{"id":"6098b56a-be6c-4a1c-a8db-adcc9e3fb0a0","arxiv_id":"1908.07436","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The WYSIWYG compactification is not an algebraic variety or complex analytic space, yet it supports an unconditional proof of the multi-component boundary tangent space formula for GL(2,R) orbit closures.","lead":"This paper proves that a natural geometric compactification of flat surfaces, the WYSIWYG space, cannot be given an algebraic or complex analytic structure. It also provides a complete, unconditional proof of a central formula for the boundaries of GL(2,R) orbit closures, a tool used to classify these dynamical systems.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.2's unconditional proof rests on Theorem 6.4, whose key period-extension claim is delegated to the BCG+ multi-scale compactification; omitted finite-cover and higher-polar-order details are the weakest link.","rationale":"I read the paper in good faith and found the main line of argument coherent. The strongest claim, Theorem 1.2, is proved by reducing to Lemma 6.5 and Lemma 6.7, which depend on Theorem 6.4. The proof of Theorem 6.4 is not self-contained: it relies on the BCG+ construction of the moduli space of multi-scale differentials, specifically on the existence of continuous perturbed period coordinates and on the smooth orbifold structure. The paper itself flags two omitted technical details in that section, and the continuity of the node-crossing terms r_{h,l}(s) is asserted on the authority of [BCG+]. This is a genuine load-bearing dependency, because a failure there would invalidate the unconditional proof of the multi-component boundary formula. However, this is an external dependency rather than an internal inconsistency, and the rest of the paper, including Theorem 1.1 and Theorem 1.3, appears well-supported by the arguments given. I agree with the reader's identification of the weakest assumption; I would make the acceptance conditional on verification of the BCG+ machinery in the specific cases used here. The proposed concrete test directly checks the most delicate part of that machinery, and a successful check would justify upgrading the verdict to ACCEPT.","tokens_in":25870,"tokens_out":32035,"duration_ms":299731,"concrete_test":"Verify the period-extension formula of Theorem 8.1 for a concrete two-level degeneration: one lower-level component whose twisted differential has a pole of order 3 at a vertical node, and a top-level component with a simple zero there. Using the plumbing and perturbed period coordinates of [BCG+, Section 11.2], write the smoothing parameter t(s), the ratios λ_k(s), and the node-crossing term r_{h,l}(s) for a homology class whose representative passes through the node. Check explicitly that r_{h,l}(s) tends to 0 as s→0 and that the formula for the period is continuous. Also confirm that the finite-cover lift used to define the sectors S_j exists and that the period functions descend to the quotient. If both checks pass, the dependency is sound; if not, Theorem 6.4(3) and hence Theorem 1.2 are in jeopardy.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The core of the unconditional proof of Theorem 1.2 is Lemma 6.5, which asserts that the limit differential ω∞ lies in T(M)∩Ann(V). The proof of Lemma 6.5 requires Theorem 6.4, specifically property (3): relative and absolute periods of the family over each simply connected set Si extend continuously to the boundary point. Theorem 6.4 is proved in Section 8 by passing to the BCG+ compactification ıH(κ) and invoking: (a) ıH(κ) is a smooth complex orbifold; (b) the 'perturbed period coordinates' of [BCG+, Section 11.2] give continuous period functions; and (c) the terms r_{h,l}(s) coming from homology classes crossing nodes converge to zero as the smoothing parameter s→0. Each of these is cited from the preprint [BCG+] rather than proved here. The paper explicitly flags two omitted technical details: the passage to a finite cover to avoid orbifold issues ('we omit this distinction here', Section 8), and the adjustment t_j^{a_j} for polar nodes of pole order greater than two (footnote 4). If the period-extension statement has a hidden failure — for example, if r_{h,l}(s) does not converge when the lower-level differential has a higher-order pole — then Theorem 6.4(3) fails, Lemma 6.5 and Lemma 6.7 do not follow, and the multi-component case of Theorem 1.2 is not proved. The proof is otherwise coherent; the Cylinder Finiteness Theorem (5.3) is presented as an outline but builds on the published [MW17, Theorem 5.1].","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the WYSIWYG partial compactification ~H(κ) of strata of Abelian differentials, continuing the program of Mirzakhani and Wright. Its first main result, Theorem 1.1, shows