{"id":"340e576a-872c-48ea-8bc7-9ea3e09f7999","arxiv_id":"1908.03088","paper_version":3,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For finite-type spaces with a C2-action, having a conjugation frame is equivalent to being homologically pure, meaning X ∧ HF splits as a wedge of sign-representation suspensions of HF.","lead":"The paper proves that conjugation spaces, spaces with a reflection symmetry whose fixed points have cohomology in half the degrees, are exactly the spaces whose equivariant cohomology splits into simple building blocks called pure pieces. This recasts an older computational definition in modern stable homotopy theory and gives conceptual proofs of the known structural properties of these spaces.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'only if' direction of Theorem 7.1 silently assumes H^{odd}(X)=0 for conjugation spaces; Definition 1.1 does not state or prove this, and a trivial-action wedge S^1∨S^2∨S^4∨… with an explicit frame appears to satisfy the definition but is not pure.","rationale":"The reader's weakest_assumption was Proposition 4.7, the geometric fixed-point splitting of HF. The concern identified here is different and more directly fatal: the proof of the 'only if' direction of Theorem 7.1 (conjugation implies purity) requires that the assembled map f be an equivalence after forgetting the action, which is true only if the chosen basis of H^{2*}(X) spans all of H^*(X), i.e., if H^{odd}(X)=0. The paper never proves this. Lemma 6.1 simply asserts surjectivity of the restriction ρ in integral grading as a consequence of being a conjugation space, but the frame in Definition 1.1 only gives a section on even degrees. The example with a trivial-action wedge of spheres S^1∨S^2∨S^4∨⋯ seems to satisfy every clause of Definition 1.1 while having nontrivial H^1, and it is not pure because the underlying homology is not even-concentrated. If the example is valid, the central theorem is false as stated; if the original literature includes an evenness condition, then the paper's Definition 1.1 is incomplete and the proof still omits the necessary argument. In either case, the current claim and proof are not sound. The concrete test asks for a direct verification of the frame axioms on that example, which would settle the issue.","tokens_in":29658,"tokens_out":30112,"duration_ms":305429,"concrete_test":"Verify the proposed frame on X = S^1∨S^2∨S^4∨S^8∨⋯ with trivial C2-action against Definition 1.1, checking the conjugation equation degreewise for each generator: if κ0(x_k)=x_{k-1} and σ(x_k)=x_k+b^{2^{k-1}}x_{k-1} satisfy all clauses, then Theorem 7.1 is false because X is not pure; if the example fails, identify which axiom rules it out.","verdict_should_be":"REJECT","load_bearing_attack":"Definition 1.1 only requires κ0 and σ on even-degree cohomology. Lemma 6.1 then asserts that 'by hypothesis, X being a conjugation space, the restriction ρ ... is surjective in integral grading' (Section 6), and Theorem 6.3 concludes that the assembled map f is 'an equivalence after forgetting the action.' Both statements would follow only if H^{odd}(X)=0. No such vanishing is proved or cited before these uses; the later vanishing of HF^{n(1+α)+1}(X) in Section 7.3 already relies on Theorem 7.1 and Lemma 5.3(4), so it cannot be used here. Under Definition 1.1 as written, take C2 to act trivially on X = S^1∨S^2∨S^4∨S^8∨⋯ (one sphere in each power-of-two degree) and let x_k be the fundamental class of S^{2^k}. Set κ0(x_k)=x_{k-1} for k≥1 (with x_0 the class of S^1) and σ(x_k)=x_k + b^{2^{k-1}}κ0(x_k). Then κ0 is an additive degree-halving isomorphism H^{2*}X→H^*X^{C2}, σ is an additive section of ρ, and rσ(x_k)=x_k+b^{2^{k-1}}κ0(x_k) has leading term κ0(x_k)b^{2^{k-1}} with lower b-degree 0. The frame axioms hold, so X is a conjugation space of finite type. But X∧HF has underlying homology in degree 1 and cannot split as ∨Σ^{n_i(1+α)}HF. Thus Theorem 7.1 is false as stated, or Definition 1.1 is missing an evenness hypothesis that the proof needs.