{"id":"785c83b2-472f-4ae7-90a2-cb97fa66cd3c","arxiv_id":"2505.13871","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper gives a finite graph claimed to separate C^3 from R^3 orthogonality spaces and proves universal embeddability into orthogonality spaces of finite and atomic orthomodular lattices.","lead":"This paper studies graphs of perpendicular lines in quantum spaces, claiming a small graph exists in complex 3D but not real 3D, and proving every finite graph can be embedded in a finite quantum logic's perpendicular-line graph. The complex 3D construction contains a cross product error, so that result is not established as written.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The complex cross-product Lemma 2.2 is false: the defined u×v is not Hermitian-orthogonal to u, so the C^3 subdiagram construction and Theorem 2.6 are unsupported.","rationale":"The reader identified the same load-bearing concern: Lemma 2.2's claim that the un-conjugated cross product is orthogonal in C^3 is false under the Hermitian inner product. I checked the example u=(1,i,0), v=(0,0,1) and found ⟨u×v,u⟩ = 2i ≠ 0, confirming the lemma fails. I also traced the failure into Lemma 2.3, where b1 and c1 are not orthogonal for the MUB vector (1,ω,ω^2) used in Theorem 2.6, and where b12 and b1 fail orthogonality for generic y,z. Since Theorem 2.5 and the C^3 half of Theorem 2.6 rely directly on Lemma 2.3, the paper's first advertised result is unsupported as written. The second theorem (Theorem 3.3 and Corollary 3.8) appears substantially correct, though the compactness step in Corollary 3.8 needs an explicit axiom linking the edge predicate E to orthogonality, e.g. ∀x∀y(E(x,y) ↔ x∧y = 0); without such an axiom the compactness model need not give a strong embedding. This is a smaller, fixable gap. Because the false lemma invalidates a central advertised claim, the REJECT verdict is appropriate; no change to the reader's verdict is needed.","tokens_in":9502,"tokens_out":6563,"duration_ms":62702,"concrete_test":"Take the paper's definitions with x=1, y=1, z=i (or y=ω, z=ω^2 as in Theorem 2.6), and compute the Hermitian inner products ⟨b12,b1⟩ and ⟨c1,b1⟩ using the inner product that appears in the angle formula. If these are nonzero, as the calculation gives ⟨b12,b1⟩ = 2ix and ⟨c1,b1⟩ = 2i for y=1,z=i, then Lemma 2.3's claimed orthogonalities fail and Figure 2 is not a sub-Greechie diagram of P(C^3); this directly invalidates the proof of the C^3 half of Theorem 2.6.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 2.2 is false for C^3 under the Hermitian inner product implied by the angle formula and by the paper's own references to conjugates. For u=(1,i,0) and v=(0,0,1), the defined product gives u×v=(i,-1,0), and ⟨u×v,u⟩ = 2i (up to convention), not 0. Thus the asserted orthogonality of u×v to u fails. Since Lemma 2.3 builds all of Figure 1 from such cross products, the claimed orthogonalities among the listed vectors do not hold in general. For example, with b1=(0,y,z) and b12=(yz,xz,-xy), ⟨b12,b1⟩ = x(z\\bar y - y\\bar z), which vanishes only when y/z is real up to a phase; for y=1,z=i it equals 2ix. Even more directly, c1=(0,z,-y) is not orthogonal to b1=(0,y,z) unless z\\bar y = y\\bar z. For the second MUB center (1,ω,ω^2) used in Theorem 2.6, this fails: ⟨c1,b1⟩ = ω - ω^2 ≠ 0. Consequently Figure 2 is not realized as a sub-Greechie diagram for the vectors specified, and the C^3 half of Theorem 2.6 is not proved. The construction may be repairable by using a conjugate-cross-product definition and adjusted vector lists, but as written the central separating-graph result is unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript contains two independent results. The first constructs a finite Greechie diagram from two copies of a configuration based on mutually unbiased bases and claims it embeds into the orthogonality graph of P(C^3) but not into that of P(R^3); the proof relies on a 'cross product' in C^3 (Definition 2.1) and on Lemmas 2.2 and 2.3 asserting that this product is Hermitian-orthogonal to both factors. The second result shows that every finite graph can be strongly embedded into the orthogonality graph of a finite orthomodular lattice (Theorem 3.3) via a sequence of lemmas using powersets and Kalmbach's coatom extension, and uses compactness to extend this to every graph into an atomic orthomodular lattice (Corollary 3.8).","tokens_in":1464,"tokens_out":2484,"duration_ms":125031,"significance":"The intended first result is a concrete finite combinatorial separation between the orthogonality spaces of complex and real 3-dimensional quantum logics, which would be a valuable addition. The second result provides a finite-dimensional version of the Tau-Tserunyan embedding theorem and appears to be a genuine contribution if the lemmas in Section 3 are correct. The paper is clearly organized and the