{"id":"c94a126a-5ef4-466d-bb75-fd9d6476f3e9","arxiv_id":"2607.19193","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"An abelian group has a topologically Hopfian profinite completion exactly when A/pA is finite for every prime p; this answers Kourovka Problem 6.30 in the negative.","lead":"This paper proves that the profinite completion of an abelian group is topologically Hopfian if and only if, for every prime p, the quotient A/pA is finite. It also supplies a counterexample to a long-standing open problem in group theory, showing that a Hopfian residually finite group can have a non-Hopfian profinite completion.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the proof is sound given the standard Kulikov theorem; no load-bearing gap found.","rationale":"The reader's verdict of ACCEPT is warranted. The proof of the main theorem is built from standard structural ingredients, and the claimed equivalences follow rigorously. I specifically re-examined the two places where the argument is most exposed: the existence of p-basic subgroups (Theorem 2.4, cited from Fuchs) and the extension/isomorphism lemma (Lemma 2.5). Both are applied correctly, and the internal proofs of Lemma 2.5 and Proposition 3.1 are sound. The counterexample to Kourovka Problem 6.30 is also correct: the direct sum of distinct localizations is Hopfian and residually finite, while its 2-completion admits the standard shift endomorphism. No load-bearing concern surfaced, so no change to the reader's verdict is needed.","tokens_in":4015,"tokens_out":31323,"duration_ms":308747,"concrete_test":"As a standalone verification, compute Lemma 2.5 for A=Z[1/3]⊕Z(2^∞), p=2, B=Z, F=C_4: confirm Hom(A,F)→Hom(B,F) is bijective by hand and verify \\hat B_2≅\\hat A_2. For the counterexample, check that the left-shift map on ∏_{i≥1} Z_2, the 2-completion of ⊕ Z[1/q_i], is continuous, surjective, and non-injective.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on external p-basic subgroup theory (Theorem 2.4) and Lemma 2.5, especially the bijection Hom(A,F)≅Hom(B,F) for finite p-groups and the resulting isomorphism \\hat B_p≅\\hat A_p. I checked these carefully. The extension proof in Lemma 2.5(i) is valid: p-purity gives well-definedness, p-divisibility of A/B gives existence, and the exponent of F kills the p^m term. The density and injectivity arguments in Lemma 2.5(ii) are also valid: B maps onto every finite p-quotient of A, and every finite p-quotient of B extends to A. Proposition 3.1's shift construction is correct for any infinite p-basic direct sum; the endomorphism is continuous because the preimage of an open subgroup has p-power index, and the compact-image argument is standard. The counterexample's residual finiteness and Hopficity, though asserted as 'clearly', are true: distinct q_i make different summands not q_i-divisible, so endomorphisms preserve summands. I found no internal inconsistency or unsupported step in the main theorem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper gives an exact criterion for the profinite completion of an abelian group to be topologically Hopfian. Main Theorem: \\(\\widehat A\\) is topologically Hopfian iff \\(A/pA\\) is finite for every prime \\(p\\) iff every Sylow pro-\\(p\\) subgroup of \\(\\widehat A\\) is topologically finitely generated. The proof passes through p-basic subgroups: after reducing to a p-basic subgroup \\(B\\), the quotient \\(A/pA\\) controls finite generation of \\(B\\); if \\(B\\) is infinite, the authors construct an explicit shift endomorphism of \\(\\widehat B_p\\) that is surjective but not injective. The paper also gives a counterexample to Kourovka Problem 6.30: for distinct odd primes \\(q_i\\), the group \\(G=\\bigoplus_{i\\ge 1}\\mathbb Z[1/q_i]\\) is residually finite and Hopfian, but \\(\\widehat G\\) is not topologically Hopfian because \\(G/2G\\) is infinite.","tokens_in":4306,"tokens_out":18773,"duration_ms":166186,"significance":"The characterization is complete and checkable, and the counterexample resolves a long-standing question of Mel'nikov. The proof is self-contained modulo standard theorems (Kulikov's p-basic subgroup theorem, Hopficity of finitely generated profinite groups). No parameters are