{"id":"999f8a41-4fa9-4b8d-8d75-cf6a8c03f6fc","arxiv_id":"2411.14379","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For real singular cubic threefolds with no real singular points, rationality is determined by singularity type and by the existence of real lines, planes, or cubic scrolls, with cohomological obstructions in four exceptional cases.","lead":"This paper classifies which real cubic threefolds with singularities are rational, meaning secretly reparametrizable as ordinary projective 3-space. It gives exact criteria involving real lines, planes, and cubic scrolls, and finds cases where a cohomological obstruction blocks rationality.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.1 asserts without proof that a P1-bundle over P2 obtained from a real normal cubic scroll admits a section over R; if this fails, the 6A1 rationality criterion collapses.","rationale":"I agree with the reader that the paper is a careful and largely credible classification note, and that the central claim is likely correct. The reader's CONDITIONAL verdict is justified. Among the candidate weak points, I identify the unproved section assertion in Proposition 5.1 as the most load-bearing, because a failure there would directly invalidate the rationality criterion in the 6A1 case, one of the main classification statements. The reader also flagged this as a second fragile point; I agree with that assessment and rank it above the general reliance on [CTZ24]/[CMTZ24], which is a normal and acceptable form of mathematical credit. The Proposition 6.1 misstatement is real and should be corrected, but the proof in Section 6 recovers the intended content, so it is less damaging. The paper deserves credit for explicit normal forms, concrete examples, and a substantial real-form analysis; no machine-checked proof is present, but the parameter-free derivations and explicit examples are meaningful evidence. My recommendation is UNCHANGED: the reader's CONDITIONAL verdict is the right one, and the concerns I raise are reasons to require a fix, not reasons to reject the paper outright.","tokens_in":19567,"tokens_out":8442,"duration_ms":87749,"concrete_test":"Take the explicit 6A1 cubic with a real normal cubic scroll given in Remark 5.2, form the P1-bundle over P2 arising from |O_X(2)-S|, and determine whether its generic fiber is a split conic over R(x1,x2,x3). Concretely, write the relative conic explicitly and compute whether the associated quadratic form represents zero over the real function field, or compute its Hasse invariant. If the generic fiber has no real point, Proposition 5.1's assertion that the bundle admits a section over R is false, and the rationality claim for this case is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest single point of exposure is Proposition 5.1. The paper's classification says that a real cubic with 6A1 singularities in linearly general position is rational over R if and only if it contains a real normal rational cubic scroll. The 'if' direction is proved by passing to a P1-bundle over P2 via |O_X(2)-S| and then asserting: 'Such a P1-bundle over R admits a section and is therefore rational.' This is not proved, and it is not an automatic fact. A P1-bundle over P2_R need not be the projectivization of a split rank-2 bundle; its Brauer class can be a nontrivial quaternion algebra over R(P2), and a real section exists exactly when that algebra is split. If the generic fiber has no real point, the P1-bundle has no real section, so the rationality conclusion does not follow from the bundle structure alone. The same assertion is used in [CTZ24, Section 7] in the equivariant setting, but over R the Serre obstruction is not automatically zero. Without this section, the 'if' direction of the 6A1 criterion is unsupported, and the headline 'rational iff contains a real normal cubic scroll' would fail in that direction. A secondary issue, also worth fixing, is Proposition 6.1: its statement 'rational iff it satisfies (H1)' is inconsistent with (1.1), since satisfying (H1) means nonvanishing H1 and hence non-stable-rationality. The proof indicates the intended statement is about 'only one real plane' and H1 != 0, but the proposition as written is false and is cited in Proposition 2.1.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the (stable) rationality of real singular cubic threefolds X ⊂ P^4 that are not cones and have no real singular points. Working over R with Γ = Gal(C/R), the authors combine the equivariant classification of [CTZ24] and [CMTZ24] with two standard obstructions: disconnectedness of X(R) and non-vanishing of H^1(Γ, Pic(X~_C)). The main output is a classification, summarized in the introduction: for each possible geometric singularity type of X_C (2, 4, 6, 8 isolated singularities, plus non-isolated cases), the paper gives criteria for rationality and stable rationality in terms of real linear-algebraic data such as existence of real lines disjoint from certain planes, real planes, or real normal cubic scrolls. The paper also provides explicit normal forms and examples realizing connected and disconnected real loci, and leaves open the rationality of cubics singular along a conic with connected real locus.","tokens_in":19809,"tokens_out":15550,"duration_ms":125890,"significance":"If correct, the classification would provide a complete rationality picture for a natural class of rationally connected threefolds over