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a stable equivariant characterization of conjugation spaces: a finite-type C2-space is a conjugation space if and only if its equivariant homology is pure, i.e. X ∧ HF splits as a wedge of suspensions of HF by multiples of the regular representation. The authors construct the conjugation frame maps κ0 and σ from geometric fixed points and from the collapse map EC2+ ∧ X → X, and use the purity description to recover uniqueness, functoriality, multiplicativity, the Franz–Puppe Steenrod-square formulas, and the Lannes–Zarati description of Borel cohomology.","tokens_in":30055,"tokens_out":13099,"duration_ms":149397,"significance":"If the main theorem were correct as stated, it would be a valuable reformulation: it gives geometric meaning to the otherwise algebraic conjugation frame, explains uniqueness and functoriality conceptually, and recovers known structural results in a unified way. The paper also contains useful computations, such as the geometric fixed points of HF and the equivariant Steenrod algebra actions. However, the theorem as stated is not correct: the definition of conjugation space used in the paper does not imply vanishing of odd-degree cohomology, and there are spaces satisfying that definition that are not homologically pure.","major_comments":[{"comment":"The paper never states or proves that a conjugation space has H^{odd}(X)=0, and Theorem 7.1 is false without such a hypothesis. Let C2 act trivially on X = S^1∨S^2∨S^4∨S^8∨⋯ and write x_k for the fundamental class of S^{2^k}. Define κ0(x_k)=x_{k−1} for k≥1 and σ(x_k)=x_k+b^{2^{k−1}}κ0(x_k). Then κ0 is an additive degree-halving isomorphism H^{2*}X≅H^*X^{C2}, σ is an additive section of ρ on even degrees, and rσ(x_k)=κ0(x_k)b^{2^{k−1}}+x_k, so the conjugation equation holds. However X∧HF has an odd-dimensional homology class in degree 1 after forgetting the action, so it cannot split as a wedge of spectra Σ^{n_i(1+α)}HF. Thus Theorem 7.1 is false exactly as stated.","section":"Definition 1.1 and Theorem 7.1"},{"comment":"The assertion \"By construction it is an equivalence after forgetting the action\" is not justified. The map f is assembled from lifts of a basis of H^{2*}(X) only; after forgetting the action, its target is a wedge of spectra Σ^{2n_i}H, whose underlying homology is concentrated in even degrees. Unless one knows H^{odd}(X)=0, the map on odd-dimensional homology need not be an isomorphism, and the counterexample in the previous comment is exactly such a case. The argument needs H^{odd}(X)=0, either as part of the definition of conjugation space or as a separate hypothesis.","section":"Proof of Theorem 6.3"},{"comment":"In the proof of Lemma 6.1 the authors state that \"by hypothesis, X being a conjugation space, the restriction ... is surjective in integral grading.\" Definition 1.1 only provides a section of ρ on even-degree cohomology; it gives no information about odd degrees. The cited sentence is therefore not a consequence of the definition, and the proof of the lemma should be rewritten to use only the even-degree surjectivity that the frame actually supplies.","section":"Lemma 6.1"}],"minor_comments":[{"comment":"The vanishing of HF^{n(1+α)+1}(X) is justified by Lemma 5.3(4), but Lemma 5.3 is stated for homologically pure spaces; the proof should explicitly cite Theorem 7.1 when applying it to a conjugation space.","section":"Section 7.3"},{"comment":"References [39] and [40] appear to be identical entries for tom Dieck's \"Orbittypen und äquivariante Homologie. II\"; the citations should be checked and the duplicate or mislabeled entry corrected.","section":"References"},{"comment":"The Lewis diagrams in Section 4 and the displayed formulas in Appendix A are reproduced with typographical artifacts (for example \"one the one hand\" in the proof of Lemma 5.9 and several broken diagram types); a careful proofreading pass would improve readability.","section":"Section 4 and Appendix A"}],"recommendation":"major_revision","confidential_remarks":"The counterexample in the report is a genuine counterexample to Theorem 7.1 as stated, not a stylistic quibble. The most likely repair is to add an evenness hypothesis to Definition 1.1 or to Theorem 7.1 and to make the proof of Theorem 6.3 use that hypothesis explicitly. Because this is a local fix rather than a failure of the overall strategy, I would not reject the paper outright, but I cannot recommend acceptance before the statement is corrected."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper's stable characterization is a good idea, and the \"pure implies conjugation\" half is solid and instructive. But as written Theorem 7.1 is false: the converse needs an even-cohomology hypothesis on X that Definition 1.1 never states and the proof never proves.