{"id":"31b8ac92-d51b-404e-a071-32d82e5f1174","arxiv_id":"1908.02803","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For fixed Fano index e, a very general complex Fano hypersurface of dimension n has degree of irrationality at least sqrt(n)/4 for all sufficiently large n; this is the first construction of rationally connected varieties with degree of irrationality at least 4.","lead":"Very general Fano hypersurfaces of large dimension have degree of irrationality at least a constant times the square root of the dimension, even when the Fano index is fixed. This gives the first known rationally connected varieties whose degree of irrationality exceeds 3, and it shows that non-rationality of these varieties is quantitatively severe.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem A's proof for degrees not divisible by a useful prime rests on an unproved degeneration to a union of a degree-pd hypersurface and f hyperplanes; this step needs explicit verification before the bound covers all sufficiently large n.","rationale":"The reader's CONDITIONAL verdict is well-founded. The main theorem's proof has two pillars: the char-p degeneration (Construction 2.4, via Kollár) and the specialization theorem (Theorem 1.1). The former is a citation and is presumably sound; the latter is proved in §1. The genuinely unproved part is the second paragraph of the proof of Theorem A, where degrees not of the form pd are handled by an asserted degeneration to a union of a degree-pd hypersurface and f hyperplanes. This is the only step that makes the lower bound apply to all sufficiently large n, since for fixed e the degrees n+2-e are not all multiples of a useful prime. The paragraph states the numerical condition (∗) and invokes Theorem 1.1(2), but it does not construct the family or verify normality, very generality of the component, or the component-wise recomputation of the separating line bundle. The concrete check proposed would either turn this paragraph into a proof or reveal that the theorem only covers a sparse set of degrees. I agree with the reader's verdict, though my emphasis is on the D=pd+f step rather than on Kollár's construction itself.","tokens_in":7458,"tokens_out":45659,"duration_ms":498093,"concrete_test":"Work out the omitted degeneration for one non-pd degree. Take e=1, n=100, p=5: degree 101, f=1, d=20. Define F_t = G·H + t·F_X in P^{101}_{A^1}, with G a very general smooth degree 100 hypersurface and H a general hyperplane. Verify: (1) the hypersurface X_T is normal and its geometric generic fiber is smooth; (2) the component G=0 has codegree 2, so the pulled-back line bundle is O(18); (3) applying Lemma 2.3 and the p-divisibility of inseparable degrees to this component gives no maps of degree ≤4 to a ruled variety; (4) Theorem 1.1(2) transfers this to the geometric generic fiber, so a very general complex degree 101 hypersurface in P^{101} has irr ≥5. If any of (1)-(4) cannot be verified, the proof of Theorem A for non-pd degrees is incomplete.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim Theorem A must bound irr for every sufficiently large n, not just for degrees of the form pd. The only argument covering the remaining degrees is the second paragraph of the proof of Theorem A: 'This can be proved by degenerating X to a union of a very general degree pd hypersurface and f hyperplanes. The claim then follows from Theorem 1.1(2) and the previous paragraph.' This is an omitted proof, not a routine citation. For it to work one needs (i) a flat family X_T over a DVR whose total space is normal and integral, with smooth geometric generic fiber and special fiber G ∪ H_1 ∪ ... ∪ H_f; (ii) the component G=0 to be a very general degree pd hypersurface so the previous paragraph's no-low-degree-map conclusion applies to it; and (iii) the numerical hypothesis on that component to be d-e-f ≥ 2p-2, because its codegree is e+f, so the separating line bundle is O(d-e-f), not O(d-e). None of (i)-(iii) is checked. In particular, if the component G=0 cannot be chosen very general while keeping X_T normal, or if Theorem 1.1(2) does not transfer the no-map property from the component to the geometric generic fiber when the special fiber is reducible, then Theorem A only covers degrees divisible by a useful prime, and the 'arbitrarily large degrees' statement for all sufficiently large n would not follow. This is exactly the step the referee must ask the authors to expand.