{"id":"e8240b6e-0c77-4d39-8263-233d89d7f778","arxiv_id":"2507.15891","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Transverse groups preserving proper domains in flag manifolds must have limit triples of a single type (Maslov index zero in the tube-type case), but the proof of this key claim contains a false step.","lead":"The paper derives a necessary condition for subgroups of Lie groups to preserve proper domains in flag manifolds, via a new notion of causal convexity on Shilov boundaries. A key proof step appears to be wrong, so the main theorem is not established as written.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 4.5's claim that limit points lie in Ω contradicts (4.3), since a∈Z_a; Theorem 1.2 rests on this invalid step.","rationale":"The reader's rejection is justified. The paper develops a substantial framework—causal convexity in Shilov boundaries, the openness result Corollary 7.1, and explicit Zariski-dense examples in Section 7—and these contributions do not depend on the faulty step. However, the advertised topological obstruction is specifically Theorem 1.2, whose proof is routed through Proposition 4.5. The assertion a∈Ω is not a harmless omission: it is contradicted by the stated hypotheses. A sequence in an open set converging to a boundary point is the standard way limit sets arise; properness plus (4.3) excludes exactly that membership. Since the same impossible membership is used to place pairwise transverse limit points in one connected component O, the constancy of the type and the Maslov-index-zero conclusion lose their support. I also note a secondary concern that deriving (4.3) from Lemma 2.17 in case (2) applies the lemma to the wrong (dual) limit, and that Lemma 5.8 contains an unsupported photon-chain assertion; both reinforce the need for a rewritten proof. A repaired argument might save the theorem, but the version under review has not supplied it. Verdict remains REJECT as in the reader's report.","tokens_in":44552,"tokens_out":19185,"duration_ms":241323,"concrete_test":"Analytic check in the model of Example 2.9(1): take G=Sp(2r,R), Sb(g)=Lag(R^{2r}), and any proper invariant domain Ω satisfying hypothesis (4.3) of Proposition 4.5. For any a∈ΛΘ(H), a is a Lagrangian subspace; since a∩a=a≠0, a is non-transverse to itself, so a∈Z_a. Equation (4.3) then gives a∉Ω, while the proof's first claim asserts a∈Ω. This one-line check isolates the invalid inference and shows Proposition 4.5 cannot be accepted as written; any repair must establish type constancy without using a∈Ω.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Proposition 4.5 is the only proof of Theorem 1.2. Its first step states: 'First note that a ∈ Ω for all a ∈ ΛΘ(H) ... Since Ω is H-invariant, we have a ∈ Ω.' The inference is invalid: from h_n·x → a with h_n·x ∈ Ω one only gets a ∈ Ω̄. Under the paper's own hypothesis (4.3), Z_a ∩ Ω = ∅ for every a ∈ ΛΘ(H), and in the self-opposite flag manifolds in question a point is non-transverse to itself (e.g. a Lagrangian L satisfies L∩L = L), so a ∈ Z_a and therefore a ∉ Ω. Thus (4.3) forces the exact opposite of the proof's claim. The later steps—choosing a component O with Ω ⊂ O, placing x,y ∈ ΛΘ(H) in O, and deriving sΘ-invariance of O—all depend on placing limit points inside Ω. In the proper-domain case (2), the derivation of (4.3) from Lemma 2.17 is also not immediate, since the lemma applies to the iΘ-limit b of a contracting sequence rather than to its Θ-limit a. Hence Corollary 4.6 and Theorem 1.2 are not established as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies subgroups H of a real semisimple Lie group G that preserve a proper domain in a self-opposite flag manifold G/P. Its central theorem, Theorem 1.2, asserts a topological restriction: if the limit set Λ_P(H) contains at least three pairwise transverse points, then all triples of pairwise transverse limit points have the same type, represented by an s-invariant connected component of the complement of the two standard Schubert cycles; in the Hermitian tube-type case this forces the real rank to be even and the Maslov index of every such triple to be zero. The paper also develops a notion of causal convexity in Shilov boundaries of Hermitian tube-type groups, proves that dually convex proper domains are causally convex (Proposition 1.6), gives an equivalence between transverse groups preserving proper domains and groups