{"id":"6a1a01b2-2239-4379-803d-665c8d788c3d","arxiv_id":"2512.01086","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Automatic homeomorphicity lifts from automorphism groups to elementary-embedding monoids for all countable saturated structures, and a new ω-stable counterexample delimits the failure for endomorphism monoids.","lead":"This survey proves a new transfer theorem in topological reconstruction: for countable saturated structures, automatic homeomorphicity of the automorphism group implies automatic homeomorphicity of the monoid of elementary embeddings. It also constructs a very tame ω-categorical structure whose endomorphism monoid has a non-continuous automorphism despite its automorphism group having automatic continuity.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.7 depends on the unproved free Galois-closed factorization [47, Prop 5.1]; if that lemma fails for saturated structures with nontrivial center, the transfer collapses.","rationale":"The reader's verdict is CONDITIONAL and the weakest assumption is exactly where I would put the pressure. I agree: the transfer theorem is assembled from three external results (Lemma 4.1 from [18], Proposition 4.3 and Proposition 5.1 from [47]); among these, [47, Prop. 5.1] is the one that does work in the proof and is not even stated in this survey, so the reader cannot check it. If that factorization is not valid for arbitrary countable saturated structures — particularly those with nontrivial center, the case the paper claims to add — Proposition 4.4's φ cannot be constructed, and Corollary 4.6 is moot. I verified the paper's own Lemma 4.5 appears correct (the terse chain construction can be made precise). I also checked the assertion in §2.2 that EEmb(A)=Aut(A) for countable saturated structures; this is false (e.g., (N,=) has non-surjective elementary embeddings), but it is not used in the proof of Theorem 4.7, so I do not base the verdict on it. Since the reader's CONDITIONAL verdict already reflects exactly this dependency, I recommend keeping the verdict unchanged.","tokens_in":23046,"tokens_out":34627,"duration_ms":338441,"concrete_test":"Extract the proof of [47, Proposition 5.1] and verify it applies to every countable saturated structure. Then test the factorization concretely on the saturated structure (Ω,E) with countably many E-classes of size 2 (whose automorphism group has nontrivial center): for the non-surjective elementary embedding that omits one E-class, construct free Galois-closed g,h with g = h∘f. If the factorization is impossible (or the lemma's hypotheses fail), Proposition 4.4 and Theorem 4.7 are false.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The main proof of Theorem 4.7 is not self-contained: Proposition 4.4 uses [47, Proposition 5.1] to factor every f ∈ Ḡ as g = h∘f with g,h free Galois-closed, and Proposition 4.3 imports [47, Cor. 3.12/Prop. 3.13] to conclude Φ(h) = h∘z_h for such h. If [47, Prop. 5.1] actually requires stronger hypotheses than countable saturation — e.g., superhomogeneity in the sense of [47], or trivial center — then Proposition 4.4's construction of the homomorphism φ:Ḡ→Z(G) with Φ(f) = φ(f)∘f has no basis, and Corollary 4.6 cannot be applied. The paper itself defines 'free Galois-closed' by translating [47]'s 'superhomogeneity', which invites exactly this worry. That is the single chokepoint: all new content (Prop 4.4, Cor 4.6, Theorem 4.7) relies on the truth of that factorization for the full class C^sat_G. The rest of the argument (Lemma 4.5) is proved in the text and appears sound. This is an external dependency, not an internal inconsistency, but it is load-bearing: a counterexample to [47, Prop. 5.1] would invalidate Theorem 4.7.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper is a survey of topological reconstruction for automorphism groups, monoids of elementary embeddings, endomorphism monoids, and polymorphism clones of countable first-order structures, with emphasis on the omega-categorical case. It organizes the notions of automatic continuity, automatic homeomorphicity, the unique Polish property, the Zariski topology, and automatic action compatibility, and it provides a diagram of known implications. The main new result is Theorem 4.7: for a countable saturated structure whose automorphism group has automatic homeomorphicity with respect to closed permutation groups, the monoid of elementary embeddings has automatic homeomorphicity with respect to closed transformation monoids. A secondary contribution is Theorem 4.9, an omega-categorical, omega-stable structure whose automorphism group has strong automatic continuity but whose endomorphism monoid has a discontinuous automorphism. The remaining sections survey existing techniques and open problems.","tokens_in":23355,"tokens_out":23405,"duration_ms":238847,"significance":"If Theorem 4.7 is correct, it is a genuine advance: it removes the trivial-centre/superhomogeneity assumptions that were present in earlier work of Pech and Pech and provides a clean transfer from