algebraic constructions with the powerset and the coatom extension are elegant. However, the cross-product lemma that drives the first result is false as stated, and Corollary 3.8 omits a necessary definitional axiom; these issues affect the two central claims as written.","major_comments":[{"comment":"Lemma 2.2 is false for C^3 under the Hermitian inner product used in Definition 2.4. For u=(1,i,0) and v=(0,0,1), the product defined in Definition 2.1 gives u×v=(i,-1,0), and ⟨u×v,u⟩=2i (up to the convention for which argument is conjugated), not 0. Consequently the vector identities in Lemma 2.3 do not produce the orthogonalities claimed in Figure 1, and the reduction in Theorem 2.5 to the three conditions c1⊥b23, c2⊥b13 and c3⊥b12 is invalid. For the two centers used in the proof of Theorem 2.6, namely (1,1,1) and (1,ω,ω^2), the vector c1=(0,z,-y) is not Hermitian-orthogonal to b1=(0,y,z) unless z times conjugate(y) equals y times conjugate(z); for (y,z)=(ω,ω^2) this inner product is ω - ω^2, which is not 0. Thus the asserted realization of Figure 2 as a sub-Greechie diagram of P(C^3) fails for the specified vectors, and Theorem 2.6 is unsupported as written. The construction may be repairable by using a conjugate cross product such as conjugate(u)×v and recomputing the vector table, but the current argument does not go through.","section":"§2, Definition 2.1, Lemma 2.2"},{"comment":"The first-order language introduces a binary predicate E for edges, but the theory Σ does not contain a sentence linking E to the orthogonality relation of the oml, for instance E(x,y) iff x ≤ y′. The sentences E(c_u,c_v) and ¬E(c_u,c_v) constrain E only on the constants naming the vertices. A compactness model can therefore interpret E arbitrarily on other pairs of elements, and even on the constants the connection between E and the orthogonality relation of the oml is not enforced. Without such a definitional axiom, the conclusion that G is strongly embedded into the orthogonality graph of an atomic oml does not follow. Adding the axiom E(x,y) iff x ≤ y′ to Σ repairs the proof, since every finite subset can then be modeled by Theorem 3.3 with E interpreted as orthogonality.","section":"§3, Corollary 3.8"}],"minor_comments":[{"comment":"The sentence 'we will not reproduce this in detail here since it is not as pertinent to our investigation, but only provide a brief a brief sketch' contains a duplicated 'a brief'.","section":"§3, after Theorem 3.1"},{"comment":"The name 'Tao and Tserunyan' should be 'Tau and Tserunyan' to match the reference and the rest of the paper.","section":"§3, Corollary 3.8 proof"},{"comment":"The displayed property (2) is dimensionally inconsistent as printed: the upset of a in M intersected with L is a subset of L, while the upset of e in M is a subset of M. The intended statement is presumably that the intersection equals the upset of e in L, which is the form used in Lemmas 3.6 and 3.7.","section":"§3, Lemma 3.5"},{"comment":"In the sentence 'obtain a finite oml N with M≤N and a∈N\\M such that a is an atom of M', the phrase 'atom of M' should read 'atom of N', since a is explicitly outside M.","section":"§3, Lemma 3.7 proof"},{"comment":"After defining the map h by h(x_i)=a_i, the sentence 'Then f is one-one' should refer to h rather than f.","section":"§3, proof of Theorem 3.3"}],"recommendation":"major_revision","confidential_remarks":"The headline first theorem is currently unsupported because its core orthogonality lemma is false for the Hermitian inner product. The fix is likely local: replacing the cross product with a conjugate version and recomputing the vector table in Lemma 2.3 should recover Theorem 2.6. The gap in Corollary 3.8 is also easily fixed by adding the definitional axiom for E. I therefore recommend major revision rather than rejection, provided the authors can verify the repaired configuration."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The second advertised result—every finite graph strongly embeds into the orthogonality graph of a finite orthomodular lattice—is a genuine, useful theorem, and the proof via Kalmbach coatom extensions is correct in its main line. The first advertised result—a finite Greechie diagram separating P(C^3) from P(R^3)—is not supported by the argument given, because Lemma 2.2 is false for complex vectors under the Hermitian inner product implied by the paper's own angle formula.\n\nFor u=(1,i,0) and v=(0,0,1), the paper's cross product gives u×v=(i,-1,0), and ⟨u×v,u⟩=2i, not 0. So the asserted orthogonality fails. The vector lists in Lemma 2.3 inherit the problem: c1=(0,z,-y) is not orthogonal to b1=(0,y,z) unless ybar z = zbar y, which fails for the mutually unbiased basis vectors used in Theorem 2.6. Consequently Theorems 2.5 and 2.6 are unproven. This is not a nitpick; it is the headline construction. The repair is plausible—redefining the cross product with conjugation and adjusting the vector lists might restore the configuration—but it is not in the paper.