acting cocompactly on convex cores with transverse ideal boundary (Theorem 1.7), and constructs Zariski-dense P-Anosov subgroups preserving proper domains (Theorem 1.4).","tokens_in":44843,"tokens_out":7043,"duration_ms":81908,"significance":"If Theorem 1.2 were valid, it would be a substantial new obstruction to the existence of proper domains in flag manifolds and would connect naturally with the Property I program of Dey--Greenberg--Riestenberg; it would also rule out, for example, maximal representations preserving proper domains in Shilov boundaries. The paper contains original technical material of independent interest: the causal convexity framework in Section 3, the openness result for domains preserved under Anosov deformations (Corollary 7.1), and the explicit examples in Section 7. However, the proof of the advertised main topological obstruction contains a false inference, and the central theorem is therefore not established as written.","major_comments":[{"comment":"The first step of the proof asserts that a ∈ Ω for every a ∈ ΛΘ(H). This is not a consequence of the preceding argument: from h_n·x ∈ Ω and h_n·x → a one only obtains a ∈ Ω̄, and H-invariance of Ω does not imply that Ω is closed. More seriously, the hypothesis (4.3) forces exactly the opposite conclusion. Indeed, for a self-opposite flag manifold a point is non-transverse to itself, so a ∈ Z_a; since (4.3) gives Z_a ∩ Ω = ∅, it follows that a ∉ Ω. The subsequent steps of the proof—choosing a connected component O with Ω ⊂ O, placing x,y ∈ Ω, deriving sΘ(O) = O, and concluding that typ(a,b,c) is constant—all depend on placing limit points inside Ω. This is a load-bearing error, and Corollary 4.6 and Theorem 1.2 are not established by the argument given.","section":"§4.2, Proposition 4.5, proof of part (1)"},{"comment":"The reduction of the proper-domain case to condition (4.3) via Lemma 2.17 is not justified. For p ∈ ΛΘ(H), a Θ-contracting sequence has an associated pair (p,b), where p is the Θ-limit and b is the i(Θ)-limit. Lemma 2.17, applied to the group Aut(Ω), says that b ∈ Ω*, i.e. Z_b ∩ Ω = ∅; it does not say that Z_p ∩ Ω = ∅. The proof, however, needs Z_p ∩ Ω = ∅ for every p ∈ ΛΘ(H). Thus the implication from the proper-domain hypothesis to equation (4.3) is not obtained, and the proof of part (2) fails independently of the issue raised in the previous comment.","section":"§4.2, Proposition 4.5, proof of part (2)"},{"comment":"The proof of point (2) invokes an unproved assertion: 'there exists a chain of photons between y and the point x0 determined in Point (1), contained in Ω.' No reference or proof is supplied for this photon-chain connectivity of arbitrary proper domains in Sb(g). The argument uses the chain to conclude that the limits a and a′ are non-transverse, which is essential for identifying the limit of the orbit of y with the limit of the orbit of x0. Since Lemma 5.8 is used in the proof of the implication (2) ⇒ (1) of Theorem 1.7, this missing justification also affects Theorem 1.7 as written.","section":"§5.3, Lemma 5.8(2)"}],"minor_comments":[{"comment":"In the final sentence of the proof, 'idx(x,y,z) = 0 for every triple of distinct points' should read 'for every triple of pairwise transverse points', since the Maslov index is defined for pairwise transverse triples.","section":"§4.3, Corollary 4.6"},{"comment":"The proof states that Ω* is open; for a proper open domain Ω, the dual Ω* is compact with nonempty interior, but not necessarily open. The density argument works with the interior of Ω*, so this is a presentation issue rather than a mathematical obstruction.","section":"§5.3, Lemma 5.8(4)"},{"comment":"In the continuity argument, the displayed inequality 'δ(x, x_k) ≤ HΩ(x0, x_k)' should presumably be 'δ(x0, x_k) ≤ HΩ(x0, x_k)'; as written the inequality is not the one used in the following line.","section":"§5.2.2, Lemma 5.6"},{"comment":"There are numerous typographical errors, including 'Aknowledgements', 'ommit', 'POints', 'eXamples', and 'F act 2.2'; these should be corrected in a revision.","section":"Throughout"}],"recommendation":"reject","confidential_remarks":"The central theorem of the paper, Theorem 1.2, rests on Proposition 4.5, whose first step is false under the paper's own hypotheses: hypothesis (4.3) implies that limit points are not in the preserved domain, while the proof requires them to be in it. The second major issue is that Lemma 2.17 gives information about the i(Θ)-limit of a contracting sequence, not the Θ-limit, so the proper-domain case is not reduced to condition (4.3). These are not local or cosmetic defects. The causal convexity material and the example constructions may be salvageable in a substantially revised manuscript, but the advertised topological restriction is unsupported as written."