automorphism groups to monoids of elementary embeddings for all countable saturated structures. The paper also performs a useful service by collecting and polishing results scattered across the literature, making the landscape of topological reconstruction more accessible. The proof of Theorem 4.7 is not circular and the external dependencies are clearly cited. However, the proof rests on at least two points that are not fully justified in the manuscript: the exact scope of a cited factorization lemma from [47], and the claim that central automorphisms are central in the monoid of elementary embeddings. These need clarification before the paper can be accepted.","major_comments":[{"comment":"The proof of Proposition 4.4 invokes [47, Proposition 5.1] to factor every f in Gbar as g = h ∘ f with free Galois-closed g,h in Gbar. This is the single load-bearing external step of Theorem 4.7: if that factorization carries hidden hypotheses (e.g., superhomogeneity in the stronger sense of [47], or applicability only to the canonical homogeneous companion A' rather than to the arbitrary saturated structure A), then the construction of the homomorphism φ has no basis and Corollary 4.6 cannot be applied. Since Definition 4.2 explicitly translates [47]'s notion of superhomogeneity, the authors should state [47, Proposition 5.1] in full, with its hypotheses, and explain why it applies to every f ∈ EEmb(A) for an arbitrary G ∈ C^sat_G. As written, this is an unverified chokepoint, not an internal inconsistency.","section":"Section 4.1, Proposition 4.4"},{"comment":"After cancelling the injective map h, the proof uses the assertion that z_h, z_g belong to Z(Gbar) 'being in Z(G)'. For arbitrary saturated structures it is not immediate that an automorphism central in Aut(A) commutes with every elementary embedding of A. The manuscript gives no proof or reference for this commutation. If Z(G) is not contained in Z(Gbar), then the step from f ∘ z_g = z_h ∘ Φ(f) to Φ(f) = z_h^{-1} z_g ∘ f is unjustified and the whole argument collapses. The claim is plausible and can be proved by an orbit argument for canonical structures, but it needs to be stated and proved in the generality used here.","section":"Section 4.1, Proposition 4.4, displayed equation"},{"comment":"The model-theoretic chain construction is too abbreviated. The sentence 'Consider the type p(z;A0):=tp(A1/A0) ... Let A2 be a realisation of p(x;α0(A0))' does not, as written, ensure that the resulting isomorphism α1 extends α0; one needs to take the pushforward of tp(A1/A0) under α0 and then use saturation or compactness to obtain an extension. In addition, the assertion that the union of a countable elementary chain of countable saturated models is again saturated is used without proof or citation. Since Corollary 4.6 and hence Theorem 4.7 depend on this lemma, the construction should be written out in detail.","section":"Lemma 4.5"}],"minor_comments":[{"comment":"In the statement of the 'Moreover' part, the notation is inconsistent: θ is said to be an injective homomorphism from G to Ω^Ω, but in the preceding sentence G denotes the monoid of elementary embeddings; the overline on G appears to be missing. Please fix the notation.","section":"Theorem 4.7"},{"comment":"Typo: 'Polishtopological group' should be 'Polish topological group'.","section":"Section 2.1"},{"comment":"The notation C^sat_G, C^sat_G, C^sat_T, C^sat_C is dense and the subscript/superscript conventions are easy to confuse. A small table with the classes and their inclusions would help the reader, especially in the diagram in Figure 1.","section":"Definition 2.12"},{"comment":"The proof states that α commutes with all elements of Aut(A'). This is true for the flip of all two-element E-classes, but it would be helpful to spell out the centralizer argument, since the rest of the verification that η is an automorphism depends on it.","section":"Theorem 4.9"}],"recommendation":"major_revision","confidential_remarks":"The main risk is the dependence of Theorem 4.7 on [47, Proposition 5.1] and on the unproved claim that Z(Aut(A)) ⊆ Z(EEmb(A)). If the authors can supply precise statements and proofs for these two points, the paper is likely publishable. The survey aspect is solid and the new counterexample in Theorem 4.9 is interesting. I would be happy to see a revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a useful survey of topological reconstruction for endomorphism monoids and clones, and there are two genuinely new items: Theorem 4.7, which claims to lift automatic homeomorphicity from Aut(A) to EEmb(A) for all countable saturated structures, and Theorem 4.9, which gives a tame counterexample to automatic homeomorphicity for End(A). The survey portions are well organized, Figure 1 is a helpful map, and the authors are honest about open questions. Credit is due for that.