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem A: for every fixed Fano index e, there is an N such that for all n > N, the very general complex Fano hypersurface of dimension n and index e has degree of irrationality at least sqrt(n)/4. The method combines a specialization theorem (Theorem B / Theorem 1.1: existence of a degree-d rational map to a ruled variety on the geometric generic fiber forces the same on suitable components of the special fiber) with Kollár's characteristic-p degeneration. In positive characteristic, a purely inseparable cover of a degree-d hypersurface carries a line bundle M, coming from O(d-e), which separates many points and embeds into a wedge of the cotangent sheaf; Lemma 2.3 converts this into the nonexistence of low-degree separable maps to ruled varieties. The main theorem then follows by applying the specialization theorem in the contrapositive and by estimating suitable primes via Bertrand's postulate. The paper also derives Proposition C (specialization of degree of irrationality for irregularity-zero surfaces and strict Calabi-Yau threefolds) and Corollary D (every complex abelian surface has degree of irrationality at most 4).","tokens_in":7749,"tokens_out":26510,"duration_ms":313051,"significance":"If the main theorem is correct, it gives the first known rationally connected varieties with degree of irrationality at least 4, and the sqrt(n) growth is a new phenomenon for Fano hypersurfaces. The specialization theorem Theorem B/1.1 is a clean and reusable statement about ruled covers, and the point-separation Lemma 2.3 is a natural and potentially useful tool. The proof is largely self-contained and has no visible free parameters: the numerical estimate is explicit and parameter-free. The main caveat is that the proof of Theorem A as printed covers degrees divisible by a suitable prime in detail, while the extension to arbitrary degrees is asserted in a single sentence; that step must be expanded before the stated theorem is fully supported. I regard the central idea as sound and the paper as publishable after the missing argument is supplied.","major_comments":[{"comment":"The paper invokes Kollár's degeneration, Construction 2.4, for a family whose special fiber is a reduced degree-p inseparable cover of a smooth degree-d hypersurface and for which M := ν^*O(pd+d-n-2) injects into ∧^{n-1}Ω_{X'_κ}. The hypotheses under which [8, §5] applies are not stated. In the numerical range used at the end, d ≈ (n+2)/p and p ≈ sqrt(n)/4, so d is comparable to or larger than p; the paper should state the exact conditions on p and d (for example, any requirements on d relative to p or on the singularities of Y) and confirm that every pair (p,d) produced by the prime-counting estimate satisfies them. If Kollár's construction imposes additional inequalities, they must be inserted into condition (∗).","section":"§2, proof of Theorem A, second paragraph"}],"minor_comments":[{"comment":"In the trace argument, the separation property is used to choose a section vanishing on all but one of the 2b preimage points. Please specify that the nonvanishing point is chosen to lie over z_2, so that the traced form vanishes at z_1 and does not vanish at z_2; as written, the direction of the vanishing at the two points is ambiguous.","section":"§2, Lemma 2.3"},{"comment":"After base changing the no-map statement from the countable field η to C, the conclusion is drawn for a very general complex hypersurface, rather than just for the one complex member coming from Kollár's degeneration. This is standard because the locus of hypersurfaces admitting a rational map of degree at most p-1 to a ruled variety is a countable union of closed subvarieties, but the argument should be stated once.","section":"§2, proof of Theorem A"},{"comment":"The numerical example appears inconsistent: for p=5, e=1, and n=35 one has n+2 = 37 and D = 36, so with f = 1 and d = 7 the quantity d-e-f equals 5, which is less than 2p-2 = 8. The displayed statement that a very general degree 35 hypersurface in P^{35} has irr(X) ≥ 4 should be corrected, or the intended parameters (possibly n=34 or a different d,f) should be spelled out.","section":"Remark 2.6"},{"comment":"In the case dim B = 2, the sentence \"As X_0 is simply connected, the map from X_0 to B factors through B'\" is terse. The factorization through the étale cover B' follows from the vanishing of π_1(X_0) once basepoints are chosen, but the step should be made explicit, especially because the lifting statement for rational maps to a cover is not automatic.","section":"§1, proof of Proposition C"}],"recommendation":"major_revision","confidential_remarks":"The core specialization theorem and the point-separation lemma are solid and interesting, and I do not see circularity in the main argument. The decisive issue is the one-sentence treatment of arbitrary degrees in the proof of Theorem A; this needs to be expanded, not only for completeness but because the current proof genuinely does not cover all n if the asserted degeneration fails. The example in Remark 2.6 also needs correction. I recommend major revision rather than rejection because the missing step is plausibly standard and the overall architecture is coherent."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper has a genuinely new result and a clean tool. Theorem B—minimal degree of a dominant map to a ruled variety can only drop under specialization—is independently useful and appears correct. Lemma 2.3 is the other key piece: it turns point-separating line bundles into degree lower bounds via a trace map. That argument is standard and I don't see a problem. The claimed application to Fano hypersurfaces would give the first rationally connected varieties with irr ≥ 4, which would be a real step.