{"id":"bbfebe89-32b1-42bb-aadf-89756c7ceca4","arxiv_id":"2608.04793","paper_version":1,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The zeroth birational motivic homotopy category of a unibranch scheme is the product of those of its function fields.","lead":"Birational motivic homotopy theory studies shapes up to removing small subsets, and this paper builds a formal framework showing that, for well-behaved schemes, the entire theory is determined by the fields of functions at the generic points. The main payoff is a decomposition theorem and fiberwise criteria that could simplify computations in motivic homotopy theory and the slice filtration.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Cor. 4.3.3 assumes a qcqs scheme has finitely many connected components; Spec of an infinite product of fields is a qcqs unibranch counterexample, so the reduction to irreducible components in the proof of Thm. 6 is not justified.","rationale":"The reader is right that Theorem 4.3.2 contains an unjustified codimension inequality: for a flat map, the closure of the image of a closed subset can have strictly smaller codimension than the original subset, so the displayed inequality in the n-birational descent step is not valid in the stated generality. But that step is only needed for n > 0. The generic decomposition Theorem 6 / Cor. 4.3.3 is about H^0, and the H^0 case of Theorem 4.3.2 is completed earlier in the proof using dominance of the descended open immersion, without the codimension inequality. So the reader's identified flaw, while real, is not the load-bearing flaw for the paper's headline claim. The genuinely load-bearing gap for Theorem 6 is the assertion that a qcqs scheme has finitely many connected components. The affine scheme Spec of an infinite product of fields is qcqs and unibranch but has infinitely many connected components, which are generally not open. Consequently the reduction to irreducible components by finite Zariski additivity is not available, and the proof as written does not establish the generic decomposition for all Qcqs unibranch schemes. This does not change the overall verdict: the contribution remains potentially valuable, the headline theorem is not fully proven as written, and a conditional verdict is still appropriate.","tokens_in":39215,"tokens_out":14752,"duration_ms":182406,"concrete_test":"Run the proof of Cor. 4.3.3 on X = Spec(prod_{n in N} k). Verify that X is qcqs, reduced, unibranch (each local ring is a field), and that X^(0) is infinite with non-open connected components. Then check whether the stated proof supplies any mechanism — beyond finite additivity Cor. 4.1.2 — to pass from the irreducible case to this X. If no such mechanism appears, the proof has an unpatched gap: the canonical map H^0(X) -> prod_{eta in X^(0)} H^0(k(eta)) is not derived by the given argument. A minimal companion check is to rerun the proof with the hypothesis 'Noetherian qcqs' replacing 'qcqs'; all cited reduction steps then go through, confirming that the missing finiteness condition is exactly what the proof relies on.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's codimension inequality in Thm. 4.3.2 is indeed not justified for flat morphisms — e.g. the projection A^2 -> A^1 with Z = V(x) has cod_X(Z)=1 but the closure of its image has codimension 0. However, the H^0 case of Thm. 4.3.2, which is what Cor. 4.3.3 actually invokes, is proved by the preceding dominance argument and does not require that inequality; the problematic paragraph concerns general n and is not load-bearing for the headline H^0 decomposition. A more direct gap in the proof of the central claim is the sentence in Cor. 4.3.3: 'Since X is qcqs, it has finitely many connected components.' This is false. Let X = Spec(prod_{n in N} k). This is affine, hence qcqs; it is reduced, and every local ring is a field, so X is unibranch. Its connected components are the points of beta N, which are infinite and generally not open. Therefore the reduction to the case of a reduced irreducible scheme via finite Zariski additivity (Cor. 4.1.2) is unavailable as written. The generic decomposition theorem is thus not established for arbitrary Qcqs unibranch schemes; the proof only covers schemes that are finite disjoint unions of irreducible components, e.g. Noetherian or finitely presented cases. If the theorem is meant to include all Qcqs unibranch schemes, an additional descent or continuity argument for infinite families of components is needed.