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: the paper's headline result, Theorem 1.2, is not established as written. Proposition 4.5 contains a false step—the claim that limit points lie in the preserved domain Ω. From h_n·x → a with h_n·x ∈ Ω you only get a ∈ closure Ω, and the paper's own hypothesis (4.3) says Z_a ∩ Ω = ∅ for every a in the limit set; since a is non-transverse to itself, a ∈ Z_a, so (4.3) forces a ∉ Ω. The rest of the proof—choosing a component O containing Ω, placing x,y in O, deriving sΘ-invariance—depends on putting limit points inside Ω. So the constancy of type and the Maslov index zero conclusion are not proven.\n\nWhat the paper does well: the notion of causal convexity for Shilov boundaries (Definition 3.12) is new in this generality, and Proposition 3.21 (independence of affine chart) looks correct. The connection to dual convexity (Proposition 1.6) is a useful structural result. The constructions in Section 7 of Zariski-dense Anosov subgroups preserving proper domains are genuinely interesting and appear to rest on standard deformation arguments. The paper is well written and the literature is engaged honestly.\n\nThere is a second soft spot: Lemma 5.8 uses an unproved 'chain of photons' connecting any point of Ω to a limit point, and this assumption feeds into Theorem 1.7. The author may be able to prove it, but as it stands it is a gap in a different load-bearing argument. One note: the stress-test's secondary worry about Lemma 2.17 does not land, since i(Θ)=Θ in the self-opposite case, so (4.3) does follow from that lemma. The problem remains the first step.\n\nOverall: the framework and examples are likely salvageable, but the paper currently overclaims. A serious referee should see it, because the ideas matter, but I would not accept it in this form. The authors need to either fix Proposition 4.5 or weaken Theorem 1.2 accordingly.","headline":"Main theorem not proven as written—Proposition 4.5 makes a false inference placing limit points in Ω—but the causal convexity framework and examples are valuable enough to deserve referee time.","tokens_in":45349,"tokens_out":4496,"would_cite":true,"duration_ms":50450,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["22E46","22E40","53C35","57S30"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper's central claim is that preserving a proper domain in a self-opposite flag manifold forces one fixed relative-position type for all limit triples, with even rank and zero Maslov index in the Hermitian tube-type case.","keywords":["proper domains in flag manifolds","transverse subgroups","causal convexity","Shilov boundary","Maslov index","Anosov representations","Hermitian tube type","self-opposite flag manifolds"],"falsifier":"The paper's own hypotheses give a direct check: for a limit point a, the set Z_a contains a, while the hypothesis Z_a ∩ Ω = ∅ forces a ∉ Ω; if this contradiction cannot be resolved, the proposition is not proved. Separately, one concrete experiment that would settle the theorem is to find a transverse subgroup with a triple of nonzero Maslov index that nevertheless preserves a proper domain in an even-rank tube-type Shilov boundary.","tokens_in":44317,"feed_emoji":"📐","tokens_out":13039,"duration_ms":149416,"temperature":0.7,"pith_summary":"This paper studies subgroups of a real semisimple Lie group that preserve a proper domain in a flag manifold $G/P$, meaning a nonempty connected open subset whose closure avoids at least one Schubert divisor. The central claim is a necessary condition (Theorem 1.2): if $H$ preserves such a domain and its limit set in $G/P$ contains at least three pairwise transverse points, then every triple of limit points has the same \"type\" — the same connected component of the complement of the two Schubert divisors, up to the Levi symmetry. In the Hermitian tube-type case, where $G/P$ is the Shilov boundary and types are carried by the classical Maslov index, this forces the rank to be even and the Maslov index of every triple to vanish. The paper