{"id":"a30344df-7033-435f-b925-7e3c765c6906","arxiv_id":"1908.01366","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A model-categorical foundation for filtered and stratified spaces and simplicial sets is constructed, with filtered homotopy groups and a filtered Whitehead theorem, while the full filtered Kan-Quillen equivalence remains open.","lead":"This mathematics thesis builds formal model categories for spaces with stratified singularities, and introduces filtered homotopy groups together with a filtered Whitehead theorem. It gives topologists a language for comparing singular spaces up to stratification-preserving homotopy, with applications to knots and singular varieties.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The anodinity of the fibrant replacement in Theorem 3.3.25 is the load-bearing step that needs an independent audit.","rationale":"The reader's weakest_assumption correctly identifies the fibration characterization in Theorem 3.3.25 and the supporting Appendices A and B as the load-bearing technical premise. My reading of the introduction, the statement of Theorem 1, and the surrounding discussion confirms that without the conclusion that X -> Ex^∞_P(X) is anodyne and that Ex^∞_P(X) is fibrant, the explicit Kan-style description of the model structure is unavailable. In that case, even if Theorem 3.2.15 provides a model structure via Cisinski's theorem, the filtered homotopy groups are not known to classify its weak equivalences, and the filtered Whitehead theorems would not follow. I did not find an internal contradiction or a circular argument in the portions of the text available to me; the concern is one of unverified technical depth. The concrete test I propose targets the specific combinatorial subclaim where an error or a hidden hypothesis would have maximal impact: the presentation of subdivided admissible horn inclusions as cell complexes built from admissible horns. This is the step that the appendices are designed to prove, and it is the step that connects the abstract Cisinski model structure to the concrete fibrations used throughout the paper. Because the reader's verdict of ACCEPT with MODERATE confidence already reflects this uncertainty, my stress-test does not change the verdict, but it sharpens the reason for the uncertainty and points to a concrete check that would increase or decrease confidence.","tokens_in":75064,"tokens_out":18339,"duration_ms":209098,"concrete_test":"Independently verify the key combinatorial claim in Appendix B: for every admissible horn Λ^φ_k -> Δ^φ, the subdivided inclusion sd_P(Λ^φ_k) -> sd_P(Δ^φ) is a finite relative cell complex whose attaching maps are admissible horn inclusions. Run this check explicitly for P = {0<1} with n = 2, k = 1, φ = [0,1,1] by enumerating all non-degenerate simplices of sd_P(Λ^2_1) and sd_P(Δ^2) and constructing the cell decomposition. Then repeat the same check for P = ℕ with a higher-dimensional simplex whose filtration is [0,1,1,2,3,...] to test whether the decomposition ever requires filling a horn with p_k distinct from both p_{k-1} and p_{k+1}. If such a non-admissible horn is needed, the anodinity result, and with it Theorem 3.3.25, fails for general P.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that sSetP carries a Kan-style model category and that filtered homotopy groups classify weak equivalences between fibrant objects rests on Theorem 3.3.25, which identifies the fibrations of the Cisinski model structure with the naive fibrations, i.e., maps with the right lifting property against admissible horns. The proof is deferred to Appendices A and B and depends on showing that the fibrant replacement map j_X : X -> Ex^∞_P(X) is anodyne, not merely a weak equivalence. If j_X is only a weak equivalence, the standard retract argument cannot show that every trivial cofibration is a retract of an anodyne extension; then fibrations could be strictly finer than naive fibrations, and the filtered homotopy groups might not detect the true weak equivalences. The most delicate point in Appendices A and B is the claim that, for every admissible horn inclusion Λ^φ_k -> Δ^φ, the subdivided inclusion sd_P(Λ^φ_k) -> sd_P(Δ^φ) admits a cellular decomposition built entirely from admissible horns. This is where the filtered subdivision and the diagonal functor must interact with the admissibility condition p_k = p_{k-1} or p_k = p_{k+1}. The text states Theorem 3.3.25 for an arbitrary poset P, but the appendix summary in the introduction does not explicitly state whether the combinatorial arguments require a finiteness