R, complementing recent results on intersections of two quadrics. The paper contains a number of explicit, checkable normal forms and concrete examples (e.g., Examples 2.2–2.5, Remark 5.2, Section 7), and it sharply contrasts the real Galois action with the equivariant setting, which is a useful conceptual contribution. However, the central classification relies at several load-bearing points on unproved assertions (the P1-bundle section in Proposition 5.1, the disjointness of a line and a plane in the 2D4+2A1 case) and on an internally inconsistent statement in Proposition 6.1. The dependence on the preprints [CTZ24] and [CMTZ24] for the completeness of the H1-obstruction should also be made explicit. These issues are local and repairable, so the manuscript is promising but not yet ready.","major_comments":[{"comment":"The proof asserts that the P1-bundle over P2 obtained from the linear system |O_X(2)-S| 'admits a section over R and is therefore rational', with no justification. Over R, a P1-bundle over P2 is a conic bundle whose Brauer class is a quaternion algebra over R(P2); a section exists exactly when that algebra is split. The existence of a real normal cubic scroll S does not by itself guarantee such a section. This step is load-bearing for the 'if' direction of the 6A1 rationality criterion, so it must be proved or replaced by a reference that handles the real case.","section":"§5, Proposition 5.1"},{"comment":"The statement 'X is (stably) rational over R if and only if it satisfies (H1) if and only if X contains three planes over R' is internally inconsistent: by (1.1), satisfying (H1) means H^1(Γ, Pic(X~_C)) ≠ 0, which obstructs stable rationality and therefore cannot be equivalent to rationality. The proof's three cases show that rationality occurs exactly in the cases with three real planes, while non-stable-rationality occurs in the case with only one real plane; the proposition should be restated accordingly. This error is cited in Proposition 2.1(4) and must be corrected.","section":"§6, Proposition 6.1"},{"comment":"The claim that 'the line passing through p3 and p4 is disjoint from Π3 and Π5' is false in the normal form (4.5): the line L = {x1 = x2 = x5 = 0} intersects Π3 = {x3 = x5 = 0} at [0:0:0:1:0], and it also meets Π4 and Π5. Since the rationality argument for the three-plane case of 2D4+2A1 in Proposition 4.5 depends on the existence of a real line disjoint from a real plane, the proof as written is invalid. The authors should either correct this geometric assertion or supply an alternative rationality argument.","section":"§4, paragraph before Proposition 4.5 and Proposition 4.5"},{"comment":"The 'only if' direction of Proposition 2.1 is not proved in the paper; it is delegated to [CTZ24] and [CMTZ24] by the sentence 'In all other cases, the (H1)-obstruction is trivial, as shown in...'. Since these cited works are preprints by overlapping authors and since the completeness of the equivariant classification is exactly what is needed for the equivalence, the classification as presented is conditional on those preprints. The authors should state this dependence explicitly in the introduction and, ideally, include a proof or a more detailed case-by-case verification of the vanishing of H1 in the configurations not listed.","section":"§2, Proposition 2.1"}],"minor_comments":[{"comment":"The condition for 4A2 singularities appears misprinted: 'b1 = b2 = b3 = b4 − (t1−t2)^2/8' should presumably read 'b1 = b2 = b3 = b4 = −(t1−t2)^2/8'.","section":"§4, after (4.2)"},{"comment":"The phrase 'where the Galois action fails (H1)' is ambiguous; it seems to mean that the (H1)-obstruction is nonzero, so the cubic is not stably rational. Please rephrase for clarity.","section":"Remark 5.2"},{"comment":"The description of the image π(X(R)) uses a square root inside an inequality; indicating the affine chart and whether the region is bounded would improve readability.","section":"Example 2.2"},{"comment":"The proof mentions a 'trivial double cover' without defining it; a brief explanation would help the reader understand the application of [BW20] or [KP24a, Theorem 6.10].","section":"Proposition 3.1"}],"recommendation":"major_revision","confidential_remarks":"The paper's main classification is plausible but depends heavily on two closely related preprints by the same research group. The referee recommends sending the revised manuscript back to a referee with expertise in real birational geometry, and possibly asking the authors to include the missing cohomology computations or to clearly mark the conditional status. The current presentation is otherwise careful and contains many useful explicit examples."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me give you my take on this paper. The new content is real: for real singular cubic threefolds with no real singular points, it gives a systematic rationality classification, with explicit normal forms, real-line/plane/scroll criteria, and concrete examples of connected and disconnected real loci. The four cohomological obstructions are clearly identified, and the central classification is likely correct. This is a useful contribution to birational geometry over R.