{"id":"40dab7da-3b8c-4f16-a59d-72ff0b935a71","arxiv_id":"1908.03415","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a precompact abelian torsion group G, a nontrivial convergent sequence exists in the pointwise dual G_p^ exactly when G has a countably infinite Hausdorff quotient.","lead":"Mathematicians found exactly when the dual object of a precompact abelian torsion group contains a sequence of characters that steadily approaches the trivial character: this happens precisely when the group has a countably infinite quotient. They also built a dense subgroup of the Cantor group that is first category and measure zero, yet whose dual has no nontrivial convergent sequences, yielding a first-category reflexive abelian group.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Prop 2.5(3)⇒(2) hinges on an unquoted [25] biduality theorem; if that theorem gives only a dense embedding, the quotient argument and Theorem 2.7 fail as written.","rationale":"The reader and I converge on the same external dependency. The rest of the proof of Theorem 2.7 is internally coherent: (4)⇒(2) uses torsion to force finite-order values into finite subgroups, and (1)⇒(3)/(3)⇒(4) are standard. The abstract's 'No infinite quotient of G is countable' sentence is inconsistent with condition (1); it is probably a typo but should be fixed. I do not see an internal mathematical error in the main construction. The verdict remains CONDITIONAL because the [25] theorem is load-bearing and unquoted; if verification shows the theorem applies as stated, the paper should be accepted.","tokens_in":10395,"tokens_out":37151,"duration_ms":449858,"concrete_test":"Retrieve [25] and check the exact statement used: for a precompact abelian group G, is the canonical evaluation e_G : G → (G_p^)_p^ a topological isomorphism, or only an injective dense embedding? If the citation is correct, verify surjectivity directly on G = Q/Z (dense in T): G^ = Z with t_p(G), the subgroup Γ = Z is metrizable, Γ^⊥ = {0}; compute whether e_G followed by the restriction map sends G onto Γ^ = Z. If the image is a proper subgroup, the proof of (3)⇒(2) collapses for this example. If the image is all of Γ^, the step survives for this test case, and the citation should be inspected for any additional hypotheses (e.g. completeness, torsion, or dual-closedness).","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing step is in Proposition 2.5, implication (3)⇒(2). The proof takes a closed infinite metrizable subgroup Γ of G_p^, uses the surjectivity of i^ : (G_p^)_p^ → Γ_p^, and then invokes [25] to identify G with (G_p^)_p^. From this the authors conclude that G/Γ^⊥ is isomorphic to Γ^. Everything after that depends on this isomorphism. The manuscript neither states the theorem from [25] nor verifies its hypotheses; it is used again in Theorem 3.9. If [25] provides only a canonical topological embedding or dense image of G into its second dual — not a surjective isomorphism — then the composed map G → Γ_p^ need not be onto, and Γ^ need not be a quotient of G. In that case Proposition 2.5(3)⇒(2) and the (2)-(4) cycle of Theorem 2.7 would not follow as written. This is a genuine external-dependency concern rather than an internal contradiction; the underlying theorem may well be true. There is also a separate abstract/statement mismatch: the abstract says the torsion characterization is 'No infinite quotient group of G is countable', which negates condition (1) of Theorem 2.7 (G has a countably infinite Hausdorff quotient); this appears to be a typo but would mislead a reader.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies precompact abelian groups whose dual group, equipped with the pointwise convergence topology, contains a nontrivial convergent sequence. The main theorem (Theorem 2.7) states that for a precompact torsion abelian group, four conditions are equivalent: (1) G has a countably infinite Hausdorff quotient, (2) G has a countably infinite Hausdorff homomorphic image, (3) the dual group G_p^ contains an infinite metrizable subgroup, and (4) G_p^ contains a nontrivial convergent sequence. The paper also constructs, in Theorem 2.3, a dense subgroup of Z(2)^ω that is of the first category, has measure zero, and whose dual contains no infinite compact subsets; as a consequence it obtains a first-category precompact reflexive abelian group in Theorem 3.9. Section 3 develops a general criterion for the presence of compact subsets in dual groups in terms of separating subgroups of Cp(K, T).","tokens_in":10671,"tokens_out":63802,"duration_ms":675288,"significance":"If the main results hold, Theorem 2.7 gives a clean internal characterization of convergent sequences in duals of torsion precompact abelian groups, reducing the phenomenon to a countable quotient property of the original group. The construction in Theorem 2.3 and the reflexive first-category group in Theorem 3.9 are interesting and complement the earlier work of Hart and Kunen. The paper is largely self-contained in its combinatorial construction, uses standard duality tools, and contains no fitted parameters or circular arguments. The proofs are mostly direct, and the overall contribution is a useful step in understanding compactness and convergence in duals of precompact groups.","major_comments":[{"comment":"The abstract and the introduction both state that the torsion characterization is 'No infinite quotient group of G is countable.' This is the negation of condition (1) of Theorem 2.7, which asserts that the existence of a nontrivial convergent sequence is equivalent to G having