papers relevant to “Joint Cone Property” and closely related cone-based notions across fields. “Joint cone property” is not a uniformly standardized formal term in the surveyed arXiv literature. The available work instead uses closely related cone-based structures to encode coordinated behavior across several objects at once: shared invariant cones for matrix semigroups (Jungers, 2012), coordinated second-order cone block geometry through critical cones and reduction mappings in nonlinear SOCP (Chen et al., 2024), interpolating spectral cone conditions for the curvature operator of the second kind on Einstein manifolds (Cheng et al., 10 Jan 2026), rigidity statements for cones, suspensions, and joins in Alexandrov geometry (Su et al., 2013), and wall-and-reflection cone structures in birational geometry via well-clipped cones (Gachet, 2 Apr 2025). A separate line of work on cone-compactness shows that sequence-based cone behavior and covering-based cone behavior can diverge sharply in nonseparable spaces (Durea et al., 6 Jan 2025). Taken together, these papers suggest that the phrase is best understood as an umbrella label for properties in which a cone organizes simultaneous structure across multiple constraints, operators, or geometric components, rather than as a single canonical definition.

1. Terminological scope and recurring structural pattern

Several of the relevant papers explicitly state that they do not introduce a named notion called “joint cone property.” In nonlinear second-order cone programming, the nearest analogue is the combined treatment of all cone blocks through the critical cone, its affine hull, the reduction matrix $H/\Xi$, and a cone alternative lemma (Chen et al., 2024). In the Einstein-manifold setting, the relevant notion is the one-parameter family of cones $\mathcal C(\alpha,\theta)$, not a formally named “joint cone” (Cheng et al., 10 Jan 2026). In birational geometry, the closest exact match is the hierarchy of well-clipped, neatly clipped, and perfectly clipped cones (Gachet, 2 Apr 2025). In discrete convexity, the paper on the joint winner property states that the structural analogue is $M^\natural$-convex completion rather than a cone property (Iwamasa et al., 2017).

Across these settings, a common pattern nevertheless appears. A cone is used to encode a coordinated restriction on an entire family: a family of matrices acting on a common ordered region (Jungers, 2012), a family of SOCP constraint blocks reduced to a common subspace (Chen et al., 2024), the lower tail of a curvature spectrum relative to its mean (Cheng et al., 10 Jan 2026), or a family of reflection walls cutting a chamber from a larger homogeneous cone (Gachet, 2 Apr 2025). This suggests a broad research usage in which “joint” refers not to a binary operation on cones, but to a cone-based device that simultaneously controls several interacting components.

2. Shared invariant cones for matrix semigroups

The most direct cone-based “joint” framework in the corpus is the invariant-cone theory for matrix semigroups. A proper cone $K\subset \mathbb R^n$ is closed, convex, has nonempty interior, and contains no line. A matrix $A$ is $K$-nonnegative if $AK\subseteq K$, $K$-positive if $AK\subseteq \operatorname{int}(K)$, and $K$-primitive if there exists $\mathcal C(\alpha,\theta)$0 such that $\mathcal C(\alpha,\theta)$1 is $\mathcal C(\alpha,\theta)$2-positive (Jungers, 2012). When a set of matrices preserves the same cone, the paper shows that several joint spectral characteristics become more tractable.

A stronger configuration is an embedded pair of cones: $\mathcal C(\alpha,\theta)$3 is embedded in $\mathcal C(\alpha,\theta)$4 if

$\mathcal C(\alpha,\theta)$5

and $\mathcal C(\alpha,\theta)$6 is invariant if every matrix in the set preserves both cones (Jungers, 2012). This additional structure is central to the continuity theorem for the joint spectral subradius. If $\mathcal C(\alpha,\theta)$7 is a compact set of matrices, $\mathcal C(\alpha,\theta)$8 in the Hausdorff metric, and $\mathcal C(\alpha,\theta)$9 leaves an invariant embedded pair $M^\natural$0 invariant, then

$M^\natural$1

as $M^\natural$2 (Jungers, 2012). The paper emphasizes that $M^\natural$3 is not continuous in general, even for nonnegative matrices, so the invariant embedded cone pair is a substantive structural hypothesis rather than a cosmetic one.

The same paper establishes an asymptotic convergence result for the joint spectral radius under cone invariance plus primitivity. If $M^\natural$4 shares an invariant cone $M^\natural$5 and one $M^\natural$6 is $M^\natural$7-primitive, then both

$M^\natural$8

converge to the joint spectral radius $M^\natural$9 as $K\subset \mathbb R^n$0 (Jungers, 2012). In this setting, a plausible interpretation of “joint cone property” is the existence of a cone or embedded pair jointly preserved by all matrices in the semigroup, since this shared order structure is exactly what enables continuity and convergence statements that fail without it.

