{"id":"a135e014-2359-4e03-8070-00664dd1f105","arxiv_id":"1908.02162","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"MGL-modules over a scheme are equivalent to motivic spectra with finite syntomic transfers, and the infinite P^1-loop space of MGL is the A^1-homotopy type of the moduli stack of finite syntomic schemes.","lead":"This paper proves a structural equivalence in motivic homotopy theory: modules over the algebraic cobordism spectrum MGL are the same as motivic spectra with finite syntomic transfers. It also identifies the infinite P^1-loop spaces of MGL with moduli stacks of finite schemes, giving algebraic analogs of classical cobordism spaces.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.4.1(i) invokes Theorem 3.3.10 for the rank-zero summand ι_0, but Theorem 3.3.10 is stated and proved only for β: B→K_{>0}; the MGL (rank-0) case is not covered, leaving a gap in the proof of Theorem 4.1.3.","rationale":"The reader's weakest assumption points to the left Kan extension criterion (Proposition A.0.4) as load-bearing for base change. That is a legitimate concern about an external, difficult theorem. However, a more immediate and internal gap appears earlier: the proof of Theorem 3.4.1(i), which identifies MGL with the framed suspension spectrum of FSynS, applies Theorem 3.3.10 to the rank-zero summand ι_0, even though Theorem 3.3.10 is stated and proved only for β:B→K_{>0}. This is not a subtle failure of a cited result but a direct mismatch between the hypothesis of the main geometric theorem and its use. The paper's own definitions and comparison theorems (hfr, Theorem 2.4.9, Theorem 3.2.1) restrict to rank >0, so the rank-zero case cannot be obtained by a purely formal consequence. Remark 3.3.15 acknowledges the rank-zero difficulty and provides only a partial loop-space statement over perfect fields, explicitly noting that it does not suffice for the module theorem. Since Theorem 4.1.3 is the central claim, this gap is the most load-bearing concern. If the rank-zero equivalence MGL≃Σ∞_{T,fr}FSynS can be established by a separate argument (e.g., via the reconstruction theorem for ξ=0 and a limit argument for general rank-0 classes), the paper's conclusion would stand; hence the verdict should be conditional rather than outright rejection.","tokens_in":40758,"tokens_out":27565,"duration_ms":267048,"concrete_test":"Verify whether Theorem 3.2.1 can be extended to rank-0 classes by proving directly that for Y smooth over S and ξ∈K(Y) with rank 0, there is a natural equivalence Th_{Y/S}(ξ) ≃ Σ∞_{T,fr}hfr_S(Y,ξ), at least when ξ=0 (then hfr_S(Y,0) is the framed correspondence presheaf and the equivalence should follow from the reconstruction theorem [Hoy20, Theorem 18]). If this rank-0 extension fails or requires new hypotheses, the proof of Theorem 3.4.1(i) and hence of Theorem 4.1.3 is incomplete as written.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The main equivalence Mod_MGL(SH(S)) ≃ SH_fsyn(S) (Theorem 4.1.3) relies on the identification MGLS ≃ Σ∞_{T,fr}FSynS (Theorem 3.4.1(i)). The proof of Theorem 3.4.1(i) says it is an instance of Theorem 3.3.10 where β is the inclusion of the rank n summand of K-theory. But for (i), n=0: MGL is the Thom spectrum of the rank 0 summand ι_0: K_0→K (by [BH20, Theorem 16.13] as cited in the paper), and FSynS = FQSm^{ι_0}_S (Example 3.3.2). Theorem 3.3.10 is explicitly stated only for smooth β:B→K_{>0}, and its proof uses Theorem 3.2.1, which is proved only for ξ∈K(Y) of rank >0. The rank-0 case is also excluded from the definitions: hfr_S(Y,ξ) is introduced in §2.2 for rank >0, and the comparison equivalences of Theorem 2.4.9 and Theorem 3.2.1 assume rank >0. Remark 3.3.15 gives only a loop-space-level argument for rank 0 over perfect fields, not the spectrum-level framed-suspension equivalence needed for Theorem 3.4.1(i), and not over arbitrary S. Thus the written proof of Theorem 3.4.1(i) leaves the rank-zero case (MGL itself) unjustified, and the main theorem depends on it.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves that for every scheme S, the ∞-category of MGL-module spectra is equivalent to the ∞-category of motivic spectra with finite syntomic transfers (Theorem 4.1.3), and similarly for MSL with an orientation condition (Theorem 4.2.1). The proof is built on a description of motivic Thom spectra of positive-rank virtual vector bundles as framed suspension spectra of moduli stacks of finite quasi-smooth derived schemes with the corresponding tangential structure (Theorem 3.3.10). Over perfect fields, the motivic recognition principle is used to deduce explicit models for infinite P^1-loop spaces, for instance identifying Ω∞_P^1 MGL with the group completion of the moduli stack of finite syntomic schemes, and the paper