that for g ≥ 3 with positive κ and for g = 2, κ = (1,1), the projectivized space P~H(κ) is neither an algebraic variety nor a complex analytic space in any structure for which the natural map from PH(κ) is a morphism. The second main result, Theorem 1.2, gives an unconditional proof of the Mirzakhani–Wright formula for the tangent space to the boundary of a GL+(2,R) orbit closure: the boundary in the finite-area locus is an algebraic variety locally described by finitely many linear subspaces in period coordinates, and the tangent space to a branch of the boundary equals the intersection of the tangent space of a branch of the orbit closure with the tangent space of the boundary stratum. The previous multi-component case was conditional on a conjectural generalization of the Eskin–Mirzakhani–Mohammadi isolation results; here the proof uses Filip's algebraicity of orbit closures and the BCG+ multi-scale compactification instead. A third result, Theorem 1.3, establishes structural properties of prime invariant subvarieties of multi-component strata, including that the absolute periods of one component locally determine those of any other component.","tokens_in":26188,"tokens_out":6140,"duration_ms":64778,"significance":"If the results stand, the paper makes a substantial advance: it removes a long-standing conditional assumption from the boundary theory of affine invariant manifolds, thereby sharpening the available tools for the classification of orbit closures. The non-algebraicity theorem is a striking and clearly explained negative result that separates the WYSIWYG compactification from algebro-geometric compactifications. The paper is careful in crediting prior work and in isolating the new ingredients: Filip's algebraicity, the BCG+ compactification, and the foundational properties of ~H from Mirzakhani–Wright. The Cylinder Finiteness Theorem is used in an essential way and is presented as an outline; Theorem 6.4 relies on the cited preprint [BCG+] for the construction of the multi-scale compactification and perturbed period coordinates. The paper is generally well written, with explicit proofs for the main new steps, including the non-algebraicity argument, the reduction to bounded cylinder modulus, and the structure theorem for prime subvarieties. The inclusion of cautionary examples is a service to the reader and sharpens the statement of Theorem 1.2.","major_comments":[{"comment":"The proof of Theorem 6.4, property (3), is the load-bearing step for Lemma 6.5 and hence for the unconditional proof of Theorem 1.2 in the multi-component case. This proof relies entirely on the cited preprint [BCG+] for the smooth orbifold structure of the multi-scale compactification and for the continuity of perturbed period coordinates, and it explicitly omits two technical details: the passage to a finite cover to avoid orbifold issues ('we omit this distinction here') and the adjustment of the smoothing parameters to powers t_j^{a_j} when polar nodes of pole order greater than two occur. Because a failure of the period-extension statement would break Lemma 6.5, I ask the authors to provide a complete verification of these two details or, failing that, to state clearly that Theorem 6.4 is conditional on the precise statements in [BCG+] and adjust the word 'unconditional' accordingly.","section":"§8, proof of Theorem 6.4 and footnote 4"},{"comment":"The proof of Theorem 5.3 is only an outline, and the statement is slightly stronger than [MW17, Theorem 5.1]. The key step that 'There are also only finitely many equations fixing the ratios of these large modulus cylinders' is asserted without proof, and the earlier warning in Remark 5.5 that the limit of cylinder deformations may lie in a higher-codimension subvariety of the boundary makes this assertion non-obvious. Since Theorem 5.3 feeds directly into Lemma 5.6 and Theorem 5.2, and Theorem 5.2 is used to pass to the bounded-modulus situation in Corollary 6.3 and Lemma 6.5, I request a complete proof of Theorem 5.3 or a precise reference to a complete proof of the strengthened statement.","section":"§5, Theorem 5.3 (Cylinder Finiteness Theorem)"},{"comment":"For disconnected strata, the proof says 'The other cases are similar and we omit the details', after discussing hyperelliptic and spin components for H(2g−2). Since the theorem asserts the result for each connected component, including the spin components where the spin parity of the attached elliptic tails must be controlled, I ask for a few sentences explaining why the same construction applies to all the remaining components and how the parity condition is preserved under the degeneration used.","section":"§3, proof of Theorem 1.1"}],"minor_comments":[{"comment":"The notation 'H1(X, Σ) ≃ H1(X′, Σ′)' is used to identify relative