3. Cone-compactness versus sequential cone-compactness

A different use of cone structure appears in cone-compactness. Let $K\subset \mathbb R^n$1 be a normed vector space, $K\subset \mathbb R^n$2 a closed convex cone, and $K\subset \mathbb R^n$3 nonempty. The set $K\subset \mathbb R^n$4 is $K\subset \mathbb R^n$5-compact if every cover

$K\subset \mathbb R^n$6

admits a finite subcover (Durea et al., 6 Jan 2025). It is $K\subset \mathbb R^n$7-sequentially compact if for every sequence $K\subset \mathbb R^n$8 there exists $K\subset \mathbb R^n$9 such that $A$0 has a convergent subsequence toward an element of $A$1 (Durea et al., 6 Jan 2025).

The paper proves that cone-compactness implies sequential cone-compactness, but the converse fails in general without separability (Durea et al., 6 Jan 2025). The counterexample is built in

$A$2

the Banach space of all bounded functions $A$3 with the supremum norm, with cone

$A$4

the closed, convex, pointed cone of all nonnegative bounded functions (Durea et al., 6 Jan 2025). The set $A$5 consists of all $A$6-valued functions whose $A$7-set is countable. The paper notes that there are $A$8 such functions, the norm distance between any two distinct elements of $A$9 is $K$0, and hence $K$1 is not separable (Durea et al., 6 Jan 2025).

The non-$K$2-compactness argument uses the cover

$K$3

where $K$4 is the open ball of radius $K$5 centered at $K$6. If a finite subcover existed, the finite union of the relevant countable supports would still be countable, permitting the construction of a new function $K$7 that contradicts membership in every set $K$8 at a point outside that union (Durea et al., 6 Jan 2025). The paper adds that the same argument actually shows that even a countable subcover cannot suffice.

By contrast, $K$9 is $AK\subseteq K$0-sequentially compact in an unusually strong sense. For any sequence $AK\subseteq K$1, one defines $AK\subseteq K$2 by taking $AK\subseteq K$3 for all $AK\subseteq K$4, and then

$AK\subseteq K$5

Hence $AK\subseteq K$6 for every $AK\subseteq K$7, so the adjusted sequence is constant (Durea et al., 6 Jan 2025). Remark 2.2 states that for every sequence $AK\subseteq K$8 there exists $AK\subseteq K$9 such that $K$0 is stationary at some point in $K$1 (Durea et al., 6 Jan 2025).

This example is significant for any generalized “joint cone” perspective because it isolates two fundamentally different roles of a cone: a covering role and a sequential adjustment role. The paper’s conclusion is that sequence-level cone control does not recover covering-level cone control in nonseparable spaces, so a single umbrella phrase can conceal materially different compactness mechanisms (Durea et al., 6 Jan 2025).

4. Coordinated cone-block geometry in nonlinear SOCP

In nonlinear second-order cone programming, the relevant geometry is explicitly multi-block. The basic SOCP has the form

$K$2

with each

$K$3

a second-order cone (Chen et al., 2024). The more general conic form is

$K$4

where $K$5 is a closed convex cone that is $K$6-cone reducible at every point (Chen et al., 2024). The paper states that this reducibility is crucial because it allows chain rules and tangent/normal calculus for the cone constraints.

The perturbed KKT system is written as

$K$7

with solution map

$K$8

The main theorem states that

$K$9

for nonlinear SOCP at a locally optimal solution, and this equivalence is obtained without strict complementarity (Chen et al., 2024).

The cone-geometric machinery is organized around the critical cone

$AK\subseteq \operatorname{int}(K)$0

its affine hull $AK\subseteq \operatorname{int}(K)$1, and a reduction matrix $AK\subseteq \operatorname{int}(K)$2 such that

$AK\subseteq \operatorname{int}(K)$3

with $AK\subseteq \operatorname{int}(K)$4 of full row rank (Chen et al., 2024). Another matrix $AK\subseteq \operatorname{int}(K)$5 satisfying

$AK\subseteq \operatorname{int}(K)$6

is used to rewrite positivity on the critical cone as positivity on a lower-dimensional space (Chen et al., 2024). The paper explicitly describes this reduction as the closest structural analogue to any “joint cone” condition because the analysis is not about one cone alone, but about a coordinated family of cone constraints reduced to a common linear subspace.