also gives Hilbert-scheme models and identifies HZ as an MGL-module.","tokens_in":41129,"tokens_out":6981,"duration_ms":73228,"significance":"If the main theorems hold, the paper gives a complete and base-scheme-independent description of MGL-modules in terms of coherent finite syntomic transfers, without any resolution-of-singularities hypothesis. This is a major advance in motivic homotopy theory and provides a structured, Quillen-style universal property for algebraic cobordism. The paper is careful about functoriality and multiplicative structures, and it contains substantial appendices (A and B) that lay out the technical foundations, including a proof of the left Kan extension criterion used for arbitrary base schemes. These strengths make the paper a likely important reference for the field.","major_comments":[{"comment":"Theorem 3.4.1(i) asserts the equivalence MGL_S ≃ Σ∞_{T,fr}FSyn_S, but the proof says this is an instance of Theorem 3.3.10 with β the inclusion of the rank-n summand of K-theory. For (i) the relevant summand is the rank-0 summand ι_0: K_0→K, whereas Theorem 3.3.10 is stated and proved only for β: B→K_{>0}. The proof of Theorem 3.3.10 relies on Theorem 3.2.1, which is proved only for ξ∈K(Y) of rank>0, and the comparison theorems in §2.4, notably Theorem 2.4.9, equally assume rank>0. Remark 3.3.15 gives only a loop-space-level equivalence over perfect fields and explicitly states that it does not suffice for the module theorem. Since Theorem 4.1.3 depends on Theorem 3.4.1(i), the proof of the main theorem is incomplete as written. Please supply a proof of the rank-zero spectrum-level equivalence, or give an alternative argument that covers the rank-0 summand.","section":"§3.4, Theorem 3.4.1(i) and its proof"}],"minor_comments":[{"comment":"The paper repeatedly says that certain proofs are 'exactly the same' as in [EHK+19] or 'almost identical' to results in that paper; given the technical weight of these results, it would be helpful to include a short dictionary or at least spell out the changes needed in the twisted-rank setting.","section":"Throughout"},{"comment":"The proof should explicitly state why the rank-0 inclusion satisfies the hypotheses of Theorem 3.3.10, or else redirect the reader to a separate argument; the current one-line appeal to Theorem 3.3.10 is misleading because that theorem is restricted to positive-rank structures.","section":"§3.4, proof of Theorem 3.4.1(i)"},{"comment":"There are minor formatting issues in the abstract and in the author affiliation line (for example, the corrupted 'F akult¨at'); these should be corrected in the final version.","section":"Abstract and §1.6"}],"recommendation":"major_revision","confidential_remarks":"The rank-zero gap in the proof of Theorem 3.4.1(i) is a genuine load-bearing issue, but it is likely fixable by extending the comparison results to ξ of rank 0 or by a separate argument using known framed-correspondence results. I am not recommending rejection because the rest of the paper is carefully developed and the gap seems local rather than fatal. The self-citation burden is high but the referenced companion papers are independent and appear to be the appropriate sources for the technical tools used here."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper's main theorem — Mod_MGL(SH(S)) ≃ SH_fsyn(S) — is the kind of result people in motivic homotopy theory have wanted for a while, and the framework they build to get it is genuinely useful. The positive-rank cases are proved carefully: Theorem 3.3.10 is a real step forward, and the moduli-stack models for Σ^n_T MGL (n>0) are concrete and convincing. The authors are also honest about how much they lean on EHK+19 and Hoy20, and those dependencies are legitimate, not circular.\n\nBut the stress-test note is right, and it lands on a load-bearing spot. Theorem 3.4.1(i) — the equivalence MGLS ≃ Σ∞_{T,fr}FSynS — is stated as an instance of Theorem 3.3.10 with β the inclusion of the rank-n summand of K-theory. For (i), n=0, so β lands in K_0, not K_{>0}. Theorem 3.3.10 is only proved for β: B → K_{>0}, and the whole chain leading to it (h_fr with rank>0, Theorem 2.4.9, Theorem 3.2.1) excludes rank 0. Remark 3.3.15 gives a loop-space-level argument over perfect fields, but not the spectrum-level framed-suspension equivalence over arbitrary S that Theorem 4.1.3 uses. So as written, the proof of the main equivalence has a gap.\n\nThis is a fixable gap, I suspect. The untwisted framed-correspondence machinery from EHK+19 presumably handles the rank-zero case, and the authors may have simply been sloppy in citing Theorem 3.3.10. But the referee needs to ask for a complete proof of Theorem 3.4.1(i), or for a version of Theorem 3.3.10 that includes rank 0, before the headline result is fully supported. The rest of the paper — the Hilbert scheme models, the MSL variants, the loop-space computations — holds up fine and is well worth having.