{"id":"05a5758e-a66c-406e-9bc2-85c6d406fb5d","arxiv_id":"2412.03408","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A new class of twisted curves with colliding marked points is defined, their moduli stack is constructed, and a characterization shows some tame abelian orbicurves, including a μ4 example, do not arise from it.","lead":"This paper expands twisted curves, curves with extra group data at special points, so that marked points may collide. It builds a moduli space for the new objects and proves contraction maps extend automatically.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the μ4 counterexample in Example 7.26 is terse but its eigenspace argument can be justified, so the advertised characterization and main stack/contraction theorems stand.","rationale":"After a good-faith read, the central constructions are coherent. Theorem 2.31's proof that M^glt is a stack uses standard stack criteria and an open cover by M_{g,n}; the key openness of the Isom locus follows from Lemma 2.17 and 2.22 and appears sound. Theorem 6.15's initial contractions are the most intricate part; the construction of N^D as a fiber product in (6.19.1) is admissible because the arrows are injective and N^D is the preimage of a saturated, finitely generated submonoid in a Q-vector space, hence saturated and with finitely generated groupification. The universal property then follows from the saturation of M'_{D→S}. The reader's flagged μ4 argument is terse, but the concerning deduction is valid for the reasons in load_bearing_attack; the sentence about y/w is a shorthand for the contradiction y=w, while the character contradiction 1=3 gives the same conclusion even more directly. The unpublished [10] reference is a presentation issue, not a load-bearing correctness issue for the object-level characterization. Therefore I do not see a reason to change the reader's conditional verdict, though I would not elevate the μ4 issue to a correctness risk.","tokens_in":45771,"tokens_out":43218,"duration_ms":429665,"concrete_test":"Prove or disprove the missing basis lemma needed in Example 7.26: for an admissible inclusion N^r ⊂ N, the ideal generated by N^r in k[N] is monomial, hence the quotient k[N]/(N^r) has a k-basis indexed by the monomials not in the ideal. If the lemma holds, the one-dimensional eigenspaces force 2z1=z2 and 2z3=z2 in N, giving z1=z3 by torsion-freeness and contradicting classes 1≠3. If the lemma fails, construct an explicit N where distinct monomials become linearly dependent and check whether that N yields the algebra k[y,w]/(y^2-w^2,yw).","verdict_should_be":"UNCHANGED","load_bearing_attack":"I reviewed the weakest point highlighted by the reader: Example 7.26's claim that X=[Spec(k[y,w]/(y^2-w^2))/μ4] is not a quotient of a toric stack for any admissible monoid. The step the reader questions, deducing 2z1=z2 and 2z3=z2 in N from one-dimensional eigenspaces, is justified because the presentation (7.26.2) gives A = k[N]/(N^r) with a monomial ideal, so distinct monomial classes are k-linearly independent. A one-dimensional eigenspace therefore forces 2z1=z2 in N, and similarly 2z3=z2; saturatedness makes N^gp torsion-free, so z1=z3, contradicting the distinct classes 1 and 3 in N^gp/Z^r. The reliance on the unpublished [10, 7.10] only affects the 2-category versus 1-category framing of the functor, not the object-level essential image characterization in Corollary 7.22 or Theorems 1.3 and 6.15. I find no load-bearing correctness gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces generalized log twisted curves: marked nodal curves with possibly coinciding sections, together with a simple inclusion of log structures and an admissible sheaf of monoids at the markings. For distinct markings these recover Abramovich-Vistoli twisted curves. The paper proves that the resulting moduli stack M^glt_{g,n} is a smooth algebraic stack locally of finite type with quasi-compact and separated diagonal (Theorem 2.31), constructs a canonical initial lift of any contraction of coarse curves to generalized log twisted curves (Theorem 6.15), and characterizes which tame abelian nodal orbicurves arise from this construction (Corollary 7.22, Theorem 1.7). A concrete non-example