homology, but the collapse maps fn are only defined up to automorphisms; it would help to say explicitly that the identification is well-defined on the level of homology classes, as is done later in the paragraph.","section":"§2, Corollary 2.4"},{"comment":"The subscript in 'µϵ' should be 'ε' rather than a Greek epsilon in a different font; the same symbol is used inconsistently in the surrounding text.","section":"§7.1, proof of Lemma 7.2"},{"comment":"The notation 'γ∗i ∈ H1(X, Σ)' uses the Poincaré dual of the core curve; it may be worth reminding the reader that this is a relative class and that the identification 'H1(X, Σ) ≃ H1(X′, Σ′)' from Section 2 is being used implicitly.","section":"§5, Remark 5.4"},{"comment":"The symbol for the multi-scale compactification is introduced as 'ıH(κ)' but the text also uses 'ÙH(κ)' in the footnote about notation; please unify the typography.","section":"§8, first paragraph"},{"comment":"The proof invokes the Cylinder Deformation Theorem before it is stated in Section 5; a forward reference would improve readability.","section":"§4, Lemma 4.1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's main dependency is on the unpublished preprint [BCG+] for the construction of the multi-scale compactification and perturbed period coordinates. This is a legitimate citation, and the preprint is by the same research group and available on arXiv, so there is no question of access. My recommendation of major revision is driven by the two omitted technical details in the proof of Theorem 6.4 and by the outline status of Theorem 5.3; both are load-bearing for the central claim. If the editor determines that [BCG+] has been or will be accepted in a form that covers exactly these points, then the changes I request could be made in a minor revision, but in the current state of the manuscript I would like to see the omitted details addressed explicitly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main event is Theorem 1.2: the multi-component case of the Mirzakhani–Wright boundary tangent space formula is now proved unconditionally, and the proof is genuinely different from the conditional one in [MW17]. That alone makes this paper important. The route through Filip's algebraicity is much shorter, and the paper is honest about what it uses. Theorem 1.1, the non-algebraicity and non-analyticity of the WYSIWYG compactification, is a nice negative result that clarifies what kind of object ~H is. Theorem 1.3 adds structural facts about multi-component boundaries that people working toward orbit closure classification will want.\n\nThe paper is also well written in the places that matter. The cautionary examples in Section 4 are illuminating and show real care about where the naive picture breaks. The authors flag their own omissions — the 'other cases are similar' in Theorem 1.1, the outline for Theorem 5.3, the finite-cover and higher-pole-order details in Section 8. That is a good sign.\n\nSoft spots, in proportion: the proof of Theorem 6.4 depends on the BCG+ compactification being a smooth complex orbifold with continuous perturbed period coordinates. That is cited from a preprint rather than proved here. This is a real dependency, but it is an external foundation, not an ad hoc step, and the stress-test note's worry does not land as a load-bearing flaw. The omitted finite-cover distinction and the t_j^{a_j} adjustment for higher-order poles are technical; the authors explicitly say they are omitting them, and a referee could ask for a sentence or two more but not a reworking. Theorem 5.3 is presented as an outline, but it builds on published [MW17, Theorem 5.1] and the argument is plausible. None of this undermines the central theorem.\n\nThe citation pattern is appropriate. The paper leans on the authors' own prior work with Mirzakhani, but that is the natural toolset for this problem, and the target formula is not assumed.\n\nWho this is for: specialists in flat geometry and orbit closures. A broader dynamics audience gets the punchline but not the machinery. My bottom line: this deserves peer review. I would send it out, with the expectation of minor revision focused on making the BCG+ imports explicit and filling the small gaps the authors themselves identify.","headline":"Removes the conditional from the multi-component boundary tangent space formula using a genuinely different, shorter route through Filip's algebraicity; the soft spots are technical dependencies on the BCG+ compactification, not hidden errors.","tokens_in":26764,"tokens_out":1479,"would_cite":true,"duration_ms":17144,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14H10","14H15","30F30","32G15","32G20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves the tangent-space formula for the boundary of any GL(2,R) orbit