\n\nThis paper deserves a serious referee, and I'd send it out. It's a substantial contribution with a clear, patchable gap in the written proof. The right outcome is probably acceptance after revision, not rejection. I'd bring it to a reading group, and I'd cite it once the rank-zero point is settled.","headline":"Major step toward the MGL-module picture, but the proof of Theorem 3.4.1(i) skips the rank-zero case that the main theorem depends on.","tokens_in":41676,"tokens_out":3606,"would_cite":true,"duration_ms":40595,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F42","19E15"],"pacs":[],"model":"deepseek-v4-flash","headline":"Modules over algebraic cobordism are exactly motivic spectra with finite syntomic transfers.","keywords":["algebraic cobordism","MGL","motivic homotopy theory","framed correspondences","finite syntomic transfers","motivic Thom spectra","derived algebraic geometry","modules over ring spectra"],"falsifier":"Over a perfect field $k$, the theorem predicts that the category of grouplike $\\mathbb{A}^1$-invariant sheaves with finite syntomic transfers is prestable and that the $\\mathbb{G}_m$-suspension functor is fully faithful; a direct way to falsify the theorem would be to find two non-isomorphic such sheaves whose images become equivalent after one suspension with respect to $\\mathbb{G}_m$, since the claimed cancellation theorem forbids this.","tokens_in":40558,"feed_emoji":"","tokens_out":10959,"duration_ms":105498,"temperature":0.7,"pith_summary":"This paper establishes that the $\\infty$-category of modules over Voevodsky's algebraic cobordism spectrum $\\mathrm{MGL}$ is equivalent, over any base scheme, to the $\\infty$-category of motivic spectra equipped with transfers along finite syntomic morphisms. This matters because it turns a purely algebraic module structure into a geometric structure that can be described concretely, with no resolution of singularities hypothesis. The proof identifies every motivic Thom spectrum built from a nonnegative virtual vector bundle with a framed suspension spectrum of a moduli stack of finite quasi-smooth derived schemes, and then reads off the module description. Over a perfect field, very effective $\\mathrm{MGL}$-modules are the same as grouplike motivic spaces with finite syntomic transfers.","feed_headline":"An MGL-module is exactly a finite-syntomic-transfer spectrum","feed_subtitle":"Over any base scheme, modules over algebraic cobordism are the same as motivic spectra with finite syntomic transfers.","key_machinery":"The load-bearing construction is the presheaf $h^{\\mathrm{fr}}_S(Y,\\xi)$ of $\\xi$-twisted tangentially framed correspondences: a span $X \\leftarrow Z \\rightarrow Y$ in which $f$ is finite quasi-smooth and the cotangent complex satisfies $L_f \\simeq -g^*(\\xi)$ in $K(Z)$. These presheaves assemble into framed motivic spectra via the reconstruction equivalence $\\mathrm{SH}(S) \\simeq \\mathrm{SH}^{\\mathrm{fr}}(S)$, and the paper proves that the Thom spectrum $M_\\beta$ of a nonnegative virtual bundle is the framed suspension spectrum of the corresponding moduli stack $\\mathrm{FQSm}^\\beta_S$. The module theorem then follows by identifying $\\mathrm{MGL}$ with $\\Sigma^\\infty_{\\mathbb{T},\\mathrm{fr}} \\mathrm{FSyn}_S$ and using an adjunction between framed correspondences and finite syntomic correspondences.","core_discovery":"The central discovery is a symmetric monoidal equivalence $\\mathrm{Mod}_{\\mathrm{MGL}}(\\mathrm{SH}(S)) \\simeq \\mathrm{SH}_{\\mathrm{fsyn}}(S)$ for every scheme $S$, natural in $S$ and compatible with the forgetful functors to $\\mathrm{SH}(S)$: a structure of $\\mathrm{MGL}$-module on a motivic spectrum is exactly a coherent system of finite syntomic transfers. The same pattern holds for the special linear cobordism spectrum $\\mathrm{MSL}$, with transfers along finite syntomic morphisms with trivialized canonical sheaf. Along the way, the paper shows that for any smooth stable tangential structure $\\beta \\colon B \\to K_{>0}$, the motivic Thom spectrum $M_\\beta$ is the framed suspension spectrum of the moduli stack $\\mathrm{FQSm}^\\beta_S$ of finite quasi-smooth derived $S$-schemes with $\\beta$-structure; in particular, over a regular equicharacteristic base, $\\Omega^\\infty_{\\mathbb{P}^1}\\mathrm{MGL}$ is the $\\mathbb{A}^1$-homotopy type of the moduli stack of virtual finite flat local complete intersections.","pith_inferences":["The framed-correspondence machinery used for MGL and MSL should also produce transfer descriptions for modules over other motivic Thom ring spectra attached to smooth stable tangential