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a functorial calculus for the n-birational motivic homotopy category H^n(S). It constructs a Pr^L-valued presheaf on a category of correspondences with smooth and universally generalization-lifting (UGLT) morphisms, proves Nisnevich descent, deformation invariance, and a flat-affine continuity result, and uses these to derive a 'generic decomposition' theorem: for a Qcqs unibranch scheme X, H^0(X) is the product of H^0 of the residue fields of its generic points. Consequences include a fiberwise criterion for birational equivalences, rational and purely transcendental invariance, and a partial unstable slice conjecture.","tokens_in":39511,"tokens_out":15143,"duration_ms":162010,"significance":"The intended results are significant: a schematic functoriality for birational motivic homotopy categories would be a useful tool, and the generic decomposition for H^0 would give a conceptual reduction of birational invariants to function fields. The paper contains several valuable structural results (dense open locality, Nisnevich descent, deformation invariance, smooth base change) and provides reasonably detailed proof sketches, with explicit acknowledgement of limitations such as the failure of local Cartesian-ness. If the gaps identified below are repaired, the paper would make a solid contribution. However, the headline theorem as stated is not proved, and one continuity theorem is stated in greater generality than its proof supports.","major_comments":[{"comment":"The proof reduces to the irreducible case using the assertion 'Since X is qcqs, it has finitely many connected components.' This is false: X = Spec(∏_{n∈N} k) is affine and qcqs, reduced, and unibranch (every local ring is a field), but its connected components are the points of βN, which are infinite and not open. Consequently, the finite Zariski additivity of Corollary 4.1.2 cannot be applied, and the generic decomposition is not established for arbitrary Qcqs unibranch schemes. The proof as written only covers schemes that are finite disjoint unions of irreducible components (e.g., Noetherian schemes or finitely presented schemes). An additional continuity or descent argument for infinite sets of components is needed to justify the stated generality.","section":"Corollary 4.3.3 / Theorem 6"},{"comment":"The proof of the general n-case asserts that for the flat transition morphism φ_{α1}: X → X_{α1}, the closure Z' of the image of the complement Z satisfies cod_{X_{α1}} Z' ≥ cod_X(Z). This inequality is not valid for arbitrary flat morphisms: for the flat projection π: A^2 → A^1 and Z = V(x), cod_{A^2}(Z)=1 while the closure of the image has codimension 0 in A^1. The appeal to the methods of §3.2 is therefore misplaced. The n=0 case, which is what Corollary 4.3.3 and Corollary 4.4.7 actually invoke, is established by the preceding dominance argument and does not depend on this inequality; however, Theorem 4.3.2 as stated for all n is not proved by the given argument.","section":"Theorem 4.3.2"}],"minor_comments":[{"comment":"The paragraph beginning 'Suppose X/S is smooth...' appears twice verbatim; the duplicate should be deleted.","section":"Remark 2.5.4"},{"comment":"There are several typos: 'towoer' for 'tower', 'continuuity' for 'continuity', 'Frudenthal' for 'Freudenthal', and 'unbranch' for 'unibranch'. A careful proofreading pass is recommended.","section":"Throughout"},{"comment":"This corollary has two 'Proof.' labels and a stray 'Proof.' before the flatness argument; the presentation should be cleaned up.","section":"Corollary 3.1.7"},{"comment":"The citation [Gro66, Proposition (8.10.3)] is not specific enough for the descent of dominance along flat limits; please give the exact statement or tag used.","section":"§4.3"},{"comment":"The term 'geometrically unibranch' is used without a definition in the paper; please add one.","section":"Corollary 4.3.7"}],"recommendation":"major_revision","confidential_remarks":"The paper relies substantially on the author's unpublished preprint [Mai26a] for key inputs (e.g., the identification H^0 ≃ H^b, dense/Nisnevich locality, and compact-generation properties in Proposition 2.4.1). Some of these are reproved here, but the editor should check that [Mai26a] is available and that the results are not circular. In addition, if the infinite-component counterexample cannot be handled, the scope of Theorem 6 should be narrowed to Noetherian or finitely presented schemes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nQuick take: this paper is a serious, useful framework for n-birational motivic homotopy categories, and the generic decomposition theorem for H^0 is probably true, but the proof as written has a gap in the reduction step for arbitrary qcqs unibranch schemes. The reader's inequality concern is real, though it is not the main obstruction for the headline result.\n\nWhat is actually new: the promotion of H^n to a Pr^L-valued presheaf on the correspondence category, the generic decomposition, the fiberwise criteria, and the fully faithful embeddings from rational and purely transcendental extensions. These are solid contributions, and the author is honest about what is not done (e.g., stable theory, lack of local cartesianness). The reliance on [Mai26a] for foundational identifications is a caveat, but the new results are not mere reformulations.\n\nThe soft spots: (1) The proof of Cor. 4.3.3 claims a qcqs scheme has finitely many connected components. That is false for Spec of an infinite product of fields, which is qcqs, reduced, and unibranch. So the reduction to irreducible components does not work for arbitrary qcqs unibranch schemes; the theorem is proven for Noetherian (finite-component) cases. The author needs either an infinite-descent argument or a restricted statement. (2) In Thm. 4.3.2, the inequality cod_{X_alpha} Z^1 >= cod_X(Z) for the closure of the image under a flat map is not generally true (A^2 -> A^1 example). However, the H^0 case used in Cor. 4.3.3 is handled by a separate dominance argument, so this flaw affects the general-n continuity theorem rather than the generic decomposition.