fitted, no external results are assumed beyond standard ones, and the arguments in Lemma 2.5 and Proposition 3.1 are detailed and correct. The shift construction for infinite p-basic subgroups is elegant and directly yields the non-Hopfian completion. This is a valuable contribution to the theory of profinite completions and Hopfian groups.","major_comments":[],"minor_comments":[{"comment":"The assertions that \\(G=\\bigoplus_i \\mathbb Z[1/q_i]\\) is 'clearly Hopfian and residually finite' are true, but a short justification would be helpful. In particular, for \\(i\\neq j\\) there are no nonzero homomorphisms \\(\\mathbb Z[1/q_i]\\to \\mathbb Z[1/q_j]\\), so every endomorphism preserves the direct summands; each summand is Hopfian. A one-sentence explanation would also clarify why this example is not vacuous (compare with \\(\\bigoplus_{i\\ge1}\\mathbb Z\\), which is not Hopfian).","section":"Introduction, counterexample"},{"comment":"The notation 'Zrp⊕Fp' is misleading: it should be \\(\\mathbb Z^r \\oplus F_p\\), not \\(\\mathbb Z_p^r \\oplus F_p\\). Please correct the typesetting.","section":"Theorem 3.2"},{"comment":"The proof of Lemma 2.2 is omitted. Since this lemma is used in the proof of the Main Theorem, a two-line argument (continuous endomorphisms preserve the Sylow pro-\\(p\\) subgroups, and surjectivity on the product is equivalent to surjectivity on each factor) would make the paper more self-contained.","section":"Lemma 2.2"},{"comment":"The sentence 'Clearly, \\(\\sigma\\) is continuous with respect to the pro-\\(p\\) topology on \\(B\\)' is correct, but a brief parenthetical explanation (the preimage of an open subgroup of \\(p\\)-power index has \\(p\\)-power index, hence is open) would help the reader.","section":"Proposition 3.1"},{"comment":"In the injectivity argument, the statement 'some finite \\(p\\)-quotient of \\(B\\) does not annihilate \\(x\\)' relies on the fact that the kernels of all finite \\(p\\)-quotients have trivial intersection. This standard fact could be stated explicitly.","section":"Lemma 2.5(ii)"},{"comment":"Theorem 2.4 is cited to 'the chapter on purity and basic subgroups in [1]'; a more precise theorem number or page reference would be convenient.","section":"References"}],"recommendation":"accept","confidential_remarks":"This is a clean and correct paper. The central proof is sound, the counterexample works, and the presentation is clear. The only issues are minor presentational and typographical; I recommend acceptance without further technical revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the main theorem is true and the paper earns its accept. The characterization of Hopficity of profinite completions of abelian groups in terms of finiteness of A/pA is new and answers a 1978 Kourovka problem in the negative. The proof is largely built on Kulikov's p-basic subgroup theorem, which is standard and applied correctly. The genuinely original mechanism is Proposition 3.1, the shift endomorphism on an infinite p-basic direct sum; it is the right construction and the continuity argument checks out.\n\nThe paper does well by spelling out Lemma 2.5 in detail: the extension of homomorphisms from a p-basic subgroup to finite p-groups is carefully justified, and the isomorphism of pro-p completions follows properly. The density/closed-image argument is sound. Giving the equivalence with all Sylow pro-p subgroups topologically finitely generated is a nice bonus.\n\nSoft spots are minor. The counterexample G = ⊕ Z[1/q_i] is said to be \"clearly\" Hopfian and residually finite. It is true, but a one-sentence justification would make the paper more self-contained. Also, the Hopfian-implies-finiteness direction depends entirely on Proposition 3.1; that is fine, but the paper could acknowledge this is the only non-structural input. The reliance on Kulikov's theorem is significant but standard, and the authors use it exactly as intended.\n\nI found no load-bearing flaw. The stress test checked the extension lemma and the shift construction; both hold. The citation pattern is honest: no self-citation, no fitting. The external dependencies are classical and correctly applied. If the referee verifies Lemma 2.5 (which is proved) and the application of Kulikov, the theorem is solid.