also introduces causal convexity in these Shilov boundaries, proves that dual convexity implies causal convexity, and shows that finitely generated transverse subgroups preserving proper domains are exactly those admitting a cocompact convex core with transverse ideal boundary. These results matter because they give topological obstructions to constructing $(G,G/P)$-manifolds as quotients $\\Omega/\\Gamma$, and the paper closes with Zariski-dense surface-group examples in even rank, showing the obstruction is sharp.","feed_headline":"Preserving a flag domain pins every limit triple to one type","feed_subtitle":"In Hermitian tube-type spaces this also forces even rank and zero Maslov index, blocking many group actions.","key_machinery":"The load-bearing object is the type of a triple of pairwise transverse points in a self-opposite flag manifold $G/P$. After moving two of the points to the opposite pair $P, P^-$, the third point lands in one connected component of $G/P \\setminus (Z_P \\cup Z_{P^-})$; the type is the orbit of that component under the Levi subgroup $P \\cap P^-$, and the involution $s$, which acts as negation in the standard affine chart, permutes these components. In the Hermitian tube-type case this type is the Maslov index $\\mathrm{idx}(a,b,c)$, and the conclusion of Theorem 1.2 is that only the $s$-invariant component can occur, forcing even rank and index $0$. The second machinery is causal convexity in the Shilov boundary: each affine chart carries future and past cones, two causally related points span a diamond $I^+(x) \\cap I^-(y)$, and a set is causally convex when it contains the closed diamond determined by any comparable pair; Proposition 3.21 makes this notion independent of the affine chart, so it is a genuine flag-manifold analogue of causal convexity in Lorentzian geometry.","core_discovery":"The paper's main theorem is Proposition 4.5 with Corollary 4.6, stated in the introduction as Theorem 1.2. In the author's formulation: let $G$ be a real semisimple Lie group, $P$ a self-opposite parabolic subgroup, and $H \\leq G$ a subgroup preserving a proper domain $\\Omega \\subset G/P$. If the limit set $\\Lambda_{\\Theta}(H)$ contains at least three pairwise transverse points, then there exists an $s$-invariant connected component $O$ of $G/P \\setminus (Z_{P} \\cup Z_{P^-})$ such that $\\mathrm{typ}(a,b,c) = [O]$ for every pairwise transverse triple of limit points. In the Hermitian tube-type case this says that the real rank $r$ is even and that $\\mathrm{idx}(a,b,c)=0$ for all triples of distinct limit points. The proof normalizes a pair of limit points to the base points $P$ and $P^-$, uses the fact that a preserved proper domain is contained in a single connected component of the complement of the Schubert divisors, and concludes that the type is constant and invariant under the opposition involution. The same ideas yield Theorem 1.7, an equivalence between (1) finite generation plus transverse dynamics plus preservation of a proper domain, and (2)/(3) existence of a causally convex or dually convex invariant domain with a convex core whose transverse ideal boundary has at least three points; in that situation all natural limit sets coincide.","pith_inferences":["A practical obstruction follows: for any candidate subgroup of a Hermitian tube-type group, computing the Maslov index of three transverse limit points gives a certificate — a nonzero value proves no proper domain is preserved, without constructing the domain.","The equality of limit sets in Theorem 1.7 suggests the quotient $\\Omega/\\Gamma$ carries a kind of causal compactness; if that extends to other causal or Nagano flag manifolds, it could give a general dictionary between transverse dynamics and convex core geometry.","The even-rank, zero-Maslov condition is reminiscent of \"spatial\" or acausal configurations in conformal Lorentzian geometry; one could test explicitly whether the preserved domain can always be foliated by acausal Cauchy hypersurfaces, as the Einstein-universe examples suggest."],"forward_implications":["Any discrete subgroup preserving a proper domain in a self-opposite flag manifold is forced to have one constant triple type on its limit set, so its limit points cannot realize several relative positions.","In Hermitian tube-type groups, no proper-domain-preserving subgroup can have odd real rank, and in even rank every transverse triple of