\n\nMy overall view: this deserves a serious referee. The framework is valuable and the main theorem is likely correct for varieties and Noetherian schemes. The gaps are repairable but require real work. I'd like to see a revised version before relying on the full qcqs statement.\n\nBest.","headline":"A promising framework and a likely-true main theorem, but the proof as written has a gap in the qcqs reduction step and an unjustified codimension inequality in the general continuity theorem.","tokens_in":40055,"tokens_out":3918,"would_cite":false,"duration_ms":45437,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F42","14E05"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that for any qcqs unibranch scheme $X$, the birational motivic homotopy category $H^0(X)$ decomposes as the product of the categories $H^0(k(\\eta))$ of its generic points' residue fields.","keywords":["birational motivic homotopy","n-dense open immersions","generic decomposition","unibranch schemes","Nisnevich descent","flat-affine continuity","slice filtration","function fields"],"falsifier":"Take the flat morphism $f: \\mathbb{A}^1_k \\to \\mathrm{Spec}(k)$ and the closed point $Z = \\{0\\} \\subset \\mathbb{A}^1_k$; then $\\operatorname{codim}_{\\mathbb{A}^1}(Z) = 1$ while $\\operatorname{codim}_{\\mathrm{Spec}(k)}(\\overline{f(Z)}) = 0$, contradicting the codimension inequality asserted in the proof of Theorem 4.3.2.","tokens_in":38970,"feed_emoji":"🧩","tokens_out":14744,"duration_ms":142549,"temperature":0.7,"pith_summary":"This paper promotes the $n$-birational motivic homotopy category $H^n(S)$ — the localization of the motivic homotopy category at open immersions whose complements have codimension greater than $n$ — to a $\\mathrm{Pr}^L$-valued presheaf on a category of correspondences built from smooth and generalization-lifting maps. The central result is a generic decomposition: for any qcqs unibranch scheme $X$, the zeroth birational motivic homotopy category $H^0(X)$ is equivalent to the product $\\prod_{\\eta \\in X^{(0)}} H^0(k(\\eta))$ of the categories attached to the residue fields of its generic points; in particular, for a variety $V$ one has $H^0(V) \\simeq H^0(K(V))$. From this the paper derives a fiberwise criterion for birational equivalence of smooth schemes over a unibranch base, and fully faithful embeddings along stably birational morphisms and purely transcendental field extensions. Along the way it proves Nisnevich descent, deformation invariance, and flat-affine continuity for $H^n$, and shows that $\\Omega_{\\mathbb{P}^1}$ lowers the height of birational locality, giving a partial unstable slice conjecture.","feed_headline":"Birational motivic homotopy of unibranch schemes reduces to fields","feed_subtitle":"The zeroth birational motivic homotopy category of a variety is determined by its function field.","key_machinery":"The machinery is the $n$-birational localization $L_n: H^{\\mathbb{A}^1}(S) \\to H^n(S)$ at $n$-dense open immersions — open immersions whose closed complement has codimension strictly greater than $n$ — together with the class of universally generalization-lifting (UGLT) morphisms, along which pullback preserves $n$-density. The paper assembles these into a presheaf $H^n(-)$ on the correspondence category $\\mathrm{Corr}(\\mathrm{Sch})_{\\mathrm{sm},\\mathrm{uglt}}$, where horizontal maps are UGLT pullbacks and vertical maps are smooth extensions. For the generic decomposition, the load-bearing tool is a flat-affine continuity theorem asserting that $H^n$ commutes with limits of pro-schemes with flat affine transition maps; applied to the pro-system of dense open affines with generic point, and combined with deformation invariance, Nisnevich descent, and dense locality of $H^0$, this identifies $H^0(X)$ with the product of the field-theoretic categories attached to its generic points.","core_discovery":"The central discovery is that the $n$-birational motivic homotopy category $H^n(S)$, despite not being stable under arbitrary pullbacks, has a well-defined functoriality along morphisms that universally lift generalizations (UGLT) and along smooth morphisms, packaged as a $\\mathrm{Pr}^L$-valued presheaf on $\\mathrm{Corr}(\\mathrm{Sch})_{\\mathrm{sm},\\mathrm{uglt}}$. Using this functoriality, the paper proves Nisnevich descent and deformation invariance for $H^n$, and a flat-affine continuity theorem for pro-schemes with flat affine transition maps. These combine to yield the generic decomposition: for a qcqs unibranch scheme $X$, the canonical map $H^0(X) \\to \\prod_{\\eta \\in X^{(0)}} H^0(k(\\eta))$ is an equivalence of $\\infty$-categories. The paper's main consequence is that birational motivic homotopy theory over unibranch schemes is purely field-theoretic: the category $H^0(V)$ of a variety