with a μ4 stabilizer is given in Example 7.26.","tokens_in":1093,"tokens_out":2832,"duration_ms":144147,"significance":"If the results stand, the paper supplies a logarithmic and stacky framework for Hassett-style weighted curves with colliding markings and gives the deformation-theoretic foundations needed for the companion theory of twisted stable maps. Its strengths are the precise universal property in Theorem 6.15, the reduction of the moduli theorem to prior published foundations, and the falsifiable negative statement in Theorem 1.7/Example 7.26. I specifically checked the μ4 counterexample: the eigenspace computation is justified because the presentation (7.26.2) is by a monomial ideal, so one-dimensional eigenspaces force 2z1 = z2 and 2z3 = z2, and saturatedness of N makes N^gp torsion-free, giving the required contradiction. I therefore do not see a load-bearing correctness gap.","major_comments":[],"minor_comments":[{"comment":"The admissibility of N_D is verified with the sentence 'One can check (for example using 5.13) that all the monoids appearing in the solid part of (6.19.1) are fine and saturated.' Since this is a load-bearing step for Theorem 6.15, please expand this into a short argument or a lemma with the relevant stalk computation from 5.13.","section":"6.20"},{"comment":"The identification of the 2-category of tame abelian nodal orbicurves with a 1-category is delegated to the unpublished preprint [10, 7.10]. The object-level statements in Corollary 7.22 and Theorem 1.7 are independent of this identification, but the functor in 7.17.1 and the faithfulness/non-fullness discussion in 7.18 are stated in that language; please either provide the argument or explicitly state that the later essential-image results do not depend on [10].","section":"7.17"},{"comment":"In Example 7.26, the phrase 'the element y/w is not in the ring k[y,w]/(y^2-w^2)' is somewhat informal; the contradiction is clearer if phrased as: saturatedness forces z1 = z3, so the monomials y and w would be equal in the graded algebra A, contradicting the displayed eigenspace decomposition.","section":"7.26"},{"comment":"In the proof of Theorem 2.31, the phrase 'For any admissible monoid N ⊂ Nn' should read 'N ⊂ Q^n_{>=0}', since admissible monoids are submonoids of Q^n_{>=0}.","section":"2.31"},{"comment":"There are several typographical slips: 'folllowing' in the proof of 2.17, 'surjecive' in the proof of 4.3, and 'Alqvist' in the Introduction for the author of reference [6], whose name is spelled Ahlqvist there.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"This is a serious contribution that is within the scope of the journal. My main editorial concern is the reliance on the unpublished preprint [10] for the 1-categorical framing in Section 7; this should be made explicit or replaced by a self-contained argument. The terse verification in 6.20 should also be expanded. I do not see a correctness issue in the central theorems."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Know the punchline first: this is a solid, genuinely new piece of moduli theory. Generalized log twisted curves give a unified language for Hassett's weighted pointed curves and Abramovich–Vistoli twisted curves, and they handle colliding markings. The main theorems — smooth stack, initial contraction, essential image with the μ4 boundary — are new and, as far as I can tell, proven correctly.\n\nThe paper is careful about foundations. It reduces to published work by Borne–Vistoli, Olsson, and Abramovich–Olsson–Vistoli where it can, and the proofs of 2.31, 6.15, and 7.22 are detailed rather than hand-wavy. The μ4 counterexample (7.26) is genuinely nice: the eigenspace argument that the reader flagged is terse, but I checked the graded algebra computation and it is sound. The deduction 2z1 = z2 and 2z3 = z2 from one-dimensional eigenspaces is justified because the presentation as k[N]/(N^r) makes distinct monomial classes k-linearly independent; saturatedness then forces z1 = z3, contradicting the distinct classes. So the non-fullness claim stands.