{"id":"d58c18b9-399a-4ad5-8e05-59ab53f64522","arxiv_id":"1908.03865","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Using a new geometric property, the author proves that the five known types of triples of pairwise disjoint triangles in 3-space are pairwise not combinatorially isotopic.","lead":"This paper gives a short proof that certain triples of disjoint triangles in 3D space, such as the Borromean (Valknut) triple, cannot be continuously deformed into a completely separate triple without the triangles crossing. It introduces an easy-to-check geometric invariant based on how the triangles' filled hulls intersect.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4's proof that Borromeanness is invariant under elementary moves contains unjustified positional assertions; the separation of the first two types in Proposition 3 depends on it.","rationale":"The reader correctly identifies the invariance of the Borromean property as the load-bearing part of Proposition 3 and gives a CONDITIONAL verdict. I agree that Lemma 4 is under-justified and that the paper needs revision. However, the reader's stated weakest assumption is the implication in Lemma 5, and that implication is actually true under the disjointness of triangle outlines: for two planar triangles whose convex hulls intersect in a segment, if ∂Λ1 meets ⟨Λ2⟩ in two points, then ∂Λ2 cannot meet ⟨Λ1⟩ without forcing the outlines to intersect. The more serious concern is the chain of unproved positional claims inside Lemma 4, especially the assertions about the swept tetrahedron τ and the transition from |∂∆2∩⟨∆0⟩|=2 to |∂∆2∩⟨∆0'⟩|=2. These claims are load-bearing because they establish that the Borromean property is invariant, and without them Proposition 3 does not separate the Borromean type from the disjoint-hull type. A computational exact-arithmetic check on the standard Borromean configuration and generic elementary moves would either expose a counterexample or confirm that the gap is purely expository. For that reason the reader's CONDITIONAL verdict should remain unchanged: the result is plausible and likely known, but the proof as written does not yet fully support the central claim.","tokens_in":3779,"tokens_out":30555,"duration_ms":345405,"concrete_test":"Use exact rational arithmetic to check Lemma 4 on the explicit Borromean triple of Conjecture 2. For a finite family of valid elementary moves (side faces ACC′ and BCC′ disjoint from the other two triangle outlines), compute before and after the move: ∂∆1∩⟨∆0'⟩, ∂∆2∩⟨∆0'⟩, and ⟨∆0'⟩∩⟨∆1⟩∩⟨∆2⟩. Also sweep C′ continuously from C to its final position and record every combinatorial event. If any event gives ∂∆1∩⟨∆0'⟩≠∅ or a nonempty triple intersection before the move but empty after, while the side faces avoid the other outlines, Lemma 4 is false. If all enumerated events agree with Lemma 4, the concern is reduced to a rigor gap, not a correctness failure.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim Proposition 3 distinguishes the pairwise-disjoint-hull type from the Borromean type by the Borromean property, whose invariance under elementary moves is exactly Lemma 4. The proof of Lemma 4 is not rigorous. To show ∂∆1∩⟨∆0'⟩=∅, it asserts 'The outline ∂∆1 may only intersect the face ABC′ of τ' and then infers from the geometry of d1 and ˆd1 that ∆'1 is in the interior of ⟨∆1⟩; the inference is not justified in the text. The subsequent proof that |∂∆2∩⟨∆0'⟩|=2 is even more compressed: it states that whenever ∂∆2∩⟨∆0⟩ is not two points, '∆P2 is entirely in the interior of τ', contradicting the existence of P. No argument is supplied, and the case where τ is degenerate (C′ lies in the plane of ABC) is not treated, although a combinatorial isotopy may pass through such degenerate tetrahedra. If any of these assertions fails, Lemma 4 collapses and Proposition 3's distinction between the first two bullet points is unsupported. I do not think the reader's specific complaint about Lemma 5's implication |∂Λ1∩⟨Λ2⟩|=2 ⇒ ∂Λ2∩⟨Λ1⟩=∅ is the right worry: for two disjoint planar triangle outlines whose convex hulls intersect in a segment, that implication does follow from convexity and disjointness. The real gap is in the rest of Lemma 4.