is equivalent to $H^0(K(V))$, and birational equivalences over a unibranch base are detected by the birational contractibility of generic fibers.","pith_inferences":["If the flat-affine continuity step can be repaired, the generic decomposition would likely extend from unibranch schemes to arbitrary qcqs schemes after passing to irreducible components, since the proof only needs the descent of dense open immersions at finite stages.","The field-theoretic reduction suggests that $H^0$ over unibranch schemes is a birational invariant of function fields; one could test this by comparing $H^0(K)$ for purely inseparable or imperfect field extensions, which are not covered by the purely transcendental statement.","The $\\Omega_{\\mathbb{P}^1}$ tower might stabilize to a birational stable homotopy category whose objects carry transfers, mirroring the role of stable motivic homotopy theory; the author notes this limit is unexplored.","A direct computation of $H^0$ for a non-unibranch scheme such as $\\mathrm{Spec}(k[x,y]/(xy))$ would clarify whether the product decomposition is specific to unibranch schemes or a general phenomenon."],"forward_implications":["If $X$ is a qcqs unibranch scheme, computing $H^0(X)$ reduces to computing $H^0$ of the residue fields of its generic points; for a variety $V$, $H^0(V) \\simeq H^0(K(V))$.","A morphism $f: X \\to Y$ in $\\mathrm{Sm}_S$ over a qcqs geometrically unibranch base $S$ is a birational equivalence as soon as each generic fiber of $f$ is birationally contractible.","Stably birational morphisms — for instance, projections from rational schemes — induce fully faithful embeddings $H^0(S) \\hookrightarrow H^0(X)$.","Purely transcendental field extensions $k \\subset k(t_1,\\dots,t_n)$ induce fully faithful embeddings $H^0(k) \\hookrightarrow H^0(k(t_1,\\dots,t_n))$.","The unstable birational slice tower is partially computed: $\\Omega_{\\mathbb{P}^1}$ carries $H^{n+1}$-local objects to $H^n$-local objects, reducing the unstable slice conjecture to a canonical comparison map being an $n$-birational equivalence."],"supporting_citations":[{"why":"Defines the $n$-birational motivic homotopy categories $H^n$ and their localizations, the objects whose functoriality is analyzed.","marker":"[BE21]"},{"why":"Establishes dense locality and the equivalence $H^0 \\simeq H^b$, which the paper uses repeatedly for the generic decomposition and its consequences.","marker":"[Mai26a]"},{"why":"Introduced the unstable slice filtration and $n$-dense localizations, providing the original motivation and the definition of the categories.","marker":"[Pel14]"},{"why":"Supplies the Nisnevich descent and six-operation formalism for motivic homotopy that the paper extends to $H^n$.","marker":"[Hoy17]"},{"why":"Provides the fully faithful pushforward along closed immersions used in the refined pushforward section.","marker":"[Kha16]"},{"why":"Supplies the $\\infty$-categorical foundations ($\\mathrm{Pr}^L$, accessible localizations, filtered colimits) on which all constructions rest.","marker":"[Lur09]"},{"why":"Provides the scheme-theoretic codimension, flatness, and descent lemmas used in the continuity and descent arguments.","marker":"[Sta26]"},{"why":"Supplies the lifting of open immersions along nilpotent thickenings used in the deformation-invariance proof.","marker":"[Gro66]"}],"fun_headline_variants":["Birational motivic homotopy reduces to function fields","Birational homotopy of schemes collapses to fields","Generic fibers detect birational equivalence of schemes","Stably birational maps embed motivic homotopy categories","Birational homotopy of varieties determined by function fields"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The generic decomposition rests on the assumption that a flat morphism cannot decrease the codimension of the closure of a closed subscheme; this inequality is false for arbitrary flat morphisms, and that is the step on which the flat-affine continuity proof depends.","fun_headline_variants_meta":{"raw":{"variants":["Birational motivic homotopy reduces to function fields","Birational homotopy of schemes collapses to fields","Generic fibers detect birational equivalence of schemes","Stably birational maps embed motivic homotopy categories","Birational homotopy of varieties determined by function fields"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00119,"raw_usage":{"total_tokens":4931,"prompt_tokens":987,"completion_tokens":3944,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":603,"completion_tokens_details":{"reasoning_tokens":3866}},"tokens_in":603,"tokens_out":3944,"duration_ms":28424,"temperature":1.0,"reasoning_tokens":3866,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T16:29:51.955448+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the flat morphism $f: \\mathbb{A}^1_k \\to \\mathrm{Spec}(k)$ and the closed point $Z = \\{0\\} \\subset \\mathbb{A}^1_k$; then $\\operatorname{codim}_{\\mathbb{A}^1}(Z) = 1$ while $\\operatorname{codim}_{\\mathrm{Spec}(k)}(\\overline{f(Z)}) = 0$, contradicting the codimension inequality asserted in the proof of Theorem 4.3.2.","supporting_citations":[],"review_version":2}