a countably infinite Hausdorff quotient. As written, the advertised characterization is the exact opposite of the theorem proved in the paper. This is a load-bearing misstatement of the main result and must be corrected in both places.","section":"Abstract and Introduction"},{"comment":"The proof of Proposition 2.5(3)⇒(2) invokes the unstated theorem from [25] that 'the canonical evaluation mapping of G to (G_p^)_p^ is a topological isomorphism.' The same external result is used again in the proof of Theorem 3.9. The authors neither quote the theorem nor verify its hypotheses for the groups in question. This is a load-bearing step because it is used to identify G with its double dual and then to conclude G/Γ^⊥ ≅ Γ^. If [25] provides only a topological embedding or requires additional hypotheses, the argument fails as written. In fact, a direct proof of (3)⇒(2) is available without any biduality: for Γ an infinite metrizable subgroup of G_p^, the dual Γ^ is countable because Γ is precompact metrizable, and the evaluation map Φ: G → Γ^ has kernel Γ^⊥ (closed) and image isomorphic to G/Γ^⊥; the image is infinite because otherwise Γ would embed into the finite dual of the finite quotient G/Γ^⊥, forcing Γ to be finite. The authors should either state the relevant theorem from [25] and confirm it applies to all precompact abelian groups, or replace the argument with a direct one.","section":"Proposition 2.5, proof of (3)⇒(2), and Theorem 3.9"}],"minor_comments":[{"comment":"In the final paragraph of the proof, the sentence 'This implies that G does not contain non-trivial convergent sequences at all' should refer to the dual group G_p^, not to G. The proof has just shown that no sequence in G_p^ converges to the identity.","section":"Theorem 2.3, proof"},{"comment":"There is a formatting typo in the first line: 'Let φ : G − → H' should read 'Let φ : G → H'. This does not affect the mathematics.","section":"Proposition 2.5, proof of (2)⇒(1)"},{"comment":"Reference [25] is cited for a central duality result without a theorem number or page reference. Since this citation is used in two load-bearing proofs, a more precise pointer would help the reader verify the claim.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The central claims of the paper appear to be defensible, and the main theorem is likely correct. The two major issues are both fixable: correcting the abstract/introduction statement and either properly citing the biduality result from [25] or replacing that step with the direct argument. I do not see concerns about novelty or citation patterns; the paper builds on standard tools in the field."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the torsion characterization and the reflexive first-category example are good, honest results; the paper is publishable after a clean-up of one proof step and the abstract typo.\n\nThe genuinely new content: Theorem 2.7's equivalence between having a nontrivial convergent sequence in G_p^ and having a countably infinite Hausdorff quotient, for precompact torsion abelian groups. The proof is mostly clean: the diagonal-product argument for (4)=>(2) is elegant. Theorem 2.3 is a nice explicit construction of a dense, measure-zero, first-category subgroup of Z(2)^ω whose dual has no nontrivial convergent sequences; it complements Hart-Kunen and gives a countable dual so all compact subsets are finite. Theorem 3.9, which builds a precompact reflexive abelian group of first category, is the most interesting: it kills the conjecture that all such groups are Baire.\n\nWhere the paper is soft: (i) The abstract states the torsion property as 'No infinite quotient group of G is countable,' which is the negation of condition (1) of Theorem 2.7. That has to be a typo, but it will mislead. (ii) Proposition 2.5 (3)=>(2) rests on a citation to [25] for the assertion that the evaluation map G -> (G_p^)_p^ is a topological isomorphism. For general precompact abelian groups that is not true without a further condition; a proper dense subgroup of a compact group typically embeds densely into its second dual, not onto. The stress-test note is right to flag this. However, I think the gap is repairable: the argument only needs G to be dense in the second dual, because then the image of G in Γ_p^ is dense in Γ_p^, and being a dense subgroup of an infinite countable Hausdorff group makes it infinite and Hausdorff. So (2) still follows without the isomorphism. But the authors need to state the theorem they are using and either justify the stronger claim or weaken the proof. The same citation appears again in Theorem 3.9, so it needs checking.\n\nAll in all: the results look solid, the mathematics is serious, and the citation pattern is normal. The external-dependency issue is a real review point, not a fatal one. I'd send it to a referee for a general topology journal; after a revision that fixes the abstract and patches Prop 2.5, it should be accepted.","headline":"Torsion characterization and a counterexample to the Baire-property conjecture are solid and new; one proof step leans on an unstated duality theorem and the abstract has a typo, but both are fixable.","tokens_in":11202,"tokens_out":20326,"would_cite":true,"duration_ms":215515,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["43A40","22D35","22C05","54E52","54C10"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a precompact torsion abelian group, a nontrivial convergent sequence exists in the dual group exactly when the group has a countably infinite Hausdorff quotient.","keywords":["reflexive","precompact","pseudocompact","Baire