\n\nWhat is genuinely new: encoding the frame maps κ0 and σ as natural maps coming from geometric fixed points and from EC2+∧X→X, and recasting the conjugation frame as purity of X∧HF. That is a real conceptual simplification. Uniqueness, functoriality, multiplicativity, Franz–Puppe, and Lannes–Zarati all drop out of stable naturality rather than from computation. The background sections on Mackey functors, geometric fixed points, and Stong's coefficients are useful and well organized.\n\nThe soft spot is load-bearing. In Lemma 6.1 the proof says that since X is a conjugation space, the restriction ρ is surjective in integral grading. But Definition 1.1 only gives a section on even-degree cohomology; surjectivity in odd degrees does not follow. Theorem 6.3 assembles lifts of an even-degree basis into a map f that is supposed to be an equivalence after forgetting the action; that equivalence requires H^odd(X)=0. No such vanishing is stated or derived before it is used. The later vanishing argument in Section 7.3 already assumes Theorem 7.1, so it cannot supply the missing hypothesis.\n\nThis is not a manufactured worry. Let C2 act trivially on X=S^1∨S^2∨S^4∨⋯ (one sphere in each power-of-two degree), with x_k the class of S^{2^k} for k≥1 and x_0 the class of S^1. Define κ0(x_k)=x_{k-1} and σ(x_k)=x_k+b^{2^{k-1}}κ0(x_k). This satisfies the frame axioms exactly as written, and X is of finite type. It is not pure, because X∧HF has underlying homology in degree 1. So Theorem 7.1 is false under the definition given in the paper.\n\nThe likely fix is to add H^odd(X)=0 to Definition 1.1 (or to Theorem 7.1) and use it in Lemma 6.1 and Theorem 6.3. I suspect the original Hausmann–Holm–Puppe definition contains or implies this condition, so this may be an expository gap rather than a fatal mathematical one. But it is a gap in the central theorem and needs to be fixed before the equivalence is cited.\n\nWho should read it: equivariant homotopy theorists and transformation-group people interested in conjugation spaces; the stable framework is worth knowing even with the caveat. I would send it to a serious referee, and I would want the referee to check the converse after the evenness hypothesis is added. My own next-12-month citation: no, not until the statement is corrected.","headline":"The stable characterization of conjugation spaces is a good idea and the pure-to-conjugation direction is solid, but as written Theorem 7.1 is false because the converse silently assumes H^odd(X)=0.","tokens_in":30575,"tokens_out":10578,"would_cite":false,"duration_ms":113978,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["55P91","57S17","55S10","55N91","55P42"],"pacs":[],"model":"deepseek-v4-flash","headline":"A C2-space is a conjugation space precisely when it is homologically pure.","keywords":["conjugation spaces","homological purity","geometric fixed points","equivariant stable homotopy theory","C2-spaces","Steenrod squares","Mackey functors","Borel cohomology"],"falsifier":"Compute the homotopy groups $\\pi_*(\\Phi^{C_2}(H\\mathbb{F}))$ directly from the isotropy separation sequence $EC_{2+} \\to S^0 \\to \\widetilde{EC}_2$ smashed with $H\\mathbb{F}$, and check that they are one-dimensional in each degree $k \\ge 0$ with polynomial generators $b_k$; if any degree has rank different from 1, or if the $H$-module structure forces a non-split extension rather than a wedge, the projection defining $\\kappa_0$ would not exist and Theorem 7.1 would fail.","tokens_in":29450,"feed_emoji":"🔄","tokens_out":9331,"duration_ms":89500,"temperature":0.7,"pith_summary":"The paper establishes that the classical notion of a conjugation space—a space with an involution whose fixed-point mod-2 cohomology is the degree-halved cohomology of the whole space, as in complex projective spaces under complex conjugation—is exactly the same as a stable equivariant condition called homological purity. A finite-type C2-space is pure when its mod-2 equivariant cohomology splits into basic pieces indexed by the even-degree cohomology classes, with no extra structure required. This matters because it converts a complicated algebraic structure (an H*-frame, i.e. a degree-halving map plus a section satisfying the conjugation equation) into a geometric property that can be read off the equivariant cohomology. The paper then uses the stable picture to explain conceptually why the conjugation frame is unique, multiplicative, functorial, compatible with Steenrod squares, and why the Borel cohomology of a conjugation space is entirely determined by the cohomology of its fixed points.","feed_headline":"Conjugation spaces are cohomologically pure","feed_subtitle":"A finite-type C2-space is a conjugation space exactly when