Cone domains separate FS-domains from RB-domains
Pith reviewed 2026-07-03 01:27 UTC · model grok-4.3
The pith
Cone domains D_C are RB-domains exactly when the cone C is simplicial.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Let C be a closed, convex, pointed and generating cone in a finite-dimensional real vector space V, and let D_C = (-C) ∪ {⊥} be the negative cone with a new least element, ordered by the cone order. D_C is an RB-domain if and only if C is simplicial. The proof converts the RB approximation property into finite-valued C-isotone approximations of the identity. The analytic obstruction is that cone-upper sets are represented up to null sets as Lipschitz epigraphs so that Rademacher's theorem, Fubini's theorem and integration by parts force any tested matrix to lie in the cone generated by the positive rank-one operators v ⊗ ℓ with v in C and ℓ in C*. If such maps approximate the identity, the i
What carries the argument
The cone generated by positive rank-one operators v ⊗ ℓ with v ∈ C and ℓ ∈ C*, which must contain the identity for the RB approximation property to hold.
If this is right
- Non-simplicial cones produce explicit FS-domains that fail to be RB-domains.
- The Lorentz cone and other non-simplicial proper cones serve as counterexamples to the claim that all cone domains are RB-domains.
- Keimel's question receives a negative answer precisely when the cone is non-simplicial.
Where Pith is reading between the lines
- The finite-dimensional geometric obstruction may suggest similar rank-one cone conditions in related approximation properties for ordered sets.
- Classification of domains by their geometric representations could separate further classes beyond FS and RB.
Load-bearing premise
Cone-upper sets can be represented up to null sets as Lipschitz epigraphs so that Rademacher's theorem applies and forces the tested matrix to lie in the cone generated by positive rank-one operators.
What would settle it
A direct computation showing whether the identity operator on V belongs to the cone generated by v ⊗ ℓ for v in C and ℓ in C* when C is a non-simplicial cone such as the Lorentz cone.
read the original abstract
Let $C$ be a closed, convex, pointed and generating cone in a finite-dimensional real vector space $V$, and let \( D_C=(-C)\cup\{\bot\}\) be the negative cone with a new least element, ordered by the cone order. Keimel proved that these cone domains are FS-domains and asked whether they are always retracts of bifinite domains. We give a sharp answer: \[D_C\text{ is an RB-domain}\quad\Longleftrightarrow\quad C\text{ is simplicial}. \] Thus every non-simplicial proper cone gives an FS-domain which is not an RB-domain. The proof converts the RB approximation property into finite-valued $C$-isotone approximations of the identity. The analytic obstruction is elementary and finite-dimensional: first in Euclidean space, cone-upper sets are represented, up to null sets, as Lipschitz epigraphs; Rademacher's theorem, Fubini's theorem and integration by parts then force the matrix tested against any finite-valued isotone map to lie in the cone generated by the positive rank-one operators $v\otimes\ell$, $v\in C$, $\ell\in C^*$. If such maps approximate the identity, the identity operator lies in this rank-one cone, which is possible exactly when the cone is simplicial. This answers Keimel's question in the negative for the Lorentz cone and other non-simplicial cones.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript proves that for a closed convex pointed generating cone C in a finite-dimensional real vector space V, the cone domain D_C = (-C) ∪ {⊥} is an RB-domain if and only if C is simplicial. Keimel had shown these domains are always FS-domains; the paper converts the RB-approximation property into a directed family of finite-valued C-isotone maps approximating the identity, then applies an analytic argument (Lipschitz epigraph representations of cone-upper sets up to null sets, Rademacher's theorem, Fubini, and integration by parts) to show that the identity must lie in the cone generated by positive rank-one operators v ⊗ ℓ (v ∈ C, ℓ ∈ C*), which holds precisely when C is simplicial. This yields FS-domains that are not RB-domains for any non-simplicial proper cone, such as the Lorentz cone.