closure in the WYSIWYG compactification, unconditionally for multi-component limits, while showing the compactification itself is not algebraic.","keywords":["Abelian differentials","translation surfaces","GL(2,R) orbit closures","WYSIWYG compactification","period coordinates","multi-scale differentials","invariant subvarieties","cylinder deformations"],"falsifier":"On a concrete stratum such as $\\mathcal{H}(2)$, take an explicit orbit closure $M$ defined by a linear equation in period coordinates, degenerate a sequence in $M$ to a two-component boundary point, and compute the two spaces $T M'$ and $T M \\cap \\operatorname{Ann}(V)$ in period coordinates at that point. If they are not equal, Theorem 1.2 is false.","tokens_in":25630,"feed_emoji":"📐","tokens_out":14801,"duration_ms":122390,"temperature":0.7,"pith_summary":"This paper studies the WYSIWYG compactification of strata of Abelian differentials, a partial compactification that records what flat polygons look like when a translation surface degenerates. The authors prove that, for $g\\ge 3$ with all zero orders positive and for the genus-two case $\\kappa=(1,1)$, this compactified space cannot carry an algebraic variety structure, nor even a complex analytic space structure, that is compatible with the natural projection from the stratum. Despite that negative result, they give a short, unconditional proof that the finite-area boundary of any $\\mathrm{GL}^+(2,\\mathbb{R})$ orbit closure is locally a finite union of linear subspaces in period coordinates (the local coordinates given by integrals of the differential), and that its tangent space is exactly the intersection of the tangent space of the orbit closure with the tangent space of the boundary stratum. The multi-component case, which had previously been conditional, is treated completely. The upshot is that a compactification built purely from flat geometry is not algebro-geometric, yet the boundary behaviour of orbit closures inside it is governed by algebraic linear equations.","feed_headline":"Boundary tangents of orbit closures are linear limits","feed_subtitle":"The tangent-space formula for boundaries now proven unconditionally, even for multi-component limits.","key_machinery":"The main mechanism is the interaction between period coordinates and collapse maps. Near a boundary point, a collapse map sends the relative homology of nearby smooth translation surfaces to that of the limit, so the tangent space of a stratum embeds in the tangent space of nearby strata; the subspace annihilating the vanishing cycles $V_n$ is identified with the tangent space of the boundary stratum. Period coordinates turn invariant subvarieties into loci cut out by linear equations, and the proof shows these equations push forward under the collapse map. The multi-component transition is controlled by the moduli space of multi-scale differentials $\\mathbb{I}\\mathcal{H}(\\kappa)$, which provides finitely many simply connected neighborhoods with model homology groups and continuous 'perturbed period coordinates' extending periods to the boundary (Theorem 6.4). Cylinder deformations and the Cylinder Finiteness Theorem are used to remove cylinders of large modulus, keeping the tangents constant along paths inside $M$.","core_discovery":"The paper's central claim is Theorem 1.2. Let $M$ be a $\\mathrm{GL}^+(2,\\mathbb{R})$ orbit closure in a stratum $\\mathcal{H}(\\kappa)$, and let $\\partial M^{<\\infty}$ be its boundary in the finite-area locus of the WYSIWYG compactification. If a sequence $(X_n,\\omega_n)\\in M$ converges to $(X_\\infty,\\omega_\\infty)\\in\\partial M^{<\\infty}$, then the intersection $M'$ of $\\partial M^{<\\infty}$ with the stratum of the limit surface is an algebraic variety, locally described by a finite union of linear subspaces in period coordinates; moreover, after passing to finitely many subsequences, the tangent space to a branch of $\\partial M^{<\\infty}$ equals the intersection of the tangent space of a branch of $M$ with the tangent space of the boundary stratum. This completes the multi-component case of a formula previously proven for single-component limits only under an additional hypothesis. Theorem 1.1 complements this by showing that the WYSIWYG compactification itself is not algebraic: for $g\\ge 3$ with all zero orders positive, and for $g=2$ with $\\kappa=(1,1)$, $\\mathbb{P}\\widetilde{\\mathcal{H}}(\\kappa)$ admits neither an algebraic variety nor a complex analytic space structure making the projection $\\pi$ a morphism. Theorem 1.3 then gives structural results for multi-component boundaries of invariant subvarieties, including equal rank of the projections and isogeny of associated Jacobian factors.","pith_inferences":["Editorial inference: the rigidity obstruction behind Theorem 1.1 