or well-ordering hypothesis on P. Since the examples in Chapter 6 use finite flag posets, while the theorem is stated in full generality, this is a concrete place where a hidden assumption would break the central construction. The concern is not that the thesis is inconsistent, but that this particular step has not been independently verified and is essential: every later result on filtered homotopy groups and the filtered Whitehead theorems depends on it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript develops a model-categorical homotopy theory for stratified and filtered spaces. For a fixed poset P it constructs a model structure on the category sSetP of filtered simplicial sets, proves a Kan-style description of its fibrations via admissible horns, and introduces filtered homotopy groups that characterize weak equivalences between fibrant objects. It then transfers these results to filtered topological spaces, establishes filtered Whitehead theorems for conically stratified and metrically homotopically stratified spaces, and uses Quillen bifibrations to build model structures on the global categories of stratified simplicial sets and stratified spaces. The central claims are Theorems 3.2.15, 3.3.25, 4.3.12, 5.3.7, 5.3.11, and the Quillen-adjunction comparison statements in Chapter 8.","tokens_in":75334,"tokens_out":5942,"duration_ms":72001,"significance":"If the central theorems hold, this is a substantial contribution: it provides a uniform homotopy theory for stratified spaces, new algebraic invariants (filtered homotopy groups) that detect filtered homotopy type, and a Whitehead theorem in the stratified setting. The paper makes good use of standard machinery (Cisinski's existence theorem, Quillen bifibrations, the Kan–Quillen adjunction) and states its main results with precise hypotheses. The treatment of examples, including filtered Eilenberg–Mac Lane spaces and knots, gives the theory concrete content. The paper also ships a clearly separated technical core: the delicate comparison of the abstract model structure with the explicit admissible-horn description is isolated in Appendices A and B. The author is transparent about which statements are conjectural, which is a strength. My reservations are about verifiability of the deferred arguments, not about the overall architecture.","major_comments":[{"comment":"The identification of the fibrations of the Cisinski model structure with maps having the right lifting property against admissible horns, together with the assertion that X → Ex∞_P(X) is anodyne, is the load-bearing step of the whole paper. Remark 3.2.16 explicitly notes that Cisinski's theorem alone does not give this equality, so the deferred proof is essential. In the version of the manuscript made available to me, Appendices A and B are listed in the table of contents but their proofs are not reproduced, so this central verification cannot be checked. I request that the full proofs be included, or that a precise pointer to a publicly available complete version be given, and that the exact hypotheses on P used in the cell-decomposition arguments be stated in Theorem 3.3.25 itself.","section":"Theorem 3.3.25 and Appendices A–B"},{"comment":"Theorem 3.3.25 and the surrounding results are stated for an arbitrary poset P, while the proof strategy for the subdivision sd_P of admissible horns uses last-vertex choices, ancestral orders, and filtered anodyne presentations. The concrete examples in Chapter 6 use finite flag posets, and the introduction does not state whether a finiteness or well-foundedness hypothesis on P is needed. If the argument requires P to be finite or well-founded, the statements must be restricted accordingly; otherwise an explicit verification for arbitrary P, including infinite chains, should be supplied. This is not merely a presentational point, because the functor Ex∞_P and the fibrant-replacement claim are stated in full generality.","section":"§3.3.2 and Appendix B; statement for arbitrary P"}],"minor_comments":[{"comment":"In the case analysis for the space X, the phrase “Si (x,y) ∈ X2\\X1” appears twice; the second occurrence should presumably read “Si (x,y) ∈ X1\\X0”, since the following cases treat points on the one-dimensional strata.","section":"Example 1.1.15"},{"comment":"Theorem 2 in the introduction states an equivalence between being a filtered homotopy equivalence and inducing weak equivalences on strata and homotopy links, but the relevant notion of weak equivalence on strata and homotopy links is not recalled at that point; a cross-reference to the