{"id":"17f36d37-5399-4d52-85b8-31fd21213dad","arxiv_id":"2411.16209","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An asymmetric convex cone in a topology-free infinite-dimensional real vector space is exactly characterized by the family of step-linear functions that are positive on it and zero on its associated linear subspace.","lead":"This paper finds a way to break any convex cone in an infinite-dimensional space into simple 'open components' and then uses those components to describe conical halfspaces and arbitrary cones with special functions called step-linear functions. The work extends the classical link between halfspaces and linear functions, which is useful in convex analysis and optimization.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 49's characterization of asymmetric cones depends entirely on the unproved imported separation theorem (Theorem 47); without a proof of that result, the analytical representation is conditional.","rationale":"The reader's weakest_assumption is exactly the point I would stress. The open-component and conical-halfspace results (Theorems 13, 30, 42) are proved in the text and appear correct; the representation of conical halfspaces by corteges and step-linear functions is internally coherent. The only unproved input is Theorem 47. Because Theorem 49 uses it both for the existence of a regular extension and for the pointwise separation of every exterior point, the central claim cannot be rated above conditional until that theorem is supplied or independently verified. I agree with the reader; no additional load-bearing concern was identified, and I am not recommending a change of verdict.","tokens_in":23770,"tokens_out":12651,"duration_ms":247143,"concrete_test":"Verify Theorem 47 by locating its proof in [11] and checking that the hypotheses match the statement used here: K1 asymmetric, K2 any convex cone, and X an arbitrary real vector space with no topology. Independently, attempt a direct proof by Zorn's lemma: take a maximal convex cone M containing K1 and disjoint from K2, and prove that X \\ M is a cone. If a counterexample emerges, or if the cited theorem requires an additional hypothesis not satisfied in this paper, then Theorem 49 is not established as stated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 49 is the paper's main analytical-representation claim. Its proof invokes Theorem 47 in two places: first, via Theorem 48, to regularly extend K to an asymmetric conical halfspace H by separating K from LK; second, in proving (27) and (28), to separate conv(K ∪ l_y) from LK for each y outside K ∪ (−K) ∪ LK, producing a step-linear u with u(y) < 0. Theorem 47 is stated without proof and imported from [11]; it asserts a very strong separation property in arbitrary real vector spaces with no topology. If it fails in this generality, or if the version in [11] carries extra hypotheses — for example, restrictions on K2, a nonempty algebraic interior condition, or a finite-dimensionality assumption — then U_K can be empty or fail to separate individual points, and both (27) and (28) fail. The manuscript explicitly flags Theorem 47 as 'presented here without proof,' so the conditional status is acknowledged in-text. No independent error was found in the rest of the chain: the open-component results, the structure of conical halfspaces, and the construction of step-linear functions from corteges all appear internally consistent.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies convex cones in real vector spaces with no topology. It introduces the dominance relation and defines 'open components' of a convex cone as its equivalence classes; it proves that the family of open components is an upper semilattice (Theorem 13) and that each open component is the intrinsic core of a minimal face (Theorem 18). It compares this 'internal geometric structure' with the facial structure and exhibits an infinite-dimensional example where the two differ (Example 24). For conical halfspaces, the paper shows that their open components are linearly ordered and each is itself a conical halfspace in its linear hull (Theorem 30), and that each asymmetric conical halfspace is represented exactly by a step-linear function (Theorem 42). The final theorem (Theorem 49) claims that every asymmetric convex cone K is characterized by the family U_K of step-linear functions that are positive on K and zero on L_K, via the equivalence x∈K iff u(x)>0 for all u∈U_K, and x∈L_K iff u(x)=0 for all u∈U_K. The proofs through Theorem 42 are detailed and the examples check out. The central result Theorem 49, however, relies on a separation theorem (Theorem 47) that is stated without proof and imported from a previous paper.","tokens_in":23984,"tokens_out":13405,"duration_ms":118741,"significance":"If the results hold, the paper makes a substantial contribution to convex analysis in general vector spaces: it provides a new structural decomposition of convex cones (open components and their upper semilattice), clarifies the relation to faces, and gives an exact analytical representation of arbitrary asymmetric