{"id":"ba06af3a-18ba-4099-aef0-f053943c024c","arxiv_id":"2412.06765","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The boundary of a totally geodesic subvariety of moduli space is itself totally geodesic in each Deligne-Mumford boundary stratum and decomposes into prime pieces with locally isometric projections.","lead":"Mathematicians study totally geodesic subspaces of moduli space, where any two points can be joined by a shortest path that stays inside the subspace. This paper proves that when such a subspace is extended to the boundary of moduli space, the boundary pieces are again totally geodesic and split into simple factors, a structural tool for classification.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.2 is the load-bearing black box: the proof of the quadratic boundary-linearity theorem is only sketched, and the dπ step used to transfer Abelian level-wise linearity is not justified in the text.","rationale":"The reader identified Theorem 5.2 as the most fragile load-bearing premise, and my reading agrees. The theorem is used at the crucial juncture where boundary limits are shown to be level-wise linear (Lemma 7.5), which underpins the GL(2,R)-invariance of the boundary locus (Lemma 7.13) and hence the GL(2,R)-geodesic property (Proposition 7.18). The proof of Theorem 5.2 is only a sketch, and the step 'apply dπ in [CMS23, Lemma 7.2]' is not expanded: it is not shown that dπ is linear on each level in period coordinates, nor that it maps the Abelian level-wise linear locus onto the quadratic boundary stratum. This is a genuine correctness risk rather than a disagreement with consensus: the rest of the paper's architecture is coherent, and the main theorem is plausible, but the stated proof does not yet establish the key quadratic boundary-linearity result. The concurrent independent work by Arana-Herrera and Wright, which the paper acknowledges, suggests the overall theorem is true, but that does not remove the need for a complete proof of Theorem 5.2 in this paper. I therefore keep the reader's CONDITIONAL verdict, finding no reason to strengthen or weaken it based on this review.","tokens_in":34546,"tokens_out":7663,"duration_ms":86499,"concrete_test":"Write out the local coordinate expression of the map dπ from [CMS23, Lemma 7.2] on a two-level boundary stratum of the multi-scale quadratic compactification, and independently verify Theorem 5.2 for a concrete example. For instance, take Q to be the full quadratic stratum Q(4) (or the locus of squares of Abelian differentials) and compute its closure in ΞQ(4) intersected with a two-level boundary stratum, checking directly that the closure is level-wise linear in period coordinates. Then repeat with a non-full linear subvariety, e.g. the closure of a Teichmüller curve in Q(4) meeting the boundary, and confirm that its image under dπ is cut out by real-linear equations level by level. If dπ fails to be linear on some level or fails to be surjective onto the quadratic boundary stratum, Theorem 5.2 does not follow from [Ben23] alone.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorem reduces totally geodesy of ˚∂N to the GL(2,R)-geodesic property (Proposition 7.18), and the key input is level-wise linearity of closures in the multi-scale quadratic boundary. Lemma 7.5 uses Theorem 5.2 directly to conclude b(φ(q)) ∈ T∂N, and Lemma 7.13 relies on the same structural control to prove GL(2,R)-invariance of ˚∂QN′. If Theorem 5.2 fails, the proof of Proposition 7.18, and hence Theorem 1.4, collapses. The text does not prove Theorem 5.2; it gives a sketch: lift Q to an Abelian stratum via the holonomy double cover, apply Benirschke's Theorem 5.1 to the image M, then 'apply the map dπ in [CMS23, Lemma 7.2]' to conclude that the image is level-wise linear and equals Q ∩ DΓ. Two concrete gaps appear. First, level-wise linearity is not obviously preserved under dπ: one must show dπ is linear in level-wise period coordinates and maps each level's linear subspace to a linear subspace, with the level filtration respected. Second, the equality dπ(M ∩ DΓ′) = Q ∩ DΓ requires knowing which boundary strata occur in the image and that dπ is surjective onto the relevant quadratic boundary stratum; this is asserted, not demonstrated. Since the proof of Theorem 5.2 is the only place where the quadratic multi-scale boundary is controlled, the main theorem is conditional on completing this argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a structural result for totally geodesic subvarieties of the moduli space of curves: if N is a totally geodesic complex algebraic subvariety of