The same coordinated role is played by the cone alternative lemma. If $AK\subseteq \operatorname{int}(K)$7 is self-adjoint and $AK\subseteq \operatorname{int}(K)$8 is a closed convex cone with $AK\subseteq \operatorname{int}(K)$9, and if

$K$0

then exactly one of two alternatives holds: either $K$1 for all $K$2, or there exists $K$3 and $K$4 such that

$K$5

(Chen et al., 2024). Combined with explicit coderivative formulas for projection onto the second-order cone and the Mordukhovich criterion for the Aubin property, this yields a blockwise yet unified regularity theory.

A plausible implication is that, in this literature, the most precise meaning of a “joint cone property” is not a standalone axiom but a reduction principle: many cone blocks are analyzed jointly through a single critical-cone geometry, and regularity is decided by quadratic positivity on that reduced structure.

5. Spectral cone conditions and rigidity on Einstein manifolds

For Einstein manifolds, the central object is the curvature operator of the second kind $K$6, acting on the space of traceless symmetric $K$7-tensors

$K$8

with eigenvalues

$K$9

and average

$\mathcal C(\alpha,\theta)$00

(Cheng et al., 10 Jan 2026). For an Einstein manifold, the paper records the relation

$\mathcal C(\alpha,\theta)$01

(Cheng et al., 10 Jan 2026).

The cone condition is a family $\mathcal C(\alpha,\theta)$02 indexed by

$\mathcal C(\alpha,\theta)$03

One has $\mathcal C(\alpha,\theta)$04 if

$\mathcal C(\alpha,\theta)$05

(Cheng et al., 10 Jan 2026). The paper states that this condition is equivalent to saying that the average of the smallest $\mathcal C(\alpha,\theta)$06-portion of eigenvalues is bounded below by a multiple of the mean eigenvalue $\mathcal C(\alpha,\theta)$07, and that $\mathcal C(\alpha,\theta)$08 interpolates between $\mathcal C(\alpha,\theta)$09-nonnegativity when $\mathcal C(\alpha,\theta)$10 and a more flexible pinching condition when $\mathcal C(\alpha,\theta)$11 (Cheng et al., 10 Jan 2026).

The main rigidity result is that a closed Einstein manifold of dimension $\mathcal C(\alpha,\theta)$12 satisfying suitable parameter restrictions and the cone condition is flat or a round sphere (Cheng et al., 10 Jan 2026). The analytic mechanism is a Bochner-type argument proving

$\mathcal C(\alpha,\theta)$13

Since on a closed manifold

$\mathcal C(\alpha,\theta)$14

integration yields $\mathcal C(\alpha,\theta)$15, so the manifold is locally symmetric (Cheng et al., 10 Jan 2026). The cone hypothesis then narrows the symmetric possibilities to the flat and round-sphere cases.

The paper explicitly notes that it does not introduce a “joint cone property.” What it does define is a one-parameter family of cones that combines spectral averaging over the lowest $\mathcal C(\alpha,\theta)$16 eigenvalues with a lower bound relative to $\mathcal C(\alpha,\theta)$17 (Cheng et al., 10 Jan 2026). In that sense, the closest relevant notion is a combined or interpolating cone condition rather than a formally named joint cone.

6. Cones, suspensions, and joins in Alexandrov geometry

In Alexandrov geometry, the relevant structures are geometric rather than order-theoretic. The paper studies glued spaces $\mathcal C(\alpha,\theta)$18 obtained by gluing Alexandrov spaces with isometric boundaries, and analyzes when the resulting space must be a suspension, cone, or join (Su et al., 2013). The standard constructions are recalled explicitly: $\mathcal C(\alpha,\theta)$19 with their corresponding metric formulas (Su et al., 2013).

Theorem A states that if $\mathcal C(\alpha,\theta)$20 with $\mathcal C(\alpha,\theta)$21, and there exist points $\mathcal C(\alpha,\theta)$22 such that

$\mathcal C(\alpha,\theta)$23

then either both pieces have a join/suspension decomposition

$\mathcal C(\alpha,\theta)$24

or each $\mathcal C(\alpha,\theta)$25 has a cone-type decomposition

$\mathcal C(\alpha,\theta)$26

in which case

$\mathcal C(\alpha,\theta)$27

(Su et al., 2013). Corollary 0.2 gives a particularly clear cone rigidity statement: if $\mathcal C(\alpha,\theta)$28 has nonempty boundary and there exists $\mathcal C(\alpha,\theta)$29 with

$\mathcal C(\alpha,\theta)$30

then $\mathcal C(\alpha,\theta)$31 is convex in $\mathcal C(\alpha,\theta)$32 and

$\mathcal C(\alpha,\theta)$33

(Su et al., 2013).