\n\nThe second half is better. Theorem 3.3 is new and the construction is clean; it gives a finite oml witness where Tau and Tserunyan only had finite height. Lemma 3.7's induction is a nice piece of work. The compactness step in Corollary 3.8 has a small gap: the first-order language has a predicate E, but the axioms never state that E is the orthogonality relation of the oml (i.e., ∀x,y E(x,y) ↔ x≤y′). Without that, the model's E need not be actual orthogonality, so the embedding is not into the orthogonality graph. That is a one-line fix.\n\nThe writing is clear, the literature is engaged honestly, and the paper does not overclaim the significance of the MUB separation. The current version should not be accepted: a central theorem is unsupported. But the paper is not a waste of time; the second result deserves to be published, and the first is likely salvageable. A serious editor should send it to a referee, and the referee should request major revision rather than a desk reject.","headline":"The finite-oml embedding theorem is solid, but the C^3/R^3 separating-graph result rests on a false complex cross-product lemma and is unproven as written.","tokens_in":10308,"tokens_out":3322,"would_cite":false,"duration_ms":30657,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["06C15","81P05","81P10"],"pacs":[],"model":"deepseek-v4-flash","headline":"A finite orthogonality configuration exists over $\\mathbb{C}^3$ but not $\\mathbb{R}^3$, and every finite graph is realized by a finite orthomodular lattice.","keywords":["orthomodular poset","orthomodular lattice","orthogonality space","mutually unbiased bases","graph embedding","Greechie diagram","subgraph"],"falsifier":"Evaluate the paper's Lemma 2.2 at $\\mathbf{u}=(1,i,0)$ and $\\mathbf{v}=(0,0,1)$: the formula gives $\\mathbf{u}\\times\\mathbf{v}=(i,-1,0)$, and the Hermitian inner product with $\\mathbf{u}$ is $\\langle (i,-1,0),(1,i,0)\\rangle=2i\\ne 0$, so the asserted orthogonality fails as stated. The second theorem, on finite orthomodular lattices, does not use this lemma.","tokens_in":9279,"feed_emoji":"⚛️","tokens_out":11413,"duration_ms":103996,"temperature":0.7,"pith_summary":"Orthogonality spaces record which one-dimensional subspaces of a Hilbert space are perpendicular to which. This paper aims to establish two claims. First, a finite configuration of rays built around mutually unbiased bases can be embedded into the orthogonality space of $\\mathbb{C}^3$ but not into that of $\\mathbb{R}^3$: complex 3-space can hold two rays that are each at equal angle to all three members of a basis and are orthogonal to each other, while real 3-space cannot. Second, every finite graph can be fully embedded into the orthogonality graph of a finite orthomodular lattice, and every graph whatsoever into some atomic orthomodular lattice. These results matter because they show that simple orthogonality data can encode both a coordinate-field distinction and arbitrary finite combinatorics.","feed_headline":"Complex rays host a graph that real rays cannot","feed_subtitle":"A diagram built from mutually unbiased bases separates C^3 from R^3.","key_machinery":"Greechie diagram: a graph determined by its maximal cliques, which for $\\mathcal{P}(\\mathbb{R}^3)$ and $\\mathcal{P}(\\mathbb{C}^3)$ are triples of pairwise orthogonal rays. The load-bearing device for the first result is a cross-product formula used to generate all rays of the center diagram; the diagram occurs exactly when the three inner products $\\langle c_1,b_{23}\\rangle$, $\\langle c_2,b_{13}\\rangle$, and $\\langle c_3,b_{12}\\rangle$ are all zero, which for nonzero $x,y,z$ is equivalent to $|x|=|y|=|z|$. For the second result the machinery is a powerset construction that makes the image of an element lie under the join of a set of images only when the element is in that set, the coatom-extension construction that adds a new atom beneath a chosen nonzero element, and an induction that turns graph adjacency into atom orthogonality; compactness then extends the finite theorem to arbitrary graphs.","core_discovery":"On its own terms, the paper's first claim is that the orthogonality graph of $\\mathcal{P}(\\mathbb{C}^3)$ contains a finite Greechie diagram — two copies of a 'center' configuration glued along their circular rims, with the two centers orthogonal — that the orthogonality graph of $\\mathcal{P}(\\mathbb{R}^3)$ does not contain. The second claim is that for any finite loopless graph $G$ there is a finite orthomodular lattice whose atom orthogonality graph contains $G$ as an induced subgraph, and by a compactness argument every graph occurs as an induced subgraph in some atomic orthomodular lattice. Along the way the paper characterizes the central configuration: a ray $\\langle