\n\nThe soft spots are two. First, Proposition 5.1 asserts without proof that the P1-bundle over P2 obtained from a real normal cubic scroll admits a section over R. That is not automatic: a P1-bundle over P2_R need not be split, and a real section exists exactly when the Brauer class is trivial. The paper needs to show this for the specific bundle, or the 'if' direction of the 6A1 criterion is unsupported. Second, Proposition 6.1 states that X is rational iff it satisfies (H1), which contradicts the paper's own obstruction (1.1). The proof shows the intended statement: rational iff it contains three real planes, and only one real plane gives nonvanishing H1. This typo appears in a result cited by Proposition 2.1 and must be fixed. Also, the paper leans heavily on [CTZ24] and [CMTZ24] as black boxes, especially for the 'only if' direction of Proposition 2.1; that is a citation concern but not circular, since the real-form analysis is original.\n\nWho is this for: people working on birational geometry over nonclosed fields and real algebraic geometry. It deserves a serious referee, and after the two fixes I'd expect it to be published. My recommendation: conditional accept.","headline":"Real-form classification of singular cubic threefolds: solid and useful, but two load-bearing statements need fixing before acceptance.","tokens_in":20419,"tokens_out":5618,"would_cite":true,"duration_ms":49991,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14E08","14J30","14P05"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper gives a nearly complete rationality classification for real singular cubic threefolds, ruled by a cohomological obstruction and by the presence of real lines, planes, and cubic scrolls.","keywords":["rationality","stable rationality","cubic threefold","singular cubic threefolds","real algebraic geometry","cohomological obstruction","quadric surface bundle","cubic scroll"],"falsifier":"Compute $H^1(\\Gamma, \\operatorname{Pic}(\\widetilde{X}_{\\mathbb{C}}))$ for the connected $2D_4$ example displayed after Lemma 3.9 in Section 3; a nonzero value would contradict Proposition 2.1, which places that case outside the list of configurations with nontrivial obstruction.","tokens_in":19283,"feed_emoji":"📐","tokens_out":15683,"duration_ms":122179,"temperature":0.7,"pith_summary":"Over the complex numbers, a singular cubic threefold in projective four-space is rational unless it is a cone over a smooth cubic curve. Over the real numbers the question is more delicate, and this paper gives a classification of rationality and stable rationality for real singular cubic threefolds that is complete except for two explicitly marked open cases, under the standing assumptions that the threefold is not a cone and has no real singular points. The primary tool is a cohomological obstruction: the Galois group of complex conjugation acts on the Picard group of the minimal resolution, and when the first cohomology group of this action is nonzero the threefold is not stably rational over the reals. The paper shows the obstruction vanishes except for four exceptional singularity configurations, and that in the remaining cases rationality is governed by the presence of real lines disjoint from planes, real planes, or real normal cubic scrolls. The two open cases are cubics with certain $2D_4$ singularities and connected real locus, and cubics whose singular locus is a conic.","feed_headline":"Real cubic threefolds: planes, lines, scrolls decide rationality","feed_subtitle":"For each listed singularity type, rationality over R comes down to a real line, plane, or cubic scroll.","key_machinery":"The load-bearing object is the first cohomology group $H^1(\\Gamma, \\operatorname{Pic}(\\widetilde{X}_{\\mathbb{C}}))$ of the Galois group $\\Gamma=\\operatorname{Gal}(\\mathbb{C}/\\mathbb{R})$ acting on the Picard group of the minimal resolution $\\widetilde{X}$ of $X$; nonvanishing is the $(H^1)$-obstruction and implies that $X$ is not stably rational over $\\mathbb{R}$. The complementary birational machinery consists of projection from a real line together with a disjoint real plane (giving a birational map to $\\mathbb{P}^2 \\times \\mathbb{P}^1$), unprojection from a real plane (reducing rationality to the existence of real lines on smooth intersections of two quadrics), and the quadric-surface-bundle structures obtained by projecting from real planes. For the six-node case the crucial subvariety is a normal rational cubic scroll, a degree-3 rational surface in $\\mathbb{P}^4$ defined over $\\mathbb{R}$, because a real scroll makes $X$ birational to a $\\mathbb{P}^1$-bundle over $\\mathbb{P}^2$, which is rational over $\\mathbb{R}$ once it has a section.","core_discovery":"The central claim is that stable rationality of a real singular cubic threefold $X \\subset \\mathbb{P}^4$, when $X$ is not a cone and has no real singular points, is governed by the action of complex conjugation on the Picard group of the minimal resolution. Proposition 2.1 asserts that $H^1(\\Gamma, \\operatorname{Pic}(\\widetilde{X}_{\\mathbb{C}})) \\neq 0$ holds exactly for four types: $2A_5$; $2D_4+2A_1$ with only one real plane; $6A_1$ in linearly general position with no real normal cubic scroll; and $8A_1$ with only one real plane. In each of these cases the threefold is not stably rational over $\\mathbb{R}$, and in all other cases the obstruction vanishes. The paper then proves rationality by explicit birational constructions whenever the relevant real object exists, a real line disjoint from a plane, a real plane, or a real normal cubic scroll, yielding criteria such as: $4A_1$ with a plane is rational exactly when it