precise definitions in Chapter 5 would help the reader.","section":"Introduction, Theorem 2"},{"comment":"The guide to the reader and the text of §2.2.4 mention that cofibrant generation of sSetTop_P is an open question, while Chapter 8 later transports a model structure onto sSetTop_P; the status of the transported structure with respect to cofibrant generation should be stated explicitly at the point of transport.","section":"Chapter 8, sSetTop_P"}],"recommendation":"major_revision","confidential_remarks":"This is a carefully written thesis-style manuscript with plausible and well-motivated results. My recommendation is driven by the fact that the proof of Theorem 3.3.25, on which the filtered homotopy-group characterization and the Whitehead theorems rest, is not present in the submitted file. If the full appendices are supplied and the generality of P is clarified, I would expect to be able to recommend acceptance; I do not see a mathematical error in the parts that are visible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a substantial PhD thesis, and the main results are real. Douteau constructs a cofibrantly generated, proper, simplicial model structure on filtered simplicial sets sSetP, gives it a Kan-style description via admissible horns, defines filtered homotopy groups, and proves filtered Whitehead theorems for conically stratified and metrically homotopically stratified spaces. That goes beyond the earlier work of Nand-Lal and Haine, and the thesis is careful to say what is new and what remains conjectural. The filtered Kan-Quillen equivalence is explicitly left open, which is the right call.\n\nThe strongest part is the organization: results are stated with precise hypotheses, and the hard technical work is isolated in Appendices A and B. The examples in Chapter 6 (knots, bundles, pseudo-varieties) show the machinery has traction. No circularity jumped out: filtered homotopy groups are defined after the model structure and then used to characterize weak equivalences between fibrant objects.\n\nThe soft spot is exactly what the stress-test note identifies. Theorem 3.3.25 says that fibrations coincide with maps having the RLP against admissible horns, and that X -> Ex^∞_P(X) is anodyne. That theorem is load-bearing: without it, filtered homotopy groups may not classify the actual weak equivalences, and the Whitehead theorems lose their foundation. The proof is in Appendices A and B, and the delicate combinatorial claim is that subdivision preserves admissible horn presentations. That needs an independent check. I'd also look at whether any finiteness or well-foundedness condition on P is smuggled into the appendix arguments, since the theorem is stated for arbitrary posets while the examples use finite flag posets. These are audit points, not demonstrated errors; the thesis reads coherently and the author's honesty about conjectures suggests the arguments were checked carefully.\n\nBottom line: this is a serious piece of work that deserves a serious referee. If I were editor, I'd send it to someone comfortable with Cisinski's theory and simplicial model categories, with the instruction to start in Appendices A and B. The reader who benefits most is a specialist in stratified homotopy theory or intersection homology; for that reader, the thesis will be a standard reference.","headline":"A careful, systematic model-categorical framework for stratified spaces; the load-bearing fibration characterization in Appendices A/B deserves independent audit.","tokens_in":75915,"tokens_out":2309,"would_cite":true,"duration_ms":26622,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["55U35","18G55","55P15","55N33"],"pacs":[],"model":"deepseek-v4-flash","headline":"Filtered homotopy groups classify stratified homotopy types.","keywords":["stratified spaces","filtered simplicial sets","model categories","filtered homotopy groups","Whitehead theorem","conically stratified spaces","homotopically stratified spaces","Quillen adjunction"],"falsifier":"The central claim would be falsified by exhibiting a filtered simplicial set $X$ for which $\\mathrm{Ex}^\\infty_P(X)$ fails to have the right lifting property against an admissible horn, since Theorem 3.3.25 asserts this is always a fibration; alternatively, a map between fibrant filtered simplicial sets that induces isomorphisms on all filtered homotopy groups but is not a filtered homotopy equivalence would falsify the filtered Whitehead theorem.","tokens_in":74844,"feed_emoji":"🧩","tokens_out":8168,"duration_ms":75201,"temperature":0.7,"pith_summary":"This thesis