convex cones by step-linear functions, extending the classical linear-functional representation of algebraically open halfspaces. Theorems 13, 18, 30 and 42 are proved in detail and are internally consistent; the examples (4, 11, 24, 45) are instructive and check out. The main weakness is that the final representation theorem depends on the unproved external separation theorem, so the significance is conditional unless that theorem is supplied or precisely referenced. The paper also contains a number of new technical tools (the cortege structure of linear functionals and step-linear functions) that are likely to be useful in further work.","major_comments":[{"comment":"Theorem 47 is stated without proof and is load-bearing for the paper's main theorem. It is used in Theorem 48 to separate K from L_K, and in the proof of Theorem 49 it is used twice: to separate conv(K ∪ l_y) from L_K for y outside K, and to construct the step-linear function with u(y) < 0. Since the theorem asserts a very strong separation property in arbitrary real vector spaces with no topology, I ask that the author either provide a complete proof (e.g., in an appendix) or give the exact statement as proved in [11], with all hypotheses explicitly listed and with a precise reference to the theorem number in [11]. Without this, Theorems 48 and 49 remain conditional on an unverified input.","section":"Section 6, Theorem 47"},{"comment":"In the proof of (27), the assertion 'It is easy to verify that conv(K ∪ l_y) is an asymmetric convex cone' is not demonstrated. This property is necessary to apply Theorem 47 with K1 = conv(K ∪ l_y). It is not immediate for arbitrary asymmetric K and arbitrary y in X \\ ((-K) ∪ L_K ∪ K), so the proof should justify it. Please add a short argument or a reference.","section":"Section 6, proof of Theorem 49"}],"minor_comments":[{"comment":"The phrase 'the partial ordered family' appears in the abstract and again later; it should read 'the partially ordered family'.","section":"Abstract and Section 2.2"},{"comment":"The sentence 'E1 ⊴*_K E2 holds for E1,E2 ∈ O(K) if and only if x1 ⊴_K x2 for all (some) x1 ∈ E1 and all (some) x2 ∈ E2' is ambiguous. Since the relation is well-defined on equivalence classes, it should say 'for all x1 ∈ E1 and all x2 ∈ E2 (equivalently, for some x1 ∈ E1 and some x2 ∈ E2)'.","section":"Section 2.2, definition of the partial order ⊴*_K"},{"comment":"The theorem states 'Let H ⊂ X be an asymmetric convex halfspace in X'; this should be 'asymmetric conical halfspace'.","section":"Theorem 38, first sentence"},{"comment":"In the display defining E_m, the condition 'lm(x) > 0' is used, but the quantifier on m is not shown in the displayed set; it should be made explicit that m ∈ Z.","section":"Example 24, definition of E_m"},{"comment":"The phrase 'H is algebraic open halfspace' should be 'H is an algebraically open halfspace' (the adverbial form is missing).","section":"Proof of Theorem 43, part (i) ⇒ (iii)"},{"comment":"The proof invokes Corollary 5.62 of [1] to separate E and L_E. Since the paper works in the purely algebraic setting, it would be helpful to state explicitly that this is an algebraic separation result or to give a direct proof using the algebraic Hahn-Banach extension theorem.","section":"Proposition 34"}],"recommendation":"major_revision","confidential_remarks":"The paper is well-written and the internally developed results appear sound. The main concern is the unresolved status of Theorem 47, which is essential for the final representation theorem. If the author can supply a proof or a precise reference, the paper would likely be acceptable. The fit with the journal's scope is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this for the open components. The idea of defining an equivalence relation on a convex cone from the dominance relation and showing the quotient is an upper semilattice (Theorem 13) is genuinely new and does real work. Example 24, where the internal structure differs from the facial structure in infinite dimensions, is concrete and correct. Sections 2–3 check out; the proofs are detailed enough that I could follow them without filling in much. The step-linear representation of conical halfspaces (Theorem 42) also holds up: it follows from the structure theorem plus Aliprantis–Border separation, and the characterization via corteges is clean.\n\nThe soft spot is exactly where the reader says it is: Theorem 47. It is stated without proof, imported from [11], and it carries the whole final representation. Theorem 49 uses it twice—once to regularly extend K to a halfspace, once to separate each outside point from L_K. If that theorem has hidden hypotheses, or is not true at the level of generality stated, then (27) and (28) fail. The paper does flag it as unproved, which is honest, but a referee will want either a proof or a precise statement of the version in [11] with its hypotheses spelled out. This is a conditional result, not a demonstrated one.