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines combinatorial isotopy for triples of pairwise disjoint triangles in 3-space, generated by elementary moves that replace one vertex of a triangle while keeping certain auxiliary convex hulls disjoint from the other triangle outlines. It conjectures that every such triple is combinatorially isotopic to one of five listed types, and it proves Proposition 3, the main result, that triples from different bullet points are not combinatorially isotopic. The proof introduces two invariants: the pairwise mod-2 linking numbers and a Borromean property defined by convex-hull intersections. Lemma 5 gives an equivalent characterization of the Borromean property, and Lemma 4 asserts that this property is preserved by elementary moves. The paper is an elementary alternative to an earlier proof using the Massey-Rolfsen invariant.","tokens_in":4082,"tokens_out":11651,"duration_ms":124280,"significance":"If the proof can be made rigorous, the paper offers a genuinely elementary and easy-to-compute invariant distinguishing the Borromean triple of triangles from a triple with pairwise disjoint convex hulls, and it supports a larger conjectural classification. The strength of the paper is its explicit, parameter-free geometric definitions and its connection to the classical Borromean rings and to Milnor's link-homotopy classification. However, the central invariance proof is currently presented as a sketch with several unsupported geometric assertions, so the significance is conditional on completing those arguments.","major_comments":[{"comment":"The assertion that 'in all cases where the intersection ∂∆2∩⟨∆0⟩ does not consist of two points ..., ∆P2 is entirely in the interior of τ' is stated without proof. This is load-bearing because it is the only argument establishing |∂∆2∩⟨∆0'⟩|=2, one of the properties required by Lemma 5. A detailed geometric proof is needed, including a treatment of the degenerate case where τ is not a nondegenerate tetrahedron, since an elementary move may pass through such configurations.","section":"Lemma 4, proof, second paragraph"},{"comment":"The inference 'The segment PQ intersects d̂1∩⟨∆0⟩, hence PQ∩d̂1⊂⟨∆1⟩. This means that d̂1 is in the interior of the triangle ⟨∆1⟩' is not justified. A single intersection point of PQ with a side of ∆1' does not by itself imply that the entire side d̂1 lies in the interior of ⟨∆1⟩; a convexity or separation argument is missing. The subsequent conclusion that 'the polygon ∆1' is in the interior of ⟨∆1⟩' is even stronger and is not derived from the preceding text.","section":"Lemma 4, proof, first paragraph"},{"comment":"The sentence 'Then ∂∆2 intersects two sides of ∆2' at two points each' is ambiguous and unproved, and it is unclear how the following conclusions about the faces ⟨ACC'⟩ and ⟨BCC'⟩ not intersecting ⟨∆2⟩ follow. This paragraph is needed to prove ⟨∆0'⟩∩⟨∆1⟩∩⟨∆2⟩≠∅, the remaining Lemma 5 property. As it stands, the proof does not give a verifiable chain of implications.","section":"Lemma 4, proof, third paragraph"},{"comment":"The step 'Since |∂Λ1∩⟨Λ2⟩|=2, we have ∂Λ2∩⟨Λ1⟩=∅' is asserted without proof. Under a general-position assumption on the two triangle planes, the implication can be justified by a convexity argument, but the paper does not state such an assumption, and the discussion of elementary moves can involve degenerate positions. If this implication fails, the case analysis in Lemma 5 collapses, so this missing justification affects the equivalence that underlies Lemma 4.","section":"Lemma 5, proof"}],"minor_comments":[{"comment":"The abstract in the arXiv posting says 'Two triples of triangles having pairwise disjoint outlines', while the full text says 'Two triples of pairwise disjoint triangles'. Since 'triangle' is defined as the outline in Section 1, the wording should be made consistent to avoid ambiguity.","section":"Abstract"},{"comment":"The definition of an elementary move requires (⟨ACC′⟩∪⟨BCC′⟩)∩XYZ=∅ for every other triangle XYZ, but it does not explicitly state that the new triangle ABC′ itself is disjoint from the other triangle outlines. The preservation of the disjointness condition should be stated as part of the definition or proved immediately.","section":"Section 1, definition of elementary move"},{"comment":"The notation ∆P2 is introduced as 'the part of ∆2 which is on the same side of the plane containing ∆0 as C′', but this is never made precise. It should be defined as the intersection of the filled triangle ⟨∆2⟩ with the corresponding open half-space, or with the closed half-space, and the distinction matters for the subsequent argument.","section":"Section 2, Lemma 4 proof"},{"comment":"In the coordinates for the Borromean linking, the expressions (−2√5,±1√10,0) should be typeset as ordered pairs with clear delimiters, for example (−2/√5, ±1/√10, 0), to avoid confusion about which coordinate is affected by the sign.","section":"Conjecture 2"},{"comment":"The proof of Lemma 4 relies on Figures 1 and 2, but the captions are not self-contained and the figures are not reproduced in the text. The geometric assertions that refer to the figures should be fully verbalized, since the journal version may not preserve the informal drawings.","section":"Figures"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a short research announcement whose main theorem is plausible but not yet rigorously established. The proof of Lemma 4 is more a sketch than a complete proof, and the gaps are in load-bearing places. I would encourage the editor to seek a revision in which the author supplies complete proofs of the geometric assertions in Lemma 4 and Lemma 5, or explicitly states and proves the needed general-position lemmas. If the gaps cannot be filled, the paper should be reframed as a conjecture-plus-evidence note rather than a proof of Proposition 3."