\n\nThat said, I would not take Theorem 4.7 as established. The proof of Lemma 4.5 is not just abbreviated; it appears to prove too much. The chain construction produces an isomorphism f:A'→A and derives α∘f∘h'=f∘β on A'. The authors say this proves the lemma, but the lemma requires f to be a self-embedding of A, not an isomorphism between A' and A. Conjugating the equation back to A yields h=γ^{-1}β, i.e., h is an automorphism. The argument, if valid, would show EEmb(A)=Aut(A) for every countable saturated structure. That is false: the countable dense linear order is saturated, and x↦x+1 is an elementary non-surjective embedding. The earlier statement in §2.2 that \"In countable saturated structures we have EEmb(A)=Aut(A)\" is the same false claim, and it contradicts the authors' own ω-categorical examples in §4.2, which are saturated and have non-surjective elementary embeddings.\n\nSecond, the external dependency is real: Proposition 4.4 imports [47, Proposition 5.1] verbatim, and if that factorization requires trivial center or extra homogeneity, the transfer collapses. I did not find an internal fix in the text.\n\nThe Zariski topology and Property X sections are competent and would help someone entering the area. Theorem 4.9's construction looks plausible and is less affected by the above, though it is still asserted in places.\n\nRecommendation: send it to a serious referee, but only with the clear message that Theorem 4.7 depends on repairing Lemma 4.5. The survey may be publishable after the claims are corrected.","headline":"Theorem 4.7 is the kind of result worth wanting, but the proof of Lemma 4.5 has a load-bearing error, and the survey's claim that countable saturated structures have EEmb=Aut is simply false.","tokens_in":23856,"tokens_out":16247,"would_cite":false,"duration_ms":145306,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03C35","03C50","08A40","20B27","20M20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that for countable saturated structures, automatic homeomorphicity of the automorphism group automatically transfers to the monoid of elementary embeddings, so algebraic isomorphisms between such monoids are automatically","keywords":["topological reconstruction","automatic homeomorphicity","elementary embedding monoid","endomorphism monoid","polymorphism clone","ω-categoricity","saturated structures","bi-interpretability"],"falsifier":"A single concrete counterexample would settle the central claim: a countable saturated structure A whose automorphism group has automatic homeomorphicity with respect to closed permutation groups, together with an algebraic isomorphism from EEmb(A) onto a closed transformation monoid that is not a homeomorphism. Internally, the weakest link to test is the factorization lemma: for instance, try f(x)=2x on the countable dense linear order (Q,<), and check whether f can be written as g=h∘f with free Galois-closed g,h; a failure there breaks Proposition 4.4 and with it the proof of Theorem 4.7.","tokens_in":22909,"feed_emoji":"🔁","tokens_out":15815,"duration_ms":146858,"temperature":0.7,"pith_summary":"This paper is a guide to topological reconstruction for spaces of symmetries, and its main new result is a transfer theorem. For any countable saturated structure, if its automorphism group is so rigid that every algebraic isomorphism to a closed permutation group is automatically a homeomorphism, then the same rigidity holds for its monoid of elementary embeddings: every algebraic isomorphism to a closed transformation monoid is automatically a homeomorphism. The proof isolates the monoid of elementary embeddings as an intermediate object and shows that the only endomorphism of it fixing the automorphism group pointwise is the identity. The paper also polishes several earlier results, including a very tame counterexample showing that full endomorphism monoids need not inherit this rigidity even when the automorphism group is maximally well-behaved. A sympathetic reader should care because automatic homeomorphicity is the precise sense in which purely algebraic information, the monoid structure, determines the topology of these symmetry spaces, and the theorem extends that guarantee to a naturally larger class.","feed_headline":"Automatic homeomorphicity lifts to elementary-embedding monoids","feed_subtitle":"For countable saturated structures, automorphism-group rigidity forces it on elementary-embedding monoids.","key_machinery":"Key object: the monoid of elementary self-embeddings EEmb(A), with its automorphism group G as the invertible elements. The load-bearing tool is the class of 'free Galois-closed' elementary embeddings: images whose tuple orbits under the setwise stabiliser are just the restrictions of G-orbits, and whose pointwise stabiliser fixes nothing outside. A prior factorization lemma writes any f as g=h∘f with g,h free Galois-closed, so any endomorphism Φ fixing G pointwise has the form Φ(f)=φ(f)∘f for a homomorphism φ: EEmb(A)→centre(G) vanishing on G. A saturation 'commuting-square' argument forces φ trivial, hence Φ is the identity; Lemma 4.1 turns this into automatic homeomorphicity of EEmb(A).","core_discovery":"The central claim (Theorem 4.7): for a countable saturated structure