\n\nSoft spots, in proportion: the paper cites the unpublished [10] (Bragg–Olsson–Webb) for the statement that tame abelian nodal orbicurves form a 1-category. That is a real dependence, but it only affects the categorical framing of the functor, not the object-level characterization of the essential image (7.22), and the authors flag it. A couple of proofs are terse — the admissibility check in 6.20 defers to 5.13, and the distinct-marking agreement with Abramovich–Vistoli is via [4, A.5] rather than reproved. Those are confidence issues, not correctness issues. There is also a lot of technical bookkeeping with monoids in Section 2, but that is the nature of the subject.\n\nWho is this for? Anyone working on moduli of stable maps to tame stacks, or on logarithmic/weighted curves. It is not a paper that will change how we think about curves in general, but it genuinely advances the subfield. It deserves a serious referee. I would send it out rather than desk-reject; my expectation is that the referee will ask for a bit more detail in Section 7 and a careful treatment of the [10] dependency, but the central mathematics is in good shape.","headline":"Solid new theory of log twisted curves with colliding markings; the μ4 counterexample works, and the paper deserves refereeing.","tokens_in":46534,"tokens_out":2476,"would_cite":true,"duration_ms":24521,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14A20","14H10","14D23"],"pacs":[],"model":"deepseek-v4-flash","headline":"Generalized log twisted curves allow marked points to collide while retaining stacky data, and their moduli form a smooth algebraic stack over the integers.","keywords":["generalized log twisted curves","admissible monoids","twisted curves","stacky curves","Hassett contractions","tame abelian nodal orbicurves","root stacks","log structures"],"falsifier":"Recompute the graded algebra $A = k[y,w]/(y^2-w^2, yw)$ with its $\\mu_4$ eigenspaces and check whether the identifications $2z_1 = z_2$ and $2z_3 = z_2$ force $z_1 - z_3$ to be a torsion element of $N^{\\mathrm{gp}}$ whose saturation contradicts the relation; if $z_1 - z_3$ is not torsion, or if $y/w$ actually lies in the ring $k[y,w]/(y^2-w^2)$, Example 7.26 collapses.","tokens_in":45596,"feed_emoji":"📐","tokens_out":10094,"duration_ms":88971,"temperature":0.7,"pith_summary":"Generalized log twisted curves are marked nodal curves whose sections may land on top of one another, equipped with a sheaf of admissible monoids and a simple extension of the base log structure. The paper shows these objects form a smooth algebraic stack locally of finite type over $\\mathbb{Z}$, and that the earlier theory of twisted curves with distinct markings is recovered exactly when the sections are disjoint. It also proves that any contraction of the underlying coarse curve extends to a canonical initial contraction of generalized log twisted curves, giving a lifting of the weighted-contraction procedure to the stacky setting. Finally, it characterizes which tame abelian nodal orbicurves arise from generalized log twisted curves: the local monoid at each marked point must be a pushout of an admissible monoid, and all $\\mu_2$ and $\\mu_3$ examples do arise while an explicit $\\mu_4$ orbicurve does not.","feed_headline":"Colliding marked points tamed by log twisted curves","feed_subtitle":"A moduli stack over Z handles colliding sections, with canonical contractions and a precise image theorem.","key_machinery":"The load-bearing object is an admissible monoid $N \\subset \\mathbb{Q}^n_{\\ge 0}$: a finitely generated saturated submonoid containing $\\mathbb{N}^n$, equivalently described by an admissible subgroup of $\\mathbb{Q}^n$. Packaged as a sheaf on the curve, $N$ encodes the amount of rooting at possibly coinciding sections. The associated stack $\\mathcal{C} = \\mathcal{C}_{\\mathrm{node}} \\times_C \\mathcal{C}_N$ combines the node stack obtained from the simple log inclusion with the stack of roots $\\mathcal{C}_N$ built from $N$ via the system of denominators $\\bigoplus s_{i,\\ast}\\mathbb{N} \\to N$; locally $\\mathcal{C}_N$ is the toric quotient $[\\operatorname{Spec}(R \\otimes_{\\mathbb{Z}[\\mathbb{N}^m]} \\mathbb{Z}[N]) / D(N^{\\mathrm{gp}}/\\mathbb{Z}^m)]$. The contraction theorem is carried by the canonical log enhancement of a coarse contraction and by the initial-contraction construction in which the new