M_{g,n}, then any irreducible component of the intersection of its boundary with a Deligne-Mumford boundary stratum is itself totally geodesic in the natural product sense (Theorem 1.4). It also proves that the boundary locus decomposes into prime factors whose projections to the factor moduli spaces are locally injective in an orbifold sense (Theorem 1.7) and locally isometric for the L^p Teichmüller metrics with 1<p<∞ (Proposition 9.2). The proof strategy is to pass from N to the set Q_N of quadratic differentials generating Teichmüller geodesics inside N, show that a suitable boundary locus of Q_N is GL^+(2,R)-invariant and large, then use a prime-factor argument and an infinitesimal-to-global totally geodesic criterion. The central tool is a quadratic-differential version of Benirschke's boundary-linearity theorem, stated as Theorem 5.2 with only a sketched proof.","tokens_in":34876,"tokens_out":5298,"duration_ms":57871,"significance":"If the proof is completed, the result is significant: it gives a strong inductive structural property for totally geodesic subvarieties of moduli space, with potential applications to classification problems and to counting closed geodesics. The paper is well organized, states hypotheses explicitly, and cleanly separates the geometric construction of many invariant differentials from the linear-algebraic boundary control. It also gives a concrete covering example illustrating the boundary phenomena. The main results overlap with independent work of Arana-Herrera and Wright, which is acknowledged. The manuscript does not rely on fitted parameters or circular definitions, and the proof is not a restatement of known results; however, the main theorem currently rests on a sketched external theorem, so the significance is conditional on completing that argument.","major_comments":[{"comment":"Theorem 5.2 is load-bearing: Lemma 7.5 uses it to conclude that b(phi(q)) lies in T∂N, and Lemma 7.13 uses the same structural control to prove GL^+(2,R)-invariance of the boundary locus. The proof given is only a sketch and leaves two essential steps unjustified. First, the statement 'Apply the map dπ in [CMS23, Lemma 7.2]' assumes that level-wise linearity is preserved under dπ: one must show that dπ is linear in the relevant level-wise period coordinates, sends linear subspaces to linear subspaces, and respects the level filtration; none of this is proved or cited precisely. Second, the equality dπ(M ∩ DΓ′) = Q ∩ DΓ requires a description of which boundary strata occur in the image and a surjectivity statement for dπ onto the relevant quadratic boundary stratum; this is asserted without argument. Since Proposition 7.18 and hence Theorem 1.4 collapse if Theorem 5.2 fails, the manuscript must either provide a complete proof of Theorem 5.2 or supply a precise reference where the theorem is proved in the stated quadratic form.","section":"5.2, Theorem 5.2"},{"comment":"Even granting Theorem 5.2, Lemma 7.5 does not verify its hypotheses. Theorem 5.2 is stated for a C-linear subvariety Q of a stratum Q(κ), but Lemma 7.5 applies it to Y, the closure of Q_N in a boundary stratum. The paper establishes in Proposition 7.3 only that Q_N is a GL^+(2,R)-invariant algebraic subvariety; it does not prove, or cite a theorem proving, that Q_N intersected with a stratum is C-linear in multi-scale period coordinates, nor that the closure Y inherits this property. This is another gap in the chain leading to GL^+(2,R)-geodesicity, and it should be addressed explicitly.","section":"7.4, Lemma 7.5"},{"comment":"The proof of Proposition 4.3, which is used in Lemma 8.2 and Lemma 8.5 to control area ratios, is not fully justified. The image P(Q) under the holonomy double cover is only known to be an injective immersion, and Chevalley's theorem gives constructibility rather than algebraicity of the image. The argument that if P(Q) is not prime then Q factors as a product uses preimages P_i^{-1}(M_i) that are only defined after taking closures, and the claimed factorization of Q is not established. Since the constancy of area ratios is essential for the prime-factor argument in Section 8, Proposition 4.3 needs a complete proof or a reference that covers the non-proper injective-immersion case.","section":"4.3, Proposition 4.3"}],"minor_comments":[{"comment":"The reference [A W24] lists the arXiv identifier as 'arXiv:??'; this should be completed before publication.","section":"References"},{"comment":"In the definition of a prime subset of a multi-component moduli space, the sentence 'there exists a of N to the product cover' appears to be missing a