\n\nOne more thing to ask the author: clarify the novelty boundary. Step-linear functions and their role in halfspace representation were introduced in [10,11]. What exactly is new in Theorem 49 beyond assembling those pieces? The open-component structure is clearly new, but the representation theorem's newness relative to prior work is understated. Minor: some algebraic steps in Proposition 3 and Theorem 13 are compressed, but recoverable.\n\nBottom line: This paper deserves a serious referee. The open-component semilattice and the halfspace representation are real contributions. The final representation is promising but conditional on an unproved separation theorem. I would send it to review and ask the author to include the missing proof or a precise citation, and to sharpen the novelty discussion. I would not cite the main representation theorem in my own work until that is resolved, but I would cite the open-component structure.","headline":"A solid structural theory of convex cones via open components, but the main analytical representation rests on an unproved imported separation theorem.","tokens_in":24480,"tokens_out":2817,"would_cite":true,"duration_ms":27988,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52A05","06F20"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every asymmetric convex cone in a real vector space is exactly the common positive set of its step-linear functionals, giving a topology-free analytical representation.","keywords":["convex cone","open component","conical halfspace","step-linear function","internal geometric structure","relative algebraic interior","upper semilattice","analytical representation"],"falsifier":"Find an asymmetric convex cone $K$ and a vector $y \\notin K$ such that every step-linear function positive on $K$ and zero on $L_K$ is also positive at $y$; Theorem 49 predicts that such a $y$ cannot exist. Equivalently, exhibit disjoint convex cones $K_1$ and $K_2$ with $K_1$ asymmetric that are not separated by any asymmetric conical halfspace, which would disprove the imported Theorem 47.","tokens_in":5,"feed_emoji":"📐","tokens_out":7841,"duration_ms":126313,"temperature":0.7,"pith_summary":"The paper develops a topology-free description of convex cones in infinite-dimensional real vector spaces, built from open components rather than faces. It proves that every convex cone is a disjoint union of its open components, that this family is partially ordered as an upper semilattice, and that this order structure is the cone's internal geometric structure. For conical halfspaces the components are linearly ordered, and each asymmetric conical halfspace is represented by a step-linear function whose positive set is the halfspace. The main result (Theorem 49) states that any asymmetric convex cone $K$ is exactly the set of points at which every step-linear function in a certain family $\\mathcal{U}_K$ is positive, with the associated subspace $L_K$ captured by vanishing. This gives an analytic dual description of cones in spaces carrying no norm and no topology.","feed_headline":"Step-linear functions exactly encode asymmetric convex cones","feed_subtitle":"No topology needed: a cone is exactly the set where every step-linear functional in its family is positive.","key_machinery":"The load-bearing notion is the dominance preorder on a convex cone $K$: a point $x$ dominates $y$, written $y \\unlhd_K x$, when $x - \\lambda y \\in K$ for some $\\lambda > 0$. Its equivalence classes are the open components of $K$; each is a relatively algebraically open convex cone, and the partially ordered family $\\mathcal{O}(K)$ is an upper semilattice with join $E_x \\vee E_y = E_{x+y}$. For conical halfspaces the dominance relation is total, so the open components are linearly ordered, and the linear functionals attached to them form a linearly independent cortege. That cortege generates a step-linear function, a real-valued function that selects the least nonzero linear functional in the order at each point, and Theorem 42 shows the asymmetric conical halfspace is exactly the positive set of this function.","core_discovery":"Theorem 49: for any asymmetric convex cone $K$ in a real vector space $X$, the family $\\mathcal{U}_K$ of step-linear functions $u : X \\to \\mathbb{R}$ satisfying $u(x) > 0$ on $K$ and $u(x) = 0$ on $L_K$ is nonempty, and $K = \\{x \\mid u(x) > 0 \\text{ for all } u \\in \\mathcal{U}_K\\}$ while $L_K = \\{x \\mid u(x) = 0 \\text{ for all } u \\in \\mathcal{U}_K\\}$. The proof proceeds by regularly extending $K$ to an asymmetric conical halfspace $H$, representing $H$ by a step-linear function, and then showing that for every point outside $K$ some step-linear functional in $\\mathcal{U}_K$ is nonpositive. The representation is exact: membership in the cone and membership in its associated vector subspace are both certified by the whole family of positive step-linear functionals.","pith_inferences":["If Theorem 49 is correct, the upper semilattice of open components becomes an order-theoretic invariant that could be used to distinguish convex cones that are not linearly isomorphic in infinite-dimensional spaces.","The paper leaves open whether the two elementary properties proved in Proposition 41 (homogeneity