its equivariant cohomology splits into pure regular-representation spheres.","key_machinery":"The central object is the equivariant Eilenberg-MacLane spectrum $H\\mathbb{F}$ for mod-2 ordinary equivariant cohomology, and the notion of purity borrowed from the equivariant literature: X is homologically pure when $X \\wedge H\\mathbb{F}$ is equivalent to a wedge of suspensions $\\Sigma^{n(1+\\alpha)} H\\mathbb{F}$, suspensions by multiples of the regular representation $1+\\alpha$. The load-bearing computation is Proposition 4.7, the $H$-linear splitting of geometric fixed points $\\Phi^{C_2}(H\\mathbb{F}) \\simeq H[b] \\simeq \\bigvee_{k\\ge 0} S^k \\wedge H$, obtained by computing homotopy groups and invoking Robinson's theorem on $H$-modules; this splitting makes $\\kappa_0$ a well-defined projection onto the zeroth factor and makes the Nakayama argument in Theorem 6.3 work on fixed-point homology. Alongside it, the computation of the equivariant Steenrod algebra identifies the conjugation equation with the Steinberg map, giving the Borel-cohomology determination.","core_discovery":"The paper proves Theorem 7.1: for a C2-space X of finite type, X is a conjugation space if and only if it is homologically pure of finite type, meaning that $X \\wedge H\\mathbb{F}$ is weakly equivalent, as a right $H\\mathbb{F}$-module, to a wedge of suspensions $\\Sigma^{n_i(1+\\alpha)} H\\mathbb{F}$ by multiples of the regular representation. The proof shows that the conjugation frame itself is not a choice: the degree-halving isomorphism $\\kappa_0$ is induced by geometric fixed points (projection onto the zeroth factor of $\\Phi^{C_2}(H\\mathbb{F}) \\simeq \\bigvee_{k\\ge 0} S^k \\wedge H$), and the section $\\sigma$ is induced by the collapse $EC_{2+} \\wedge X \\to X$. Consequently all known structural properties of conjugation spaces—uniqueness, functoriality, multiplicativity, Steenrod compatibility, and the determination of Borel cohomology by fixed-point cohomology—are recovered from stable equivariant homotopy theory rather than from the original conjugation equation.","pith_inferences":["If the purity criterion is accepted, the search for new conjugation spaces reduces to detecting C2-spaces whose underlying cohomology is even-concentrated and whose geometric fixed points match the pure splitting; this may yield new non-spherical examples more easily than solving conjugation equations.","The same style of argument might extend to other finite groups or other coefficient fields wherever a Stong-style coefficient computation and an H-module splitting are available, with the regular representation replaced by the appropriate $RO(G)$-graded shifts.","Identifying the conjugation equation with the Steinberg map places conjugation spaces inside the theory of unstable modules over the Steenrod algebra; this may allow purely algebraic classification of which unstable algebras can arise as the cohomology of a conjugation space.","Because $\\kappa_0$ is now a geometric fixed-point projection, any equivariant cohomology operation preserving the line $m(1+\\alpha)$ will automatically commute with the frame, suggesting a wider family of operations compatible with conjugation besides Steenrod squares."],"forward_implications":["Being a conjugation space is a property, not a structure: no choice of H*-frame is needed once homological purity holds.","The conjugation frame is unique, functorial, and multiplicative because it is induced by natural maps of spectra: geometric fixed points and the collapse $EC_{2+} \\wedge X \\to X$.","The Steenrod-square compatibility formulas $\\kappa_0(Sq^{2\\ell}x) = Sq^\\ell\\kappa_0(x)$ and $\\kappa_\\ell(x) = Sq^\\ell\\kappa_0(x)$ follow from naturality of equivariant operations of degree $\\ell(1+\\alpha)$, without case-by-case computation.","For a conjugation space X, the Borel cohomology $H^*(X_{hC_2})$ is functorially isomorphic to $R H^*(X^{C_2})$ and the conjugation equation is exactly the Steinberg map $St(x) = \\sum_j b^{n-j} \\otimes Sq^j\\kappa_0(x)$.","Every spherical conjugation space built from conjugation cells is pure, and conversely every finite-type pure space admits the stable splitting, connecting the original examples to the new characterization."],"supporting_citations":[{"why":"Supplies the original definition of conjugation spaces and H*-frames, whose structural properties this paper recovers conceptually from purity.","marker":"[12]"},{"why":"Provides the definition of homological purity and the geometric fixed-point properties