should affect any compactification that contracts entire families of collapsed components while retaining no record of them; finer compactifications avoid this by keeping level-graph and matching data.","Editorial inference: because the proof of Theorem 1.2 uses algebraicity of orbit closures as an input, the same boundary formula may extend to invariant subvarieties of multi-component surfaces once their algebraicity is established by other means.","Editorial inference: Theorem 1.3(2) suggests a concrete classification strategy: decompose a multi-component boundary stratum into prime factors and use the 'one component controls the periods of the others' property to reduce classification to a single component; this can be tested on explicit rank-two example families."],"forward_implications":["The tangent-space formula for the boundary of an orbit closure now holds for multi-component limits without any extra hypothesis, so inductive studies of orbit closures via their boundaries can proceed in full generality.","The finite-area boundary of an orbit closure in the WYSIWYG compactification is a finite union of invariant subvarieties, each locally cut out by linear equations in the period coordinates of the boundary stratum.","For a prime invariant subvariety of a product of strata, the absolute periods of any one component determine those of every other component on the prime locus, and the natural Jacobian factors of the components are isogenous.","The listed strata admit no algebraic compactification of WYSIWYG type compatible with the projection from the stratum; algebraic treatments must use finer compactifications or analytic arguments.","Boundary equations are push-forwards of the equations defining $M$ under the collapse map, so the boundary can be computed from local period-coordinate data of $M$ alone."],"supporting_citations":[{"why":"Establishes the WYSIWYG compactification, convergence criteria, collapse maps and vanishing cycles, the single-component boundary formula, and the Cylinder Finiteness Theorem the paper extends.","marker":"[MW17]"},{"why":"Supplies the multi-scale compactification with perturbed period coordinates that Theorem 6.4 uses to extend periods continuously to the boundary.","marker":"[BCG+]"},{"why":"Shows orbit closures are algebraic varieties, letting the proof treat them as invariant subvarieties cut out by linear equations in period coordinates.","marker":"[Fil16b]"},{"why":"Provides the Cylinder Deformation Theorem used to deform large-modulus cylinders along paths that stay inside the orbit closure.","marker":"[Wri15]"},{"why":"Provides the compactification of strata and the global residue condition used to build the contracted families in the proof of Theorem 1.1.","marker":"[BCG+18]"},{"why":"Gives the zero-residue meromorphic differentials needed to construct the elliptic pencil whose contracted fibers prove non-algebraicity.","marker":"[GT]"}],"fun_headline_variants":["Unconditional boundary tangent formula for orbit closures","WYSIWYG compactification not algebraic, boundary still linear","Multi-component tangent space proof completes MW formula","Boundary tangents are linear limits, now unconditional","Non-algebraic WYSIWYG compactification, but boundary algebraic"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The unconditional proof rests on the cited construction of a smooth multi-scale compactification whose local charts extend period integrals continuously to the boundary; the paper cites that construction rather than proving it, so a gap there would undo the boundary formula.","fun_headline_variants_meta":{"raw":{"variants":["Unconditional boundary tangent formula for orbit closures","WYSIWYG compactification not algebraic, boundary still linear","Multi-component tangent space proof completes MW formula","Boundary tangents are linear limits, now unconditional","Non-algebraic WYSIWYG compactification, but boundary algebraic"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00063,"raw_usage":{"total_tokens":2895,"prompt_tokens":914,"completion_tokens":1981,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":530,"completion_tokens_details":{"reasoning_tokens":1903}},"tokens_in":530,"tokens_out":1981,"duration_ms":13952,"temperature":1.0,"reasoning_tokens":1903,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:18:47.708867+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a concrete stratum such as $\\mathcal{H}(2)$, take an explicit orbit closure $M$ defined by a linear equation in period coordinates, degenerate a sequence in $M$ to a two-component boundary point, and compute the two spaces $T M'$ and $T M \\cap \\operatorname{Ann}(V)$ in period coordinates at that point. If they are not equal, Theorem 1.2 is false.","supporting_citations":[],"review_version":1}