tries to build a homotopy theory for stratified spaces that respects the stratification itself, not just the underlying topological space. The paper's route is to fix a poset $P$ of strata and study filtered simplicial sets, objects $X\\to N(P)$, with a model category whose weak equivalences are detected by newly defined filtered homotopy groups. The central payoff is a filtered Whitehead theorem: for conically stratified or metrically homotopically stratified spaces admitting filtered simplicial replacements, a stratified map is a filtered homotopy equivalence exactly when it induces isomorphisms on all filtered homotopy groups, equivalently weak equivalences on all strata and homotopy links. A sympathetic reader would care because intersection cohomology and other stratified invariants are only invariant under stratified homotopies, so a homotopy category adapted to those homotopies is the natural setting for classifying stratified spaces.","feed_headline":"Filtered homotopy groups classify stratified homotopy types","feed_subtitle":"New model categories for filtered and stratified spaces give a Whitehead theorem and a Kan–Quillen adjunction.","key_machinery":"The central object is the category $sSet_P$, whose objects are simplicial sets $X$ equipped with a filtration $\\varphi_X\\colon X\\to N(P)$, the nerve of a fixed poset of strata. The mechanism that carries the argument is the filtered subdivision–extension pair $\\mathrm{sd}_P\\dashv\\mathrm{Ex}_P$, iterated to $\\mathrm{Ex}^\\infty_P$, which provides the fibrant replacement $X\\to\\mathrm{Ex}^\\infty_P(X)$ and yields the \"à la Kan\" description of fibrations. The key combinatorial notion is the admissible horn: an inclusion $\\Lambda^\\varphi_k\\to\\Delta^\\varphi$ of a horn into a filtered simplex is admissible when the filtration repeats an adjacent value at $k$, which is exactly the condition under which filling the horn does not change the filtered homotopy type. Admissible horns generate the anodyne extensions, and the filtered homotopy groups, defined from filtered spheres, are shown to detect precisely the weak equivalences between fibrant objects.","core_discovery":"On the paper's own terms, the discovery is that the homotopy theory of filtered and stratified spaces admits the same shape as classical homotopy theory, with filtered homotopy groups playing the role of homotopy groups. Theorem 1 (synthesizing Theorems 3.2.15, 3.3.25, and 4.3.12) states that the category $sSet_P$ of filtered simplicial sets over a poset $P$ is a simplicial, cofibrantly generated, proper model category: cofibrations are monomorphisms, fibrations are maps with the right lifting property against admissible horns, and a map between fibrant objects is a weak equivalence precisely when it induces isomorphisms on all filtered homotopy groups. Theorems 5.3.7 and 5.3.11 then derive the filtered Whitehead theorem for filtered spaces: for conically stratified or metrically homotopically stratified spaces that admit filtered simplicial replacements, a filtered map is a filtered homotopy equivalence if and only if it induces isomorphisms on filtered homotopy groups, which is also equivalent to inducing weak equivalences on all strata and homotopy links. The paper further constructs a model category of filtered spaces, Quillen equivalent to a diagram category, and a Quillen adjunction between filtered simplicial sets and filtered spaces, reducing to the classical Kan–Quillen adjunction when $P$ is a point.","pith_inferences":["If Conjecture 1 (Quillen equivalence between filtered spaces and filtered simplicial sets) holds, then filtered homotopy groups would give a complete and computable classification of filtered homotopy types for the same range of spaces where classical homotopy groups classify CW complexes, and the homotopy categories would be interchangeable in practice.","The filtered homotopy groups appear to encode not only the homotopy groups of individual strata but also the way paths can approach lower strata, so they may be the stratified analogue of a Postnikov tower or of the exit-path category; a testable consequence is that they should determine the homotopy type of the exit-path $\\infty$-category for conically stratified spaces.","The admissible-horn condition suggests an algorithmic recognition procedure: a finite filtered simplicial set is fibrant if every missing face whose horn is admissible can be filled; implementing this check would give a computable way to search for filtered weak equivalences in finite examples.","The paper's example of knots suggests that