base log structure is defined as a saturation and the new admissible monoid as a fiber product, making the universal property hold.","core_discovery":"The central discovery is that the extra freedom needed to allow coinciding markings is precisely a global admissible sheaf of monoids $N \\subset \\bigoplus s_{i,\\ast}\\mathbb{Q}_{\\ge 0}$, together with a simple extension of the base log structure; from these data one canonically constructs a tame Artin stack $\\mathcal{C}$ with coarse space $C$. The functor taking a generalized log twisted curve to its associated stack is faithful but not full, and the paper determines its essential image: a tame abelian nodal orbicurve arises this way iff at every smooth point with nontrivial stabilizer the local monoid $D_x$ is a pushout $\\mathbb{N}^m \\oplus_{\\mathbb{N}^m} N$ of an admissible monoid. Consequently every $\\mu_2$ and $\\mu_3$ tame abelian nodal orbicurve lies in the image, while the stack $[\\operatorname{Spec}(k[y,w]/(y^2-w^2))/\\mu_4]$ does not. On the moduli side, the fibered category $\\mathcal{M}^{\\mathrm{glt}}_{g,n}$ is a smooth algebraic stack locally of finite type over $\\mathbb{Z}$ with a Zariski open cover by standard twisted-curve stacks, and contractions of coarse curves lift to universal initial contractions of generalized log twisted curves.","pith_inferences":["This suggests a testable classification of the essential image by finite abelian groups: for a fixed stabilizer group $G$, the obstruction to being an admissible pushout should be computable from the eigenspace decomposition, and $\\mu_4$ should be the first in an infinite family of forbidden quotients.","If the companion article's program succeeds, weighted stable maps from these curves should form proper moduli stacks whose boundary points encode exactly the colliding-marking configurations that admissible monoids describe.","The non-fullness examples imply that enumerative invariants counting maps from these curves should be built from the generalized log twisted curve itself, not merely from its associated orbicurve.","One can test the contraction formalism on explicit one-parameter families: take a weighted stable family where two markings collide and compare the initial contraction of Theorem 6.15 with the relative coarse space of the associated stack; they should agree."],"forward_implications":["Because $\\mathcal{M}^{\\mathrm{glt}}_{g,n}$ is a smooth algebraic stack locally of finite type over $\\mathbb{Z}$ with quasi-compact and separated diagonal, the usual deformation-theoretic and descent machinery applies to families of generalized log twisted curves.","Any contraction of the underlying coarse curve has a universal initial lift to generalized log twisted curves, so contractions can be performed compatibly with stacky and logarithmic data without losing information.","When the marked sections are pairwise disjoint, generalized log twisted curves coincide with the classical twisted curves, so the new theory contains the earlier distinct-markings theory as a special case.","A tame abelian nodal orbicurve lies in the essential image exactly when each local monoid is an admissible pushout, so all $\\mu_2$ and $\\mu_3$ cases occur while the explicit $\\mu_4$ stack $[\\operatorname{Spec}(k[y,w]/(y^2-w^2))/\\mu_4]$ does not.","The associated-stack functor is faithful with finite fibers but not full, so the log data genuinely distinguish objects that the underlying orbicurve forgets."],"supporting_citations":[{"why":"Supplies the foundational constructions of twisted stable maps, relative coarse spaces, and the stack of log twisted curves that this paper extends.","marker":"[4]"},{"why":"Defines the twisted curves that this theory specializes to when the markings are distinct.","marker":"[5]"},{"why":"Provides the stack-of-roots construction $\\mathcal{C}_N$ and the Deligne-Faltings chart formalism used throughout.","marker":"[9]"},{"why":"Gives the statement that tame abelian nodal orbicurves form a 1-category, against which the associated-stack functor's image is compared in Section 7.","marker":"[10]"},{"why":"Introduces weighted stable pointed curves and the coarse-curve contraction picture that Section 6 lifts to generalized log