\n\nThe main soft spot is in the proof of Theorem A, the paragraph covering degrees not divisible by p. The text says, essentially, that you can degenerate X to a union of a very general degree pd hypersurface and f hyperplanes, then apply Theorem 1.1(2). That is a whole argument in one sentence. To make it work you need a flat family over a DVR with normal total space, the degree pd component very general, and the numerical condition d−e−f ≥ 2p−2 (because that component has codegree e+f, not e). None of this is checked. The stress-test note gets this right. Without this step, the proof as written only covers degrees of the form pd for your chosen primes, not all sufficiently large n.\n\nI want to stress that this is a repairable gap, not a sign of sloppiness. The rest of the paper is clear and the citations are appropriate. The prime-counting bound is fine. The Corollary D application is separate but plausible.\n\nVerdict: send it out, but the referee should be specifically asked to fill in the degeneration construction. If that step works, the paper is a solid contribution. If not, the theorem is weaker but still interesting. I'd read it again after revision.","headline":"Genuinely new lower bounds for degree of irrationality of Fano hypersurfaces, with a clean specialization theorem, but the proof of Theorem A currently rests on an unstated degeneration argument.","tokens_in":8281,"tokens_out":4317,"would_cite":true,"duration_ms":43985,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14E08","14E05","14J70","14M20"],"pacs":[],"model":"deepseek-v4-flash","headline":"For any fixed Fano index e, very general complex Fano hypersurfaces of sufficiently large dimension n have degree of irrationality at least $\\sqrt{n}/4$.","keywords":["degree of irrationality","Fano hypersurfaces","rationally connected varieties","positive characteristic degeneration","ruled varieties","specialization of birational invariants","point-separating line bundles"],"falsifier":"One concrete test would be to compute the minimal degree of a dominant rational map from the special fiber of the p-fold inseparable cover described in the degeneration to a ruled variety. If for some p and d in the claimed range the special fiber admitted such a map of degree below $\\lfloor (d-e)/2\\rfloor$, the point-separating lemma would be contradicted; equivalently, exhibiting a very general complex Fano hypersurface $X_{n,e}$ with $\\mathrm{irr}(X_{n,e}) < \\sqrt{n}/4$ would refute the theorem.","tokens_in":7185,"feed_emoji":"📐","tokens_out":8585,"duration_ms":84185,"temperature":0.7,"pith_summary":"The paper tries to establish a quantitative lower bound on how far Fano hypersurfaces can be from being rational. For every fixed Fano index e and all sufficiently large dimensions n, the very general complex Fano hypersurface of dimension n and codegree e has degree of irrationality at least $\\sqrt{n}/4$. Since the degree of irrationality counts the smallest degree of a dominant rational map to projective space, this means these varieties cannot be covered rationally by maps of small degree. The result supplies the first examples of rationally connected varieties whose degree of irrationality is at least 4, and it is proved by specializing to positive characteristic and tracking maps to ruled varieties through the specialization.","feed_headline":"Fano hypersurfaces have degree of irrationality at least √n/4","feed_subtitle":"First rationally connected examples with degree of irrationality above 3, proved by degeneration to characteristic p.","key_machinery":"The mechanism is an injection of a line bundle into the sheaf of $(n-1)$-forms on a resolution of a positive-characteristic degeneration. In the degeneration construction, the special fiber is a purely inseparable degree-p cover of a smooth degree-d hypersurface, and the pullback line bundle $\\mathcal{O}(pd+d-n-2)$ injects into $\\wedge^{n-1}\\Omega$ on a resolution. Sections of this line bundle separate many points, and a trace-map argument shows that any separable rational map to a ruled variety would have to have degree at least one more than half the number of separated points. That converts separation of points into a lower bound on degrees of maps, and the bound is then transferred from the special fiber to the complex generic fiber by the specialization theorem.","core_discovery":"The central claim is that the irrationality of Fano hypersurfaces grows at least as the square root of the dimension. More precisely, for a fixed codegree e there is an N such that for all n > N, the very general complex Fano hypersurface $X_{n,e}$ of dimension n and codegree e satisfies $\\mathrm{irr}(X_{n,e}) \\ge \\sqrt{n}/4$. The paper proves the stronger statement that every dominant rational map from $X_{n,e}$ to a ruled variety has degree at least $\\sqrt{n}/4$. A companion specialization theorem says that, in a family over a DVR, the minimal degree of a dominant generically finite rational map to a ruled variety can only drop on special fibers; this is what lets the positive-characteristic obstruction push back to characteristic zero.","pith_inferences":["I infer that the same mechanism should give lower bounds for other rationality measures that count degrees of