limit points has Maslov index zero; highly twisted (maximal) surface representations are therefore excluded.","In Shilov boundaries, dually convex domains are causally convex and lie in an affine chart, so projective-style convexity and causal convexity coincide for domains preserved by such groups.","Theorem 1.7 says that acting cocompactly on a causally convex core with transverse ideal boundary is equivalent to being finitely generated, transverse, and preserving a proper domain; thus this natural definition of convex cocompactness does not distinguish Anosov subgroups from more general transverse subgroups.","Zariski-dense Anosov surface subgroups preserving proper domains exist in every even tube-type rank, and in some classical tube-type groups there are Zariski-dense examples that are neither free nor surface groups; the even-rank restriction is therefore sharp."],"supporting_citations":[{"why":"supplies the dual-convexity framework, the properness of Aut(Ω) actions, and Lemma 2.17 used to place limit points in the dual of a preserved domain.","marker":"[Zim18a]"},{"why":"introduces the opposition involution s and the component set EΘ whose s-invariant element Theorem 1.2 detects.","marker":"[DGR24]"},{"why":"classifies the connected components O_i of the standard chart minus a Schubert divisor, identifying the only s-invariant component in the tube-type case.","marker":"[Kan88]"},{"why":"defines the classical Maslov index that encodes the type of triples in Hermitian tube-type Shilov boundaries.","marker":"[LV80]"},{"why":"provides the orbit-structure fact (Fact 3.5) underlying the decomposition into the components O_i.","marker":"[Tak88]"},{"why":"gives the invariant properly convex cone in the unipotent radical and the projective-space necessary condition that motivates Proposition 4.5.","marker":"[Ben00]"},{"why":"supplies the representation-theoretic transfer of transversality and Anosov dynamics to projective space, used throughout the equivalence statements.","marker":"[GGKW17]"},{"why":"supplies the Anosov domain-of-discontinuity machinery and structural stability used in the openness of preservation of proper domains.","marker":"[GW12]"},{"why":"defines strong projective convex cocompactness and the Hilbert-metric tools that Section 5 adapts to flag manifolds.","marker":"[DGK24]"},{"why":"provides the Zariski-density result for surface-group deformations used to construct the examples of Theorem 1.4.","marker":"[KP15]"}],"fun_headline_variants":["Flag domain preservers lock every triple's type","Proper flag domains force constant triple types","Transverse groups: preserving a domain fixes limit triples","Even rank and zero Maslov index for tube-type domain preservers","Transverse subgroups with proper domains: rigidity in flag manifolds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of Proposition 4.5 relies on every limit point a of the limit set lying inside the preserved domain Ω, so that two limit points can be placed in the same connected component; yet the hypothesis that Ω avoids the Schubert divisor Z_a, together with a ∈ Z_a, appears to force a ∉ Ω.","fun_headline_variants_meta":{"raw":{"variants":["Flag domain preservers lock every triple's type","Proper flag domains force constant triple types","Transverse groups: preserving a domain fixes limit triples","Even rank and zero Maslov index for tube-type domain preservers","Transverse subgroups with proper domains: rigidity in flag manifolds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000953,"raw_usage":{"total_tokens":4121,"prompt_tokens":1055,"completion_tokens":3066,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":671,"completion_tokens_details":{"reasoning_tokens":2988}},"tokens_in":671,"tokens_out":3066,"duration_ms":26105,"temperature":1.0,"reasoning_tokens":2988,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T15:47:15.636730+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The paper's own hypotheses give a direct check: for a limit point a, the set Z_a contains a, while the hypothesis Z_a ∩ Ω = ∅ forces a ∉ Ω; if this contradiction cannot be resolved, the proposition is not proved. Separately, one concrete experiment that would settle the theorem is to find a transverse subgroup with a triple of nonzero Maslov index that nevertheless preserves a proper domain in an even-rank tube-type Shilov boundary.","supporting_citations":[],"review_version":1}