word or phrase and should be rewritten.","section":"1.2"},{"comment":"The statement of Proposition 4.3 refers to a 'stratum of multi-component translation surfaces' although the proof concerns quadratic differentials; the terminology should be aligned with the quadratic setting throughout the statement and proof.","section":"4.3"},{"comment":"The characterization in part (2) of Lemma 7.12 is worded awkwardly; it should say that the set consists of differentials whose product-cover components are zero exactly on the factors where all tangent vectors to the product cover of ∂N vanish.","section":"7.4, Lemma 7.12"},{"comment":"Theorem 5.2 uses the term 'C-linear subvariety' without giving a definition or a reference for the notion in the quadratic multi-scale setting; a definition or reference would help the reader verify the applicability in Lemma 7.5.","section":"5.2"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is attractive and the overall strategy is coherent, but the manuscript is currently conditional on a sketched theorem that is used in the central proof. I would be willing to reconsider after the authors either supply a complete proof of Theorem 5.2 or point to a precise published statement. The gaps in Proposition 4.3 should also be repaired, as that proposition feeds into the prime-factor arguments. The overlap with Arana-Herrera-Wright is acknowledged in the text; the incomplete arXiv number for that work should be fixed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe paper proves something that was missing from the literature: the Deligne–Mumford boundary of a totally geodesic subvariety of moduli space is itself totally geodesic in every boundary stratum, with a prime decomposition and local isometries of projections. This is the natural analogue of the Mirzakhani–Wright boundary theorem for affine invariant manifolds, and as far as I can tell it is genuinely new. The independent work by Arana-Herrera and Wright is contemporaneous, not prior, and the authors acknowledge it properly.\n\nThe main proof strategy is interesting: they introduce a GL(2,R)-geodesic property, show the boundary locus has it via the real multi-scale compactification, and then convert that into total geodesy. The reduction from infinitesimally totally geodesic to totally geodesic in the multi-component setting (Section 3) is clean and useful. The extension of Chen–Wright's ratio-of-areas result to quadratic differentials (Proposition 4.3) is plausible and carefully argued. The exposition is generally clear.\n\nThe soft spot is Theorem 5.2, the quadratic version of Benirschke's boundary linearity theorem. The proof is only a sketch, and it is load-bearing: Lemma 7.5 and Lemma 7.13 use it directly to get level-wise linearity and GL(2,R)-invariance of the boundary locus. The sketch says: lift to the Abelian stratum via holonomy double cover, apply Theorem 5.1, then apply the map dπ from [CMS23, Lemma 7.2] to conclude. Two steps are asserted without proof. First, dπ should preserve level-wise linearity in the sense needed—that is not obvious and is not checked. Second, the equality dπ(M ∩ DΓ′) = Q ∩ DΓ requires a precise statement about which boundary strata appear in the image and surjectivity onto the relevant quadratic stratum. That equality is also asserted, not demonstrated. If Theorem 5.2 fails, the main theorem collapses. It may be that these gaps are fillable; they are technical rather than conceptual, but they are real and the text does not fill them.\n\nThe rest of the proof seems coherent to me. I did not find circularity or hidden fitting; the cited [Ben23] and [Ben24] are independent, published results. The example in Section 9 is a nice check on the sharpness of local injectivity.\n\nThis paper is for people working on totally geodesic subvarieties, orbit closures, and boundary compactifications. It deserves a serious referee. I would send it out with instructions to focus on Theorem 5.2, and I would want the referee to verify that the dπ step can be made rigorous. My own verdict would be conditional until that argument is written out, but the result is important enough to warrant referee time.","headline":"A genuinely new boundary theorem for totally geodesic subvarieties, with a real but likely fixable gap in the proof of the key quadratic linearity lemma.","tokens_in":35383,"tokens_out":3030,"would_cite":true,"duration_ms":29143,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32G15","14H10","30F60","37F34"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that the Deligne–Mumford boundary of any totally geodesic subvariety of