used to construct kappa_0.","marker":"[15]"},{"why":"Gives the structure of the equivariant Steenrod algebra and of the coefficient ring $H\\mathbb{F}_\\star$, used in the computations behind Proposition 4.7 and Steenrod compatibility.","marker":"[16]"},{"why":"Supplies the splitting of free $H\\mathbb{F}$-modules for finite C2-spaces, used to identify purity and to compute cohomology of representation spheres.","marker":"[26]"},{"why":"Proves that $H$-modules decompose as products of Eilenberg-MacLane spectra, a key ingredient in the geometric fixed-point splitting Proposition 4.7.","marker":"[33]"},{"why":"Provides the RO(C2)-graded Mackey functor computation of $H\\mathbb{F}_\\star$ that feeds the homotopy-group calculation in Proposition 4.7.","marker":"[7]"},{"why":"Establishes the Steenrod-square compatibility formulas that the paper re-derives conceptually from the stable description.","marker":"[8]"},{"why":"Introduces the Steinberg map and the derived functor of destabilization, used to identify the conjugation equation and to prove the Borel-cohomology determination.","marker":"[17]"}],"fun_headline_variants":["Conjugation spaces are pure homologically","Purity characterizes conjugation spaces","Homological purity: the conjugation space test","Finite C2-spaces: conjugation iff purity","Stable homotopy proves conjugation purity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument stands or falls on the splitting of geometric fixed points $\\Phi^{C_2}(H\\mathbb{F})$ as a wedge of ordinary Eilenberg-MacLane spectra $H$ in every degree $k \\ge 0$ (Proposition 4.7), together with the coefficient computation of $H\\mathbb{F}$ that feeds it; if that splitting were not an equivalence of $H$-modules, the construction of $\\kappa_0$ as a projection would break.","fun_headline_variants_meta":{"raw":{"variants":["Conjugation spaces are pure homologically","Purity characterizes conjugation spaces","Homological purity: the conjugation space test","Finite C2-spaces: conjugation iff purity","Stable homotopy proves conjugation purity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000389,"raw_usage":{"total_tokens":2004,"prompt_tokens":854,"completion_tokens":1150,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":470,"completion_tokens_details":{"reasoning_tokens":1086}},"tokens_in":470,"tokens_out":1150,"duration_ms":11962,"temperature":1.0,"reasoning_tokens":1086,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:25:15.852469+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the homotopy groups $\\pi_*(\\Phi^{C_2}(H\\mathbb{F}))$ directly from the isotropy separation sequence $EC_{2+} \\to S^0 \\to \\widetilde{EC}_2$ smashed with $H\\mathbb{F}$, and check that they are one-dimensional in each degree $k \\ge 0$ with polynomial generators $b_k$; if any degree has rank different from 1, or if the $H$-module structure forces a non-split extension rather than a wedge, the projection defining $\\kappa_0$ would not exist and Theorem 7.1 would fail.","supporting_citations":[{"cited_title":"Hausmann, T","cited_arxiv_id":null,"evidence_quote":"Supplies the original definition of conjugation spaces and H*-frames, whose structural properties this paper recovers conceptually from purity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the definition of homological purity and the geometric fixed-point properties used to construct kappa_0."},{"cited_title":"A structure theorem for $RO(C_2)$-graded Bredon cohomology","cited_arxiv_id":"1804.03691","evidence_quote":"Supplies the splitting of free $H\\mathbb{F}$-modules for finite C2-spaces, used to identify purity and to compute cohomology of representation spheres."},{"cited_title":"Robinson","cited_arxiv_id":null,"evidence_quote":"Proves that $H$-modules decompose as products of Eilenberg-MacLane spectra, a key ingredient in the geometric fixed-point splitting Proposition 4.7."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the RO(C2)-graded Mackey functor computation of $H\\mathbb{F}_\\star$ that feeds the homotopy-group calculation in Proposition 4.7."},{"cited_title":"Franz and V","cited_arxiv_id":null,"evidence_quote":"Establishes the Steenrod-square compatibility formulas that the paper re-derives conceptually from the stable description."},{"cited_title":"Lannes and S","cited_arxiv_id":null,"evidence_quote":"Introduces the Steinberg map and the derived functor of destabilization, used to identify the conjugation equation and to prove the Borel-cohomology determination."}],"review_version":1}