Significance. If the result holds, it supplies a sharp, geometrically concrete separation between FS-domains and RB-domains, resolving Keimel's question in the negative. The finite-dimensional analytic technique for verifying approximation properties in domain theory is novel and may connect order-theoretic and measure-theoretic methods more broadly. The argument is direct, uses only standard theorems, and contains no free parameters or circular reductions.
major comments (1)
- [analytic obstruction / only-if direction] In the 'only if' direction (analytic obstruction paragraph of the abstract and the corresponding proof section): the argument establishes that each finite-valued isotone approximant satisfies the rank-one matrix condition almost everywhere with respect to its own null set arising from the Lipschitz epigraph representation. Because the directed family need not be countable and the null sets are not shown to be controlled uniformly (e.g., via a common full-measure set or cofinal countable subsequence with uniform Lipschitz bounds), it does not immediately follow that the identity operator itself satisfies the condition on a single full-measure set. An explicit argument addressing accumulation of exceptional sets is required to conclude that the identity lies in the rank-one cone.
Simulated Author's Rebuttal
We thank the referee for the careful reading of the manuscript and for identifying a point in the only-if direction that requires clarification. We address the major comment below and will revise the manuscript to incorporate an explicit argument on the control of null sets.
read point-by-point responses
-
Referee: In the 'only if' direction (analytic obstruction paragraph of the abstract and the corresponding proof section): the argument establishes that each finite-valued isotone approximant satisfies the rank-one matrix condition almost everywhere with respect to its own null set arising from the Lipschitz epigraph representation. Because the directed family need not be countable and the null sets are not shown to be controlled uniformly (e.g., via a common full-measure set or cofinal countable subsequence with uniform Lipschitz bounds), it does not immediately follow that the identity operator itself satisfies the condition on a single full-measure set. An explicit argument addressing accumulation of exceptional sets is required to conclude that the identity lies in the rank-one cone.
Authors: We agree that the manuscript would benefit from an explicit treatment of how the directed family of approximants yields a uniform full-measure set for the identity operator. Because the underlying vector space is finite-dimensional, the domain is separable, and the approximation property is defined pointwise via the way-below relation, a countable cofinal subfamily can always be extracted while preserving the directed supremum property. The exceptional null sets of this countable subfamily then have a common null set of full measure by countable additivity. We will add a short paragraph immediately after the application of Rademacher's theorem that spells out this extraction and the resulting uniform control. This change does not alter the overall argument but makes the passage from the approximants to the identity fully rigorous. revision: yes
Circularity Check
No circularity; central claim derived from external theorems without reduction to inputs
full rationale
The paper converts the RB-approximation property into a directed family of finite-valued C-isotone maps approximating the identity, then applies Rademacher's theorem, Fubini's theorem and integration by parts to Lipschitz epigraph representations of cone-upper sets (up to null sets) to conclude that the tested matrix lies in the rank-one cone generated by v⊗ℓ. This forces the identity operator into that cone precisely when C is simplicial. All steps invoke standard, externally verifiable theorems (Rademacher, Fubini) applied directly to the cone order definition; no parameters are fitted to data, no self-citations are load-bearing, and the target statement is not defined in terms of itself. The argument is self-contained against external benchmarks and receives score 0.
Axiom & Free-Parameter Ledger
axioms (2)
- standard math Rademacher's theorem: Lipschitz functions are differentiable almost everywhere
- standard math Fubini's theorem and integration by parts on Lipschitz epigraphs
Reference graph
Works this paper leans on
-
[1]
S. Abramsky and A. Jung, Domain theory, in S. Abramsky, D. M. Gabbay, and T. S. E. Maibaum (eds.),Handbook of Logic in Computer Science, Vol. 3, Oxford University Press, Oxford, 1994, pp. 1–168. Corrected expanded online version available from the author’s webpage
work page 1994
-
[2]
S. Boyd and L. Vandenberghe,Convex Optimization, Cambridge University Press, Cambridge, 2004. Online version available from the authors’ webpage
work page 2004
-
[3]
J. van Dobben de Bruyn, Tensor products of convex cones, part II: closed cones in finite-dimensional spaces,Linear and Multilinear Algebra71 (2023), no. 12, 2031–2050. Preprint version: arXiv:2009.11843
-
[4]
L. C. Evans and R. F. Gariepy,Measure Theory and Fine Properties of Functions, revised edition, CRC Press, Boca Raton, 2015
work page 2015
-
[5]
A. Jung,Cartesian Closed Categories of Domains, CWI Tracts 66, Centrum voor Wiskunde en Infor- matica, Amsterdam, 1989
work page 1989
-
[6]
K.Keimel, Bicontinuousdomainsandsomeoldproblemsindomaintheory,Electronic Notes in Theoretical Computer Science257 (2009), 35–54. 18
work page 2009
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.