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's my read of Kogan's 'Linking of three triangles in 3-space.' The paper proposes a new, easily computable invariant — the 'Borromean property' — to distinguish the first two types in Conjecture 2, and claims an elementary proof of Proposition 3 that the five types are pairwise non-isotopic. The idea is genuinely appealing: instead of the Massey-Rolfsen invariant, you just count intersection points between triangle outlines and convex hulls. The mod-2 linking coefficient part is solid, and the five-type classification is a nice conjecture worth thinking about.\n\nBut the core proof doesn't yet hold up. Lemma 4, which asserts that the Borromean property is invariant under elementary moves, is the load-bearing wall, and it's built on several geometric assertions that are neither proven nor fully justified. For example, the claim that ∂∆1 may only intersect the face ABC′ of τ, and the later claim that when |∂∆2∩⟨∆0⟩| ≠ 2 the piece ∆P2 is entirely inside τ, are exactly the kind of positional facts that need careful proof, not a figure. The degenerate case where τ is flat is also not handled. If any of these fails, the distinction between the first two bullet points collapses. The stress-test note suggests the reader's complaint about Lemma 5 might be aimed at the wrong step; I think there's something to that — the implication in Lemma 5 may be recoverable by convexity — but the gaps in Lemma 4 are real regardless.\n\nThat said, this is a short, honest paper. The author explicitly states that the non-isotopy claim is already known from Massey-Rolfsen, and that the new thing is the simpler proof. The writing is clear. It deserves a serious referee, not a desk reject, because the underlying idea could be made rigorous with moderate effort. I'd send it to review with a request that the author supply a fully detailed proof of Lemma 4 and address degenerate configurations.\n\nFor a reading group, I'd say maybe — the conjecture and invariant are worth discussing, but the current proof would require a lot of filling in. I wouldn't cite it in its present form, but I'd keep an eye on a revised version.","headline":"A promising geometric invariant for distinguishing Borromean triangle linkings, but the invariance proof is too hand-wavy to carry the main claim.","tokens_in":4568,"tokens_out":3130,"would_cite":false,"duration_ms":30385,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57K10"],"pacs":[],"model":"deepseek-v4-flash","headline":"A pair of easy invariants — the three pairwise mod-2 linking numbers and a Borromean convex-hull condition — separates the five types of triples of disjoint triangles in 3-space.","keywords":["combinatorial isotopy","triangle linking","Borromean rings","Valknut","convex hull","linking number mod 2","elementary moves","3-space linkings"],"falsifier":"Construct three pairwise disjoint triangles whose convex hulls share a point, with |∂Λ1∩⟨Λ2⟩|=2 and |∂Λ2∩⟨Λ0⟩|=2 and ∂Λ1∩⟨Λ0⟩=∅, but where ∂Λ0 meets ⟨Λ1⟩ in exactly one point instead of two; such an example would contradict Lemma 5 and would leave the invariance of the Borromean property unproved.","tokens_in":3567,"feed_emoji":"🔺","tokens_out":5540,"duration_ms":57331,"temperature":0.7,"pith_summary":"The paper proves that five specific triples of pairwise disjoint triangles in 3-space are pairwise not combinatorially isotopic: a triple with pairwise disjoint convex hulls, the Borromean (Valknut) triple, one Hopf-linked pair plus an isolated triangle, a triangle stabbed by two disjoint hulls, and a rotated-translated equilateral configuration. This is Proposition 3, the main result. The point is to show that the classification problem for triangle linkings has at least five distinct classes, and that distinguishing them requires only easy-to-compute invariants rather than the Massey-Rolfsen invariant used in earlier work. The completeness side — that every unordered triple belongs to one of these five types — remains a conjecture.","feed_headline":"Five configurations of three disjoint triangles are inequivalent","feed_subtitle":"Mod-2 linking numbers plus a convex-hull intersection check separate all five types.","key_machinery":"The machinery is combinatorial isotopy generated by elementary moves: replace one triangle ABC by ABC' when the two new triangular faces do not meet any other triangle. The invariants are the pairwise linking coefficient modulo 2, which separates three of the five classes, and the Borromean property, defined as the convex hulls