A, if Aut(A) has automatic homeomorphicity with respect to closed permutation groups, then the monoid EEmb(A) of elementary self-embeddings has automatic homeomorphicity with respect to all closed transformation monoids. Moreover, any injective monoid homomorphism from EEmb(A) into the full transformation monoid with image closed in the symmetric group is automatically a homeomorphism onto its image. The proof shows that the only endomorphism of EEmb(A) fixing Aut(A) pointwise is the identity, using a factorization through 'free Galois-closed' maps and a saturation argument forcing centre-valued homomorphisms to be trivial.","pith_inferences":["The transfer theorem suggests a division of labour: to prove automatic homeomorphicity for elementary-embedding monoids, it is now enough to prove it for the underlying automorphism group, where tools such as the small index property and ample generics are already available; this reverses the usual direction of effort.","Coupled with the folklore correspondence between topological isomorphism of elementary-embedding monoids and first-order bi-interpretability, the theorem implies that under its hypotheses an algebraic isomorphism between two such monoids already witnesses bi-interpretability, giving a purely algebraic route to a model-theoretic conclusion.","The counterexample for endomorphism monoids suggests that the elementary-embedding monoid, not the full endomorphism monoid, is the right intermediate object for general transfer: hypotheses strong enough to force automatic homeomorphicity of Aut(A) must be supplemented by extra structural information before they say anything about End(A).","A natural testable extension is whether the two internal ingredients, the factorization lemma and the commuting-square construction, hold outside the countable saturated setting, for example for saturated structures in uncountable languages or for non-saturated structures with quantifier elimination."],"forward_implications":["For every countable saturated structure, automatic homeomorphicity of Aut(A) among closed permutation groups implies automatic homeomorphicity of EEmb(A) among all closed transformation monoids.","Any injective monoid homomorphism from EEmb(A) whose image is a closed subgroup of the full symmetric group is automatically a homeomorphism onto its image; the image itself is a closed permutation group.","For countably categorical structures satisfying the G-finite condition, a mild finite-intersection hypothesis on open subgroups, the elementary-embedding monoid has automatic homeomorphicity with respect to closed transformation monoids, extending earlier results that required a trivial-centre assumption.","The full endomorphism monoid does not inherit this: there is a countably categorical, ω-stable structure, definable in a finitely homogeneous structure, whose automorphism group has automatic continuity with respect to all second-countable groups, yet whose endomorphism monoid has a discontinuous automorphism."],"fun_headline_variants":["Automatic homeomorphicity lifts to elementary-embedding monoids","Saturated A: Aut rigidity forces EEmb automatic homeomorphicity","From Aut to EEmb: automatic homeomorphicity lifts","Topological reconstruction guide for monoids and clones"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise, imported from the literature without proof, is that every elementary embedding f of a countable saturated structure factors as g = h∘f with g and h free Galois-closed; if that fails, the proof that the only automorphism-group-fixing endomorphism of the embedding monoid is the identity, and hence Theorem 4.7, collapses.","fun_headline_variants_meta":{"raw":{"variants":["Automatic homeomorphicity lifts to elementary-embedding monoids","Saturated A: Aut rigidity forces EEmb automatic homeomorphicity","From Aut to EEmb: automatic homeomorphicity lifts","Topological reconstruction guide for monoids and clones"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001233,"raw_usage":{"total_tokens":4867,"prompt_tokens":676,"completion_tokens":4191,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":420,"completion_tokens_details":{"reasoning_tokens":4122}},"tokens_in":420,"tokens_out":4191,"duration_ms":29599,"temperature":1.0,"reasoning_tokens":4122,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T19:17:47.341440+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single concrete counterexample would settle the central claim: a countable saturated structure A whose automorphism group has automatic homeomorphicity with respect to closed permutation groups, together with an algebraic isomorphism from EEmb(A) onto a closed transformation monoid that is not a homeomorphism. Internally, the weakest link to test is the factorization lemma: for instance, try f(x)=2x on the countable dense linear order (Q,<), and check whether f can be written as g=h∘f with free Galois-closed g,h; a failure there breaks Proposition 4.4 and with it the proof of Theorem 4.7.","supporting_citations":[],"review_version":1}