covers, such as covering gonality or the least degree of a map to a uniruled variety, whenever a point-separating line bundle injects into a wedge power of the cotangent sheaf.","I infer that the specialization theorem, stated for maps to ruled varieties, may extend further to the degree of irrationality itself in larger classes of families with controlled fundamental group or Hodge numbers.","I infer that a sharper point-separation count in the key lemma could improve the constant $1/4$ and possibly the exponent in the lower bound, since the current argument is optimized only through the prime-counting step."],"forward_implications":["For fixed codegree e and $n > (4e-4)^2 - 2$, the very general complex Fano hypersurface satisfies $\\mathrm{irr}(X_{n,e}) \\ge \\sqrt{n}/4$.","The same lower bound applies to the minimal degree of a dominant rational map to any ruled variety, not just to projective space.","In the families covered by Proposition C (surfaces with $H^{1,0}=0$ and strict Calabi-Yau threefolds), the degree of irrationality can only drop under specialization.","Every complex abelian surface has degree of irrationality at most 4.","The specialization theorem gives a general tool for transferring lower bounds on maps to ruled varieties from special fibers to generic fibers in arbitrary families over a DVR."],"supporting_citations":[{"why":"Supplies the positive-characteristic degeneration construction whose line-bundle injection into $\\wedge^{n-1}\\Omega$ drives the lower bound.","marker":"[8]"},{"why":"Provides the birational extension lemma used to move a rational map to a ruled variety from the generic fiber to a special component.","marker":"[10]"},{"why":"Supplies the ruledness criterion applied to the image of a special fiber component.","marker":"[9]"},{"why":"Provides the trace map for rational maps that converts separated sections into contradictory differential forms.","marker":"[5]"},{"why":"Establishes the very general abelian-surface input used to derive the bound for every abelian surface.","marker":"[4]"},{"why":"Provides the field-isomorphism lemma that lets very general fibers stand in for the geometric generic fiber in applications.","marker":"[15]"}],"fun_headline_variants":["Fano hypersurfaces: irrationality at least √n/4","First rationally connected varieties with irrationality >3","Fano hypersurfaces force irrationality ≥√n/4","Irrationality of Fano hypersurfaces grows as √n","New result: Fano hypersurfaces have irrationality ≥√n/4"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the cited positive-characteristic degeneration really produces the line-bundle injection into the sheaf of $(n-1)$-forms on a resolution; if that injection fails for the relevant primes and degrees, the lower bound on the degree of irrationality does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Fano hypersurfaces: irrationality at least √n/4","First rationally connected varieties with irrationality >3","Fano hypersurfaces force irrationality ≥√n/4","Irrationality of Fano hypersurfaces grows as √n","New result: Fano hypersurfaces have irrationality ≥√n/4"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000386,"raw_usage":{"total_tokens":1994,"prompt_tokens":855,"completion_tokens":1139,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":471,"completion_tokens_details":{"reasoning_tokens":1061}},"tokens_in":471,"tokens_out":1139,"duration_ms":10733,"temperature":1.0,"reasoning_tokens":1061,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:37:30.671600+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"One concrete test would be to compute the minimal degree of a dominant rational map from the special fiber of the p-fold inseparable cover described in the degeneration to a ruled variety. If for some p and d in the claimed range the special fiber admitted such a map of degree below $\\lfloor (d-e)/2\\rfloor$, the point-separating lemma would be contradicted; equivalently, exhibiting a very general complex Fano hypersurface $X_{n,e}$ with $\\mathrm{irr}(X_{n,e}) < \\sqrt{n}/4$ would refute the theorem.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the birational extension lemma used to move a rational map to a ruled variety from the generic fiber to a special component."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the ruledness criterion applied to the image of a special fiber component."},{"cited_title":"B/a.sc/s.sc/t.sc/i.sc/a.sc/n.sc/e.sc/l.sc/l.sc/i.sc, P","cited_arxiv_id":null,"evidence_quote":"Provides the trace map for rational maps that converts separated sections into contradictory differential forms."},{"cited_title":"For a hypersurface X ⊂ Pn+1 of degree d , recall from the introduction that the codegree of X is e := n + 2 − d","cited_arxiv_id":null,"evidence_quote":"Establishes the very general abelian-surface input used to derive the bound for every abelian surface."},{"cited_title":"K/o.sc/n.sc/t.sc/s.sc/e.sc/v.sc/i.sc/c.sc/h.sc /a.sc/n.sc/d.sc Y","cited_arxiv_id":null,"evidence_quote":"Provides the field-isomorphism lemma that lets very general fibers stand in for the geometric generic fiber in applications."}],"review_version":1}