\n\nWho is this for? Any group theorist working with Hopfian groups, profinite completions, or abelian groups. It is a clean result that will be a useful reference. I would not be surprised if the proof can be streamlined, but it is already concise.\n\nRecommendation: send it to a serious referee. The result is significant enough and the proof looks correct. I would accept with minor revisions, mostly the one-sentence justification of the example.","headline":"A clean, correct characterization that answers a 1978 problem; the proof is solid, and only minor presentation nits remain.","tokens_in":4689,"tokens_out":1905,"would_cite":true,"duration_ms":18059,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20E18","20K20","22C05"],"pacs":[],"model":"deepseek-v4-flash","headline":"For abelian groups, the profinite completion is topologically Hopfian exactly when A/pA is finite for every prime p, and this criterion refutes a long-standing open problem.","keywords":["Hopfian group","profinite completion","abelian group","pro-p completion","p-basic subgroup","topologically Hopfian","Kourovka Problem 6.30","residually finite"],"falsifier":"A single abelian group with A/pA finite for every p but with a surjective non-injective continuous endomorphism of its profinite completion would refute the theorem, as would a group with A/pA infinite for some p whose pro-p completion is nonetheless Hopfian. In the specific example A = ⊕_{i≥1} Z[1/q_i] with distinct odd primes, A/2A is infinite; the theorem predicts a non-Hopfian 2-adic completion, so either constructing or ruling out such an endomorphism would settle the prediction.","tokens_in":3974,"feed_emoji":"","tokens_out":7248,"duration_ms":59505,"temperature":0.7,"pith_summary":"The paper pins down exactly when the profinite completion of an abelian group is topologically Hopfian: precisely when the quotient A/pA is finite for every prime p, equivalently when every Sylow pro-p subgroup of the completion is topologically finitely generated. This reduces a question about a potentially enormous completion to a check on finite quotients of the original group. The proof works prime by prime via p-basic subgroups, showing that an infinite ladder of summands in a p-basic subgroup yields a surjective non-injective endomorphism of the p-adic completion. As a byproduct, the criterion answers Kourovka Problem 6.30 in the negative: a residually finite Hopfian group can have a profinite completion that is not topologically Hopfian.","feed_headline":"Profinite completion is Hopfian iff A/pA is finite for all primes","feed_subtitle":"The criterion settles Kourovka Problem 6.30: a Hopfian residually finite group can have a non-Hopfian profinite completion.","key_machinery":"The load-bearing object is the p-basic subgroup B of A (Kulikov's theorem): B is a direct sum of infinite cyclic groups and finite cyclic p-groups, p-pure in A, with p-divisible quotient A/B. Lemma 2.5 shows that restriction to B is a bijection on homomorphisms into finite p-groups, so the pro-p completions of A and B are isomorphic and A/pA ≅ B/pB. In the infinite-rank case, Proposition 3.1 constructs a surjective non-injective endomorphism of the pro-p completion by assigning each summand a 'height' (∞ for infinite cyclic, n for C_{p^n}), picking an infinite non-decreasing ladder, shifting each selected summand to the next, and letting the universal property of pro-p completion extend the","core_discovery":"The central claim is a complete classification: for an abelian group A, the profinite completion  is topologically Hopfian if and only if A/pA is finite for every prime p. Because  decomposes as the direct product of its Sylow pro-p subgroups, the proof is prime-by-prime: Â_p is Hopfian exactly when A/pA is finite, equivalently when Â_p is topologically finitely generated. The transfer device is a p-basic subgroup B of A, which shares both the pro-p completion and the mod-p reduction with A. If B has infinitely many cyclic summands, an endomorphism of the pro-p completion is built by shifting indices of non-decreasing p-heights and killing one summand, producing a surjective but non-inject","pith_inferences":["The same local