contains a real line disjoint from the plane; $6A_1$ with no plane is rational exactly when it contains a real cubic scroll; and $8A_1$ is rational exactly when it contains three real planes.","pith_inferences":["The four configurations with nonzero $H^1$ share a pattern: complex conjugation swaps pairs of divisor classes on the minimal resolution, and the same cohomological test could be applied to other real Fano threefolds with an equivariant Picard group of this form, though the paper does not make that comparison.","The paper's specialization diagram runs only in one direction over $\\mathbb{R}$; a natural extension is to construct real families with rational general fiber and non-stably-rational special fiber and test whether the obstruction jumps, since the paper notes specialization arguments do work over $\\mathbb{R}(t)$.","For the two open rationality cases, the paper's connectedness criteria come from fibers of quadric-surface bundles, which suggests a testable next step: check whether rationality in those cases is equivalent to some real fiber containing a real line, though the paper does not claim this."],"forward_implications":["A real singular cubic threefold with two $A_5$ singularities and no real singular points is not stably rational over $\\mathbb{R}$, even though its real locus is connected.","For a cubic with six $A_1$ singularities in general position, rationality over $\\mathbb{R}$ is equivalent to containing a real normal rational cubic scroll; without one, the $(H^1)$-obstruction is nonzero.","For eight $A_1$ singularities, rationality over $\\mathbb{R}$ is equivalent to containing three real planes; the case with a single real plane is not stably rational.","For several configurations, rationality reduces to the existence of a real line disjoint from a real plane, via birational equivalence to a smooth intersection of two quadrics in $\\mathbb{P}^5$.","For cubics whose complex singular locus is non-isolated, the paper shows all real forms are rational over $\\mathbb{R}$ unless the singular locus is a conic; for the conic case it gives criteria for connectedness and disconnectedness of $X(\\mathbb{R})$ and leaves rationality open."],"supporting_citations":[{"why":"Supplies the equivariant classification of singularity configurations and class-group actions that Proposition 2.1 uses to assert triviality of the obstruction in all unlisted cases.","marker":"[CTZ24]"},{"why":"Provides the specific class-group computations, for example the $C_2$-action for $2A_5$, that yield $H^1=\\mathbb{Z}/2$ in the exceptional cases.","marker":"[CMTZ24]"},{"why":"Gives the criterion that a smooth intersection of two quadrics over $\\mathbb{R}$ is rational exactly when it contains a real line, which drives the plane-plus-line rationality reductions.","marker":"[HT21]"},{"why":"Proves non-rationality of real conic bundles with smooth quartic discriminant, establishing the non-rationality of the $2A_1$ and $2A_2$ cases.","marker":"[BW20]"},{"why":"Used for the identity relating $H^1(\\Gamma,\\operatorname{Pic}(\\widetilde{X}_{\\mathbb{C}}))$ to the Brauer group and for a connected example of a cubic singular along a conic.","marker":"[CTP24]"},{"why":"Classifies the possible non-isolated singular loci of cubic threefolds (plane, line, conic, twisted quartic), setting up Section 7.","marker":"[Yok02]"},{"why":"Also used for the classification of non-isolated singular cubic threefolds and their moduli context.","marker":"[All03]"},{"why":"Shows that any real form of the Segre cubic (ten isolated singularities) is rational, so the paper may restrict to at most eight isolated singular points.","marker":"[Avi20]"}],"fun_headline_variants":["Rationality of real cubic threefolds: when a real line or scroll exists","Stable rationality over R: only four singular real cubic types fail","Real cubic threefolds: rationality iff real plane, line, or scroll found","Conjugation on Picard group decides rational real cubic threefolds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof leans on the completeness of the classification of singularity configurations and Galois actions imported from the two earlier papers; if that classification omits a configuration, the asserted equivalence in Proposition 2.1, and with it several non-rationality conclusions, would fail.","fun_headline_variants_meta":{"raw":{"variants":["Rationality of real cubic threefolds: when a real line or scroll exists","Stable rationality over R: only four singular real cubic types fail","Real cubic threefolds: rationality iff real plane, line, or scroll found","Conjugation on Picard group decides rational real cubic threefolds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000417,"raw_usage":{"total_tokens":2076,"prompt_tokens":800,"completion_tokens":1276,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":416,"completion_tokens_details":{"reasoning_tokens":1195}},"tokens_in":416,"tokens_out":1276,"duration_ms":9765,"temperature":1.0,"reasoning_tokens":1195,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:14:45.191300+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $H^1(\\Gamma, \\operatorname{Pic}(\\widetilde{X}_{\\mathbb{C}}))$ for the connected $2D_4$ example displayed after Lemma 3.9 in Section 3; a nonzero value would contradict Proposition 2.1, which places that case outside the list of configurations with nontrivial obstruction.","supporting_citations":[],"review_version":1}