moduli space is itself totally geodesic in each boundary stratum, and that each boundary locus decomposes into prime pieces whose projections to each…","keywords":["totally geodesic subvariety","moduli space of curves","Teichmüller metric","Deligne-Mumford boundary","quadratic differentials","multi-scale compactification","GL(2,R)-invariant subvariety","boundary linearity"],"falsifier":"A concrete check of the theorem would be to compute the boundary in the covering example of Section 1.1 and verify directly that every pair in $\\mathring{\\partial}N$ lies on a Teichmüller geodesic in $\\mathring{\\partial}N$; a single violation would refute Theorem 1.4. The more surgical falsifier is Theorem 5.2: find a C-linear subvariety of a quadratic-differential stratum whose closure meets a multi-scale boundary stratum in a non-level-wise-linear set, which would invalidate the proof's load-bearing step.","tokens_in":34370,"feed_emoji":"📐","tokens_out":8359,"duration_ms":78134,"temperature":0.7,"pith_summary":"The paper aims to show that total geodesicity is inherited by the Deligne–Mumford boundary. Precisely: if $N\\subset \\mathcal{M}_{g,n}$ is a subvariety in which every pair of points is joined by a Teichmüller geodesic lying in $N$, then each irreducible component $\\mathring{\\partial}N$ of the intersection of $\\partial N$ with a boundary stratum $\\Delta_\\Gamma$ is itself totally geodesic in $\\Delta_\\Gamma$, with geodesics allowed to travel at independent speeds in the different factors of the stratum. The authors further claim that each such boundary locus splits into prime factors, and that the projection of each prime factor to every moduli-space component is locally injective and, for the natural $L^p$ Teichmüller metrics with $1<p<\\infty$, a local isometry. The point of caring is that classification of totally geodesic subvarieties can then proceed inductively, cutting off boundary strata of lower complexity.","feed_headline":"Boundary of a totally geodesic subvariety is totally geodesic","feed_subtitle":"If a moduli subvariety is totally geodesic, each boundary piece inherits it, with prime factors projecting as local isometries.","key_machinery":"The load-bearing object is the pair consisting of the set $Q_N$ of quadratic differentials generating Teichmüller geodesics contained in $N$ and its boundary in the Hodge and real multi-scale compactifications. A boundary locus $\\mathring{\\partial}N$ is shown to be $\\mathrm{GL}^+(2,\\mathbb{R})$-geodesic: one can find a $\\mathrm{GL}^+(2,\\mathbb{R})$-invariant algebraic witness $Q$ of dimension at least $2\\dim \\mathring{\\partial}N$ projecting onto $\\mathring{\\partial}N$. Two tools carry the proof: constancy of ratios of areas on prime invariant subvarieties, extended from Abelian to quadratic differentials through the holonomy double cover, and the continuity of the $\\mathrm{GL}^+(2,\\mathbb{R})$-action on the real multi-scale space, which moves invariance of $Q_N$ onto its boundary locus.","core_discovery":"The paper's central claim is Theorem 1.4: a totally geodesic complex algebraic subvariety of $\\mathcal{M}_{g,n}$ has totally geodesic boundary in every boundary stratum, where the ambient boundary stratum is a multi-component moduli space with its multi-speed Teichmüller metric. Theorem 1.7 adds that the product cover of each boundary locus decomposes into prime totally geodesic pieces, and the projection of any prime piece to any factor moduli space is locally injective in the orbifold sense; Proposition 9.2 upgrades this to a local isometry for every $L^p$ Teichmüller metric with $1<p<\\infty$. The proof route is to convert the flatness of $N$ into a large algebraic family of quadratic differentials: the set $Q_N$ of differentials whose Teichmüller geodesics stay in $N$ is a $\\mathrm{GL}^+(2,\\mathbb{R})$-invariant subvariety, and the authors show that its boundary locus lying over $\\mathring{\\partial}N$ is again $\\mathrm{GL}^+(2,\\mathbb{R})$-invariant and has dimension large enough to generate the tangent space of $\\mathring{\\partial}N$; from that, every tangent direction is realized by a geodesic in the boundary, which is exactly the totally geodesic property.","pith_inferences":["Beyond the paper: the proof's setup works with multi-component moduli spaces throughout, so the same boundary-flatness conclusion should hold verbatim for totally geodesic subvarieties of products of moduli spaces; the authors use this implicitly but do not list it as a separate theorem.","Beyond the paper: if Theorem 1.7's local injectivity were asked to be global, Example 9.3 shows the statement