criterion may serve as a sufficient condition for Hopficity of profinite completions in larger classes, such as finitely generated nilpotent or polycyclic groups, where the mod-p abelianization could play the role of A/pA; the abelian proof suggests the completion is non-Hopfian whenever the mod-p abelianization is infinite.","The counterexample shows that 'residually finite plus Hopfian' is not a robust combination under profinite completion; a more natural sufficient condition for a Hopfian completion is that each pro-p Sylow subgroup is topologically finitely generated, exactly the third condition of the Main Theorem.","The height-shift construction is a general recipe for producing non-injective surjective endomorphisms in completions of infinite-rank abelian groups; a similar ladder argument might yield analogous examples in nonabelian pro-p groups with an infinite descending chain of open subgroups, though the abelian structure is used crucially.","Because the criterion is purely in terms of finite quotients A/pA, it is effectively computable for finitely presented abelian groups via Smith normal form; this suggests treating Hopficity of profinite completions as a decidable property in that class."],"forward_implications":["Topological Hopficity of the profinite completion of an abelian group is equivalent to topological finite generation of each Sylow pro-p subgroup.","The Kourovka Problem 6.30 answer is negative: Hopfian and residually finite do not imply the profinite completion is Hopfian; the explicit group is a direct sum of localizations Z[1/q_i] with distinct odd primes.","For abelian p-groups, a Hopfian profinite completion forces the completion to be finite and the group to split as D ⊕ F with F finite (Corollary 3.3).","For direct sums of localizations A = ⊕_{i} Z[S_i^{-1}], Hopficity of the completion is equivalent to finiteness, for each prime p, of the set of indices i with p not in S_i (Corollary 3.4).","In the abelian setting, non-Hopficity of the completion is detected already by an infinite elementary abelian p-quotient A/pA."],"fun_headline_variants":["Hopfian profinite completion iff A/pA finite for each prime p","Abelian group's profinite completion Hopfian exactly when mod-p quotients finite","Kourovka 6.30 answered: Hopfian group can have non-Hopfian completion","Hopfian profinite completion: A/pA finite for every prime","Problem 6.30 refuted: Hopfian group may have non-Hopfian completion"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Everything depends on Kulikov's theorem guaranteeing that every abelian group has a p-basic subgroup and on the extension lemma that every homomorphism from that subgroup to a finite p-group lifts to the whole group; if either fails, the identification of the pro-p completions of A and B breaks and the criterion may not transfer.","fun_headline_variants_meta":{"raw":{"variants":["Hopfian profinite completion iff A/pA finite for each prime p","Abelian group's profinite completion Hopfian exactly when mod-p quotients finite","Kourovka 6.30 answered: Hopfian group can have non-Hopfian completion","Hopfian profinite completion: A/pA finite for every prime","Problem 6.30 refuted: Hopfian group may have non-Hopfian completion"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000962,"raw_usage":{"total_tokens":3901,"prompt_tokens":677,"completion_tokens":3224,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":421,"completion_tokens_details":{"reasoning_tokens":3122}},"tokens_in":421,"tokens_out":3224,"duration_ms":21612,"temperature":1.0,"reasoning_tokens":3122,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T13:11:35.617407+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single abelian group with A/pA finite for every p but with a surjective non-injective continuous endomorphism of its profinite completion would refute the theorem, as would a group with A/pA infinite for some p whose pro-p completion is nonetheless Hopfian. In the specific example A = ⊕_{i≥1} Z[1/q_i] with distinct odd primes, A/2A is infinite; the theorem predicts a non-Hopfian 2-adic completion, so either constructing or ruling out such an endomorphism would settle the prediction.","supporting_citations":[],"review_version":1}