Theorem B analyzes the case where the glued space is a cone $\mathcal C(\alpha,\theta)$34, and Theorem C treats the case where the glued space is a join $\mathcal C(\alpha,\theta)$35 with $\mathcal C(\alpha,\theta)$36, concluding that each glued piece must itself be a join with the same factor (Su et al., 2013). Proposition A.1 states that for $\mathcal C(\alpha,\theta)$37,

$\mathcal C(\alpha,\theta)$38

if and only if

$\mathcal C(\alpha,\theta)$39

(Su et al., 2013).

This literature does not use “joint cone property” as a technical term. The paper’s contribution is instead a rigidity principle: extremal metric relations across a gluing boundary force global decomposition into cones, suspensions, or joins. This suggests a geometric reading of the phrase in which “joint” refers to the way the glued pieces are jointly constrained by the ambient cone or join structure.

7. Chamber structures, finite quotients, and the boundary of the terminology

In birational geometry, the new notion is that of a well-clipped cone. Let

$\mathcal C(\alpha,\theta)$40

with preferred lattice $\mathcal C(\alpha,\theta)$41. A cone $\mathcal C(\alpha,\theta)$42 is well clipped in a self-dual homogeneous cone $\mathcal C(\alpha,\theta)$43 if there is a collection of hyperplanes $\mathcal C(\alpha,\theta)$44 such that

$\mathcal C(\alpha,\theta)$45

subject to three structural conditions: each hyperplane is supported in one indecomposable factor of $\mathcal C(\alpha,\theta)$46, the associated orthogonal reflections preserve the lattice,

$\mathcal C(\alpha,\theta)$47

and the wall normals satisfy

$\mathcal C(\alpha,\theta)$48

for distinct walls (Gachet, 2 Apr 2025). The paper describes the intuition as cutting a cone $\mathcal C(\alpha,\theta)$49 from a larger, very symmetric cone $\mathcal C(\alpha,\theta)$50 using reflection hyperplanes, much like a Coxeter chamber (Gachet, 2 Apr 2025).

A key characterization states that for a well-clipped cone $\mathcal C(\alpha,\theta)$51, the following are equivalent: $\mathcal C(\alpha,\theta)$52 is perfectly clipped, $\mathcal C(\alpha,\theta)$53 is neatly clipped, and $\mathcal C(\alpha,\theta)$54 admits a rational polyhedral fundamental domain on

$\mathcal C(\alpha,\theta)$55

(Gachet, 2 Apr 2025). The paper also proves two stability properties directly relevant to any broad “joint cone” perspective: a direct sum of well-clipped cones is well-clipped, and if $\mathcal C(\alpha,\theta)$56 is well clipped and $\mathcal C(\alpha,\theta)$57 is finite, then the invariant cone

$\mathcal C(\alpha,\theta)$58

is again well clipped, possibly inside a smaller self-dual homogeneous cone $\mathcal C(\alpha,\theta)$59 (Gachet, 2 Apr 2025).

This is the paper in the corpus that comes closest to a combined geometric-and-group-theoretic cone property. The wall structure, arithmetic reflection group, and rational polyhedral fundamental domain are coordinated in a single framework, and that framework is explicitly designed to descend under finite quotients (Gachet, 2 Apr 2025).

A final terminological boundary is supplied by the paper on the joint winner property. It explicitly states that the structural analogue is not a cone property but $\mathcal C(\alpha,\theta)$60-convex completion of an associated quadratic $\mathcal C(\alpha,\theta)$61-function, characterized by a graph cycle condition on an assignment graph (Iwamasa et al., 2017). The main theorem there is

$\mathcal C(\alpha,\theta)$62

(Iwamasa et al., 2017). This serves as a useful corrective: not every “joint” structural condition in adjacent literatures is profitably recast as a cone property.

The surveyed literature therefore supports a restrained conclusion. “Joint cone property” is best treated as a contextual label rather than a settled universal definition. In matrix analysis it most naturally means a shared invariant-cone framework (Jungers, 2012); in nonlinear SOCP, a jointly reduced critical-cone geometry across cone blocks (Chen et al., 2024); in curvature and Alexandrov geometry, an interpolating or rigidity-enforcing cone condition (Cheng et al., 10 Jan 2026, Su et al., 2013); and in birational geometry, a chamber cut from a larger symmetric cone with reflection-group control (Gachet, 2 Apr 2025). The literature also shows that cone-based notions that look similar at the sequential level and at the covering level need not coincide (Durea et al., 6 Jan 2025), and that some nearby “joint” theories are fundamentally discrete-convex rather than cone-theoretic (Iwamasa et al., 2017).