would be false for arbitrary $\\mathrm{GL}^+(2,\\mathbb{R})$-geodesic subvarieties, so the orbifold-local formulation is likely the sharp one; this suggests the boundary pieces behave like local graphs over each factor, not global ones.","Beyond the paper: the reliance on the real multi-scale compactification suggests an analogous boundary-flatness result may hold for $\\mathrm{SL}(2,\\mathbb{R})$-orbit closures in strata of quadratic differentials, since the same continuity of the $\\mathrm{GL}^+(2,\\mathbb{R})$-action is the decisive input; verifying this would be a direct test of the method's reach."],"forward_implications":["Every boundary piece of a totally geodesic subvariety is itself a totally geodesic subvariety of a lower-complexity moduli space, so induction over the stratification of $\\overline{\\mathcal{M}}_{g,n}$ becomes available.","The prime decomposition of Theorem 1.7 bounds how complicated a boundary piece can be: over each factor it is locally injective and locally isometric, so it looks like a local graph rather than a spread-out family.","The local-isometry statement of Proposition 9.2 holds simultaneously for all $L^p$ Teichmüller metrics with $1<p<\\infty$, which pins down the metric content of the boundary projection.","The construction of the witness $Q$ gives a concrete algebraic family of differentials on the boundary, a tool that can be used to count closed geodesics on $N$ inside its boundary strata."],"supporting_citations":[{"why":"Gives constancy of ratios of areas on prime invariant subvarieties of multi-component translation surfaces, which Proposition 4.3 extends to quadratic differentials and Lemma 8.2 uses to control the map from differentials to tangent vectors.","marker":"[CW21]"},{"why":"Supplies Theorem 5.1, the Abelian boundary-linearity theorem that the quadratic version (Theorem 5.2) is sketched from.","marker":"[Ben23]"},{"why":"Provides the map $d\\pi$ in Lemma 7.2 used in the sketch of Theorem 5.2 to transfer level-wise linearity from Abelian to quadratic strata.","marker":"[CMS23]"},{"why":"Constructs the real multi-scale compactification and the continuous extension of the $\\mathrm{GL}^+(2,\\mathbb{R})$-action, used in Lemma 7.13 to get invariance of the boundary locus.","marker":"[BCG+19b]"},{"why":"Proposition 2.1 is used to show that sums of boundary quadratic differentials in $Q_N$ remain in $Q_N$ during the inductive construction of limits.","marker":"[Ben24]"},{"why":"With [Fil16], cited in Proposition 7.3 to conclude that the set $Q_N$ is an algebraic subvariety.","marker":"[EMM15]"},{"why":"Together with [EMM15], used in Proposition 7.3 to establish algebraicity of $Q_N$.","marker":"[Fil16]"},{"why":"Lemma 1 supplies the holonomy double cover used in Proposition 4.1 to carry the ratio-of-areas result from Abelian to quadratic differentials.","marker":"[KZ03]"}],"fun_headline_variants":["Totally geodesic boundary for totally geodesic subvarieties","Boundary of totally geodesic subvariety: prime pieces, locally isometric","Totally geodesic subvariety's boundary: prime and totally geodesic"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof relies on a theorem, given only as a sketch here, that says boundary limits of certain families of flat surfaces are structured linearly; if that theorem is false, the main argument breaks.","fun_headline_variants_meta":{"raw":{"variants":["Totally geodesic boundary for totally geodesic subvarieties","Boundary of totally geodesic subvariety: prime pieces, locally isometric","Totally geodesic subvariety's boundary: prime and totally geodesic"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002552,"raw_usage":{"total_tokens":9779,"prompt_tokens":954,"completion_tokens":8825,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":570,"completion_tokens_details":{"reasoning_tokens":8762}},"tokens_in":570,"tokens_out":8825,"duration_ms":58549,"temperature":1.0,"reasoning_tokens":8762,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T19:19:07.060914+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check of the theorem would be to compute the boundary in the covering example of Section 1.1 and verify directly that every pair in $\\mathring{\\partial}N$ lies on a Teichmüller geodesic in $\\mathring{\\partial}N$; a single violation would refute Theorem 1.4. The more surgical falsifier is Theorem 5.2: find a C-linear subvariety of a quadratic-differential stratum whose closure meets a multi-scale boundary stratum in a non-level-wise-linear set, which would invalidate the proof's load-bearing step.","supporting_citations":[],"review_version":1}