x,y,z\\rangle$ is a center for the standard basis exactly when $|x|=|y|=|z|$, i.e. exactly when it is unbiased with respect to that basis. This is what lets a pair of mutually unbiased bases in $\\mathbb{C}^3$ produce two orthogonal centers, a configuration impossible over $\\mathbb{R}^3$.","pith_inferences":["Editorial extension: the center/unbiasedness equivalence suggests that similar finite diagrams could serve as coordinate-free graph-theoretic probes for the existence of mutually unbiased bases in other dimensions and fields.","Editorial extension: because the second construction is purely order-theoretic and independent of Hilbert-space coordinates, the same compactness route may realize other first-order configurations of atoms inside atomic orthomodular lattices.","Editorial extension: if the first theorem can be restated without reliance on the complex cross-product lemma, it would supply a small explicit graph separating complex from real dimension-three quantum logics, potentially useful as a dimension witness."],"forward_implications":["The orthogonality space of $\\mathbb{C}^3$ is not graph-theoretically identical to that of $\\mathbb{R}^3$: a single finite, coordinate-free configuration witnesses the difference.","Unbiasedness with respect to an orthonormal basis is detectable from orthogonality alone, since it is equivalent to the occurrence of the center diagram.","Every finite graph is realizable as the orthogonality graph of atoms of a finite orthomodular lattice, so finite quantum-logical structures are universal for finite graphs.","Every graph, including uncountable ones, appears as an induced subgraph of the orthogonality graph of some atomic orthomodular lattice.","The finite-graph realization is produced by an explicit induction, so for any finite graph one can in principle write down a finite orthomodular lattice that realizes it."],"supporting_citations":[{"why":"Supplies the Greechie-diagram framework, the block-structure facts for orthogonality spaces of projection lattices, and the coatom-extension construction used in the second theorem.","marker":"[3]"},{"why":"Provides the detailed Greechie 'paste job' by which the coatom extension is assembled into a finite orthomodular lattice.","marker":"[4]"},{"why":"Gives the Hilbert-space induced-subgraph theorem that the paper's second result parallels and extends, along with the finite-height embedding used in the compactness corollary.","marker":"[6]"}],"fun_headline_variants":["Complex rays host a graph real rays cannot","MUB graph exists only for complex rays","Finite lattices host any finite graph","MUBs build a complex-only orthogonality graph","Every finite graph fits a finite orthomodular lattice"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The first result's load-bearing premise is that the paper's complex cross-product formula always produces a nonzero vector perpendicular to both input vectors; the center characterization and the $\\mathbb{C}^3$-vs-$\\mathbb{R}^3$ separation depend on that single claim.","fun_headline_variants_meta":{"raw":{"variants":["Complex rays host a graph real rays cannot","MUB graph exists only for complex rays","Finite lattices host any finite graph","MUBs build a complex-only orthogonality graph","Every finite graph fits a finite orthomodular lattice"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001077,"raw_usage":{"total_tokens":4463,"prompt_tokens":855,"completion_tokens":3608,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":471,"completion_tokens_details":{"reasoning_tokens":3537}},"tokens_in":471,"tokens_out":3608,"duration_ms":24944,"temperature":1.0,"reasoning_tokens":3537,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:09:51.261828+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the paper's Lemma 2.2 at $\\mathbf{u}=(1,i,0)$ and $\\mathbf{v}=(0,0,1)$: the formula gives $\\mathbf{u}\\times\\mathbf{v}=(i,-1,0)$, and the Hermitian inner product with $\\mathbf{u}$ is $\\langle (i,-1,0),(1,i,0)\\rangle=2i\\ne 0$, so the asserted orthogonality fails as stated. The second theorem, on finite orthomodular lattices, does not use this lemma.","supporting_citations":[{"cited_title":"Kalmbach, Orthomodular lattices , London Math","cited_arxiv_id":null,"evidence_quote":"Supplies the Greechie-diagram framework, the block-structure facts for orthogonality spaces of projection lattices, and the coatom-extension construction used in the second theorem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the detailed Greechie 'paste job' by which the coatom extension is assembled into a finite orthomodular lattice."},{"cited_title":"Bure s ov\\' a , K","cited_arxiv_id":null,"evidence_quote":"Gives the Hilbert-space induced-subgraph theorem that the paper's second result parallels and extends, along with the finite-height embedding used in the compactness corollary."}],"review_version":1}