of all three triangles having a common point while, after a cyclic enumeration, each triangle outline intersects the next convex hull in exactly two points. Lemma 5 is the load-bearing reformulation: it rewrites the Borromean property in terms of three outline-hull intersections and one empty intersection, and Lemma 4 then uses this form to show the property survives elementary moves.","core_discovery":"The paper's central claim is that no sequence of elementary moves can connect triples from different bullet points of Conjecture 2. The proof assigns to each triple the multiset of its three pairwise mod-2 linking numbers and, for the one ambiguous pair in which all three numbers vanish, checks whether the triple is Borromean: the convex hulls of the three triangles share a common point and, after a suitable cyclic enumeration, each triangle's outline meets the next triangle's convex hull in exactly two points. Lemma 4 states that the Borromean property is invariant under elementary moves, and Lemma 5 gives an equivalent formulation used in the proof. Since the five listed triples differ in these invariants, they occupy distinct combinatorial isotopy classes.","pith_inferences":["The same convex-hull intersection counting could be tested on four or more triangles; if the Borromean condition generalizes, it might produce a hierarchy of higher-order linking invariants.","If the completeness half of Conjecture 2 is eventually proved, the five types would give a complete enumeration of combinatorial isotopy classes, making the classification algorithmically checkable in the sense the author conjectures.","The unproved geometric implication in Lemma 5 is the place to probe first: if it admits a counterexample, the Borromean property may fail to be invariant under elementary moves, and a different convex-hull condition would be needed."],"forward_implications":["The five classes in Conjecture 2 are pairwise inequivalent, so any complete classification of triples of disjoint triangles must contain at least these five combinatorial isotopy classes.","All pairwise linking numbers zero does not mean the triple is trivial: the Borromean triple and the triple with pairwise disjoint convex hulls are both pairwise unlinked mod 2 yet are not combinatorially isotopic.","The Borromean property supplies a computable certificate that a given triple cannot be unlinked by elementary moves; checking it requires only counting outline-hull intersections.","Because the invariants are mod-2 counts, the same proof shows the five types remain distinct whether the triples are treated as ordered or unordered and whether vertices are labelled."],"supporting_citations":[{"why":"Supplies the historical Valknut example and the medieval origin of linked triangle motifs.","marker":"[Va]"},{"why":"Supplies the Borromean/Valknut linking used as the second bullet point of Conjecture 2.","marker":"[Bo]"},{"why":"Defines combinatorial isotopy via elementary moves, the mod-2 linking coefficient for triangle outlines, and the triple-linking context.","marker":"[Sk]"},{"why":"Gives Milnor's link-homotopy classification to which Conjecture 2 is compared.","marker":"[Mi54]"}],"fun_headline_variants":["Five types of disjoint triangle triples proven inequivalent","No motion links these 5 triangle triples in 3-space","Distinct invariants separate five disjoint triangle triples","Linking numbers plus Borromean test separate five triples"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of invariance depends on the unproved geometric assertion in Lemma 5 that if one outline meets the next triangle's convex hull in two points, then the third outline misses that convex hull entirely; the author states this without derivation, and disjointness of the outlines alone does not imply it.","fun_headline_variants_meta":{"raw":{"variants":["Five types of disjoint triangle triples proven inequivalent","No motion links these 5 triangle triples in 3-space","Distinct invariants separate five disjoint triangle triples","Linking numbers plus Borromean test separate five triples"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00053,"raw_usage":{"total_tokens":2469,"prompt_tokens":778,"completion_tokens":1691,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":394,"completion_tokens_details":{"reasoning_tokens":1623}},"tokens_in":394,"tokens_out":1691,"duration_ms":12523,"temperature":1.0,"reasoning_tokens":1623,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:59:37.101047+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct three pairwise disjoint triangles whose convex hulls share a point, with |∂Λ1∩⟨Λ2⟩|=2 and |∂Λ2∩⟨Λ0⟩|=2 and ∂Λ1∩⟨Λ0⟩=∅, but where ∂Λ0 meets ⟨Λ1⟩ in exactly one point instead of two; such an example would contradict Lemma 5 and would leave the invariance of the Borromean property unproved.","supporting_citations":[],"review_version":1}