and the implications $u(y)=0 \\Rightarrow u(x+y)=u(x)$, $u(x),u(y)>0 \\Rightarrow u(x+y)>0$) actually characterise step-linear functions; a positive answer would yield an axiomatic definition usable in optimization.","The regular-extension argument suggests that the collection of all conical halfspaces extending a given asymmetric cone, together with their step-linear representations, may encode boundary data that a topological dual space cannot see."],"forward_implications":["Every asymmetric convex cone in a real vector space, with no topology, has a nonempty family of positive step-linear functionals, so points outside the cone always admit an analytic certificate of nonmembership.","Asymmetric conical halfspaces and step-linear functions are dual objects: the halfspace is the positive set of the function and the subspace $L_H$ is its zero set.","The open-component semilattice gives a way to study the internal geometry of cones independently of faces, which matters because in infinite-dimensional spaces faces can have empty intrinsic cores while open components always exist.","The family $\\mathcal{U}_K$ can serve as a step-linear counterpart of the positive dual cone, extending the classical duality between algebraically open halfspaces and linear functions."],"supporting_citations":[{"why":"Supplies the separation theorem (Theorem 47) guaranteeing that disjoint convex cones, one asymmetric, are separated by an asymmetric conical halfspace; Theorems 48 and 49 depend on it.","marker":"[11]"},{"why":"Provides the proper-separation result used in Proposition 34 to assign to each open component a nonzero linear functional with $E = \\{x \\in \\mathrm{Lin}(E) \\mid l_E(x) > 0\\}$.","marker":"[1]"},{"why":"Source for the classification of halfspaces in infinite-dimensional spaces and for Lemma 2, which decomposes a conical halfspace as $(-H) \\cup L_H \\cup H$.","marker":"[9, 10]"},{"why":"Introduced step-linear functions, called conditionally linear functions there, together with corteges of linear functions used to build them.","marker":"[8]"},{"why":"Defines corteges of linear functions and step-linear functions and their use in analytically representing halfspaces.","marker":"[10]"},{"why":"Supplies the intrinsic-core and minimal-face machinery used in Theorem 18 and Remark 19 to identify open components with nonempty intrinsic cores of faces.","marker":"[6]"},{"why":"Provides the semispace theory and the linearly ordered families of linear functionals whose axioms the cortege definition extends.","marker":"[20]"}],"fun_headline_variants":["Step-linear family gives exact portrait of any asymmetric cone","Infinite-dimensional cones exactly captured by step-linear functions","No topology needed: step-linear functionals define any asymmetric cone","Every asymmetric cone equals the positive set of its step-linear family","Step-linear functionals certify cone and lineality space exactly"],"cache_read_input_tokens":26752,"weakest_assumption_plain":"The main representation theorem imports, without proof, a separation theorem stating that two disjoint convex cones, one asymmetric, can be separated by an asymmetric conical halfspace; Theorems 48 and 49 collapse if that theorem is false or needs hypotheses not present here.","fun_headline_variants_meta":{"raw":{"variants":["Step-linear family gives exact portrait of any asymmetric cone","Infinite-dimensional cones exactly captured by step-linear functions","No topology needed: step-linear functionals define any asymmetric cone","Every asymmetric cone equals the positive set of its step-linear family","Step-linear functionals certify cone and lineality space exactly"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000652,"raw_usage":{"total_tokens":3019,"prompt_tokens":1002,"completion_tokens":2017,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":618,"completion_tokens_details":{"reasoning_tokens":1936}},"tokens_in":618,"tokens_out":2017,"duration_ms":14999,"temperature":1.0,"reasoning_tokens":1936,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:23:08.670371+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find an asymmetric convex cone $K$ and a vector $y \\notin K$ such that every step-linear function positive on $K$ and zero on $L_K$ is also positive at $y$; Theorem 49 predicts that such a $y$ cannot exist. Equivalently, exhibit disjoint convex cones $K_1$ and $K_2$ with $K_1$ asymmetric that are not separated by any asymmetric conical halfspace, which would disprove the imported Theorem 47.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduced step-linear functions, called conditionally linear functions there, together with corteges of linear functions used to build them."},{"cited_title":"In: Przeworska–Rolewich, D","cited_arxiv_id":null,"evidence_quote":"Defines corteges of linear functions and step-linear functions and their use in analytically representing halfspaces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the semispace theory and the linearly ordered families of linear functionals whose axioms the cortege definition extends."}],"review_version":1}