twisted curves.","marker":"[11]"},{"why":"Supplies the theory of fine and saturated monoids, integral inclusions, and log blowups used for admissible monoids and log contraction diagrams.","marker":"[14]"},{"why":"Gives the canonical log smooth structure on a family of nodal curves used to define the base log structures.","marker":"[15]"},{"why":"Defines simple inclusions of log structures and the associated node stack $\\mathcal{C}_{\\mathrm{node}}$, along with the log twisted-curve category this paper extends.","marker":"[16]"},{"why":"Provides the Stacks Project results used for prestable curves, cohomology and base change, and contractions of rational bridges and tails.","marker":"[19]"}],"fun_headline_variants":["Log twisted curves handle colliding markings","Colliding points get new moduli stack","New curves for coincident marked points","Log stacks tame colliding marked points","Moduli for curves with merging sections"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"If the eigenspace calculation in Example 7.26 showing that the $\\mu_4$ stack $[\\operatorname{Spec}(k[y,w]/(y^2-w^2))/\\mu_4]$ cannot be written as a quotient of an admissible toric stack is wrong, then the claimed boundary of the essential image and the non-fullness of the functor are not established.","fun_headline_variants_meta":{"raw":{"variants":["Log twisted curves handle colliding markings","Colliding points get new moduli stack","New curves for coincident marked points","Log stacks tame colliding marked points","Moduli for curves with merging sections"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000195,"raw_usage":{"total_tokens":1329,"prompt_tokens":887,"completion_tokens":442,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":503,"completion_tokens_details":{"reasoning_tokens":380}},"tokens_in":503,"tokens_out":442,"duration_ms":4511,"temperature":1.0,"reasoning_tokens":380,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T22:28:00.784011+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the graded algebra $A = k[y,w]/(y^2-w^2, yw)$ with its $\\mu_4$ eigenspaces and check whether the identifications $2z_1 = z_2$ and $2z_3 = z_2$ force $z_1 - z_3$ to be a torsion element of $N^{\\mathrm{gp}}$ whose saturation contradicts the relation; if $z_1 - z_3$ is not torsion, or if $y/w$ actually lies in the ring $k[y,w]/(y^2-w^2)$, Example 7.26 collapses.","supporting_citations":[{"cited_title":"Algebraic Geom","cited_arxiv_id":null,"evidence_quote":"Supplies the foundational constructions of twisted stable maps, relative coarse spaces, and the stack of log twisted curves that this paper extends."},{"cited_title":"Abramovich and A","cited_arxiv_id":null,"evidence_quote":"Defines the twisted curves that this theory specializes to when the markings are distinct."},{"cited_title":"Borne and A","cited_arxiv_id":null,"evidence_quote":"Provides the stack-of-roots construction $\\mathcal{C}_N$ and the Deligne-Faltings chart formalism used throughout."},{"cited_title":"Bragg, M","cited_arxiv_id":null,"evidence_quote":"Gives the statement that tame abelian nodal orbicurves form a 1-category, against which the associated-stack functor's image is compared in Section 7."},{"cited_title":"Hassett, Moduli spaces of weighted pointed stable curves, Adv","cited_arxiv_id":null,"evidence_quote":"Introduces weighted stable pointed curves and the coarse-curve contraction picture that Section 6 lifts to generalized log twisted curves."},{"cited_title":"Ogus, Lectures on logarithmic algebraic geometry, Cambridge Studies in Advanced Mathematics, vol","cited_arxiv_id":null,"evidence_quote":"Supplies the theory of fine and saturated monoids, integral inclusions, and log blowups used for admissible monoids and log contraction diagrams."},{"cited_title":"Olsson, Universal log structures on semi-stable varieties, Tohoku Math","cited_arxiv_id":null,"evidence_quote":"Gives the canonical log smooth structure on a family of nodal curves used to define the base log structures."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines simple inclusions of log structures and the associated node stack $\\mathcal{C}_{\\mathrm{node}}$, along with the log twisted-curve category this paper extends."}],"review_version":1}