structures, for example symplectic or quadratic Grothendieck–Witt structures.","Because the moduli-stack and Hilbert-scheme models are explicit, algebraic cobordism computations could be attacked by studying group completions of Hilbert schemes of finite local complete intersections, a more geometric route than working with the formal spectrum $\\mathrm{MGL}$.","A further question the paper leaves open is whether the module equivalence is compatible with the six-functor formalism; if it were, all $\\mathrm{MGL}$-modules would inherit Gysin transfers and duality for finite syntomic morphisms automatically."],"forward_implications":["Every $\\mathrm{MGL}$-module carries coherent finite syntomic transfers, and this transfer structure is sufficient to characterize $\\mathrm{MGL}$-modules among motivic spectra.","Over a perfect field, very effective $\\mathrm{MGL}$-modules are equivalent to grouplike motivic spaces with finite syntomic transfers, giving a cancellation theorem for the suspension functor on $\\mathbb{A}^1$-invariant sheaves with finite syntomic transfers.","The infinite $\\mathbb{P}^1$-loop space $\\Omega^\\infty_{\\mathbb{T}}\\mathrm{MGL}$ over a pro-smooth base over a field is the group completion of the moduli stack of finite syntomic schemes, and over a field it has a smooth Hilbert-scheme model classifying finite local complete intersections in $\\mathbb{A}^\\infty$.","For $n>0$, $\\Omega^\\infty_{\\mathbb{T}}\\Sigma^n_{\\mathbb{T}}\\mathrm{MGL}$ is the $\\mathbb{A}^1$-homotopy type of the moduli stack of finite quasi-smooth derived schemes of virtual dimension $-n$.","Motivic cohomology $\\mathrm{H}\\mathbb{Z}$ is an $\\mathrm{MGL}$-module and is equivalent to the suspension spectrum of the constant sheaf $\\mathbb{Z}$ with its canonical finite syntomic transfers."],"supporting_citations":[{"why":"supplies the symmetric monoidal $\\infty$-category of tangentially framed correspondences and the motivic recognition principle over perfect fields.","marker":"[EHK+19]"},{"why":"supplies the reconstruction equivalence $\\mathrm{SH}(S)\\simeq\\mathrm{SH}^{\\mathrm{fr}}(S)$ that underlies the framed-spectrum descriptions.","marker":"[Hoy20]"},{"why":"supplies the comparison of equationally framed correspondences for vector bundles over infinite fields used to prove the Thom-spectrum theorem.","marker":"[GNP18]"},{"why":"supplies the formalism of motivic Thom spectra and the motivic $J$-homomorphism used to define $M_\\beta$.","marker":"[BH20]"},{"why":"removes the perfectness assumption in the Garkusha–Neshitov–Panin theorem, extending the comparison to arbitrary infinite fields.","marker":"[Dru20]"},{"why":"provides Gysin transfers that support the conditional geometric description of $\\Omega^\\infty_{\\mathbb{T}}M_\\beta$ and the treatment of finite syntomic transfers.","marker":"[DJK20]"}],"fun_headline_variants":["MGL-modules equal finite-syntomic-transfer spectra","Algebraic cobordism modules are transfer spectra","Equivalence: MGL-modules and syntomic transfers","MGL-modules: a moduli-stack viewpoint","Finite syntomic transfers classify MGL-modules"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof for arbitrary base schemes rests on a technical fact: algebraic $K$-theory and the $K$-theory summands used to build $\\mathrm{MGL}$ are determined by their values on smooth algebras, so that the twisted framed correspondences satisfy the required base-change comparison; if that fact failed for any of these summands, the equivalence would only be known over fields.","fun_headline_variants_meta":{"raw":{"variants":["MGL-modules equal finite-syntomic-transfer spectra","Algebraic cobordism modules are transfer spectra","Equivalence: MGL-modules and syntomic transfers","MGL-modules: a moduli-stack viewpoint","Finite syntomic transfers classify MGL-modules"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000646,"raw_usage":{"total_tokens":2996,"prompt_tokens":1000,"completion_tokens":1996,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":616,"completion_tokens_details":{"reasoning_tokens":1918}},"tokens_in":616,"tokens_out":1996,"duration_ms":24599,"temperature":1.0,"reasoning_tokens":1918,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:52:04.880516+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Over a perfect field $k$, the theorem predicts that the category of grouplike $\\mathbb{A}^1$-invariant sheaves with finite syntomic transfers is prestable and that the $\\mathbb{G}_m$-suspension functor is fully faithful; a direct way to falsify the theorem would be to find two non-isomorphic such sheaves whose images become equivalent after one suspension with respect to $\\mathbb{G}_m$, since the claimed cancellation theorem forbids this.","supporting_citations":[],"review_version":1}