filtered homotopy groups can serve as complete invariants for embedded submanifolds with their natural filtrations; one could test whether the same groups distinguish links up to stratified homotopy, or whether they refine classical link concordance invariants."],"forward_implications":["For conically stratified and metrically homotopically stratified spaces, filtered homotopy groups are complete stratified homotopy invariants: two such spaces with the same strata have the same filtered homotopy type exactly when all filtered homotopy groups agree.","Since intersection cohomology is invariant under stratified homotopies, the new model categories provide a homotopy-theoretic setting in which intersection cohomology and related perverse invariants can be studied as representable or homotopy-invariant functors.","The Kan-style characterization of fibrations makes the model structure on $sSet_P$ usable for computation: to test whether a map is a fibration one only checks lifting against admissible horns, and filtered homotopy groups can be computed from fibrant replacements.","The filtered Kan–Quillen adjunction $(\\|\\mathrm{sd}_P(-)\\|_P,\\mathrm{Ex}_P\\mathrm{Sing}_P)$ between filtered simplicial sets and filtered spaces preserves weak equivalences, so the two model categories present related homotopy theories and differ only at the level of the conjectured Quillen equivalence.","When $P$ is a singleton, all of these constructions collapse to the classical model structure on simplicial sets and the classical Kan–Quillen adjunction, so the filtered theory is a genuine extension rather than a parallel theory."],"supporting_citations":[{"why":"Supplies the general recognition theorem used to produce the model structure on $sSet_P$ once the class of anodyne extensions has been fixed.","marker":"[Cis06]"},{"why":"Shows that the singular filtered simplicial set of a conically stratified space is fibrant, one of the two geometric hypotheses in the filtered Whitehead theorem.","marker":"[Lur]"},{"why":"Shows the same fibrantness for metrically homotopically stratified spaces, the other geometric hypothesis in the theorem.","marker":"[NL19]"},{"why":"Provides the axioms, the original model structure on simplicial sets, and the Kan–Quillen adjunction that the filtered constructions generalize.","marker":"[Qui67]"},{"why":"Establishes the realization statements about simplicial sets, including compatibility with products and the fact that the unit and counit are weak equivalences, used to transfer between filtered spaces and filtered simplicial sets.","marker":"[Mil57]"},{"why":"Supplies the transport theorem from which the model structure on filtered spaces is derived from a diagram category of reduced simplicial diagrams.","marker":"[HKRS17]"}],"fun_headline_variants":["Filtered homotopy groups classify stratified homotopy types","Stratified spaces get a Whitehead theorem","Model categories for filtered and stratified spaces","A Kan–Quillen adjunction for stratified spaces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole classification rests on the claim that the concrete replacement built from filtered subdivision can in fact be used to tell fibrations from non-fibrations; if that replacement failed to be fibrant or to stay weakly equivalent, the filtered homotopy groups would stop being a reliable test for weak equivalence.","fun_headline_variants_meta":{"raw":{"variants":["Filtered homotopy groups classify stratified homotopy types","Stratified spaces get a Whitehead theorem","Model categories for filtered and stratified spaces","A Kan–Quillen adjunction for stratified spaces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000347,"raw_usage":{"total_tokens":2027,"prompt_tokens":1200,"completion_tokens":827,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":816,"completion_tokens_details":{"reasoning_tokens":766}},"tokens_in":816,"tokens_out":827,"duration_ms":12349,"temperature":1.0,"reasoning_tokens":766,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:15:05.472933+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The central claim would be falsified by exhibiting a filtered simplicial set $X$ for which $\\mathrm{Ex}^\\infty_P(X)$ fails to have the right lifting property against an admissible horn, since Theorem 3.3.25 asserts this is always a fibration; alternatively, a map between fibrant filtered simplicial sets that induces isomorphisms on all filtered homotopy groups but is not a filtered homotopy equivalence would falsify the filtered Whitehead theorem.","supporting_citations":[],"review_version":1}