property","convergent sequence","dual group","torsion abelian group","Pontryagin duality"],"falsifier":"A single counterexample would settle it: a precompact torsion abelian group $G$ such that $G_p^\\wedge$ has a nontrivial sequence converging to the identity while every infinite Hausdorff quotient of $G$ is uncountable. The paper's diagonal-product proof predicts such a sequence would produce a countably infinite quotient, so no such group exists.","tokens_in":10204,"feed_emoji":"🔁","tokens_out":13775,"duration_ms":116712,"temperature":0.7,"pith_summary":"The paper proves a duality-theoretic characterization for precompact abelian torsion groups $G$ (abelian groups that sit densely inside a compact group and whose elements all have finite order): the dual group $G_p^\\wedge$---the continuous characters of $G$ with the topology of pointwise convergence---contains a nontrivial convergent sequence exactly when $G$ has a countably infinite Hausdorff quotient. This reduces a subtle analytic property of the dual to a purely algebraic property of the original group. Along the way the authors show that this is equivalent to the dual containing an infinite metrizable subgroup and to the group admitting a countably infinite Hausdorff homomorphic image. They also construct a dense subgroup of $\\mathbb{Z}(2)^\\omega$ that is first category and has measure zero, yet whose dual has no nontrivial convergent sequences, and they use this to exhibit a precompact reflexive abelian group of the first Baire category.","feed_headline":"Dual convergent sequences are countable quotients in disguise","feed_subtitle":"A subtle analytic property of the dual becomes a plain algebraic condition on the group.","key_machinery":"The central object is the dual group $G_p^\\wedge$: all continuous homomorphisms from $G$ to the circle group $\\mathbb{T}$, with the topology of pointwise convergence on elements of $G$. Two mechanisms carry the proof. First, the canonical evaluation mapping $G\\to(G_p^\\wedge)_p^\\wedge$ is taken to be a topological isomorphism for precompact abelian $G$, so a closed metrizable subgroup $\\Gamma$ of $G_p^\\wedge$ can be converted into the quotient $G/\\Gamma^\\perp$. Second, for a convergent sequence of characters, the diagonal product $f=\\triangle_n\\chi_n:G\\to\\mathbb{T}^\\omega$ has image in the direct sum $D^{(\\omega)}$ when $G$ is torsion, which makes the image countable and yields the quotient witnessing the sequence. The paper also uses the group $c_0(\\mathbb{T})$ of sequences in $\\mathbb{T}$ converging to $1$: a dual group contains a nontrivial convergent sequence exactly when $G$ maps onto a subgroup of $c_0(\\mathbb{T})$ that separates the points of the underlying convergent sequence.","core_discovery":"The main theorem (Theorem 2.7) states that for a precompact torsion abelian group $G$, the following are equivalent: (1) $G$ has a countably infinite Hausdorff quotient; (2) $G$ has a countably infinite Hausdorff homomorphic image; (3) $G_p^\\wedge$ contains an infinite metrizable subgroup; and (4) $G_p^\\wedge$ contains a nontrivial convergent sequence. The proof of (4) implies (2) is the heart: a nontrivial sequence of characters converging to the identity gives a diagonal homomorphism $f:G\\to\\mathbb{T}^\\omega$, and because $G$ is torsion each value $f(x)$ has finite order, so $f(x)$ lies in the direct-sum subgroup $D^{(\\omega)}$ of $\\mathbb{T}^\\omega$, where $D$ is the torsion subgroup of the circle group; hence $f(G)$ is a countably infinite Hausdorff homomorphic image of $G$. The converse direction is also explicit: the dual of a countably infinite precompact group is an infinite metrizable group, which embeds in $G_p^\\wedge$ and therefore supplies a convergent sequence.","pith_inferences":["A natural extension question, not answered in the paper, is whether absence of nontrivial convergent sequences in the dual of a precompact torsion group forces every compact subset of the dual to be finite; the paper proves this for its example and cites it for bounded torsion groups, but not for all torsion groups.","The torsion hypothesis enters only to make the diagonal image $f(G)$ countable; outside the torsion setting the same argument can produce uncountable images, which suggests the equivalence between convergent sequences and countable quotients is genuinely special to torsion groups.","Because the proof of (3) implies (2) depends on the cited double-duality theorem, a reader who wants to rely on Theorem 2.7 should verify that theorem's exact hypotheses; if it only gives a dense embedding for some precompact groups, the annihilator step would need repair before the equivalence is established."],"forward_implications":["For precompact torsion groups, the search for convergent sequences in the dual is a search for countable quotients: if every infinite quotient of $G$ is uncountable, then $G_p^\\wedge$ has no nontrivial convergent sequence.","Every nontrivial convergent sequence in $G_p^\\wedge$ factors through a countably infinite Hausdorff quotient of $G$; the diagonal homomorphism built from the sequence has that quotient as its image.","Countably infinite precompact torsion groups are the extreme case: their duals are infinite metrizable groups and hence contain convergent sequences.","There exist precompact abelian groups whose duals contain infinite compact subsets but no nontrivial convergent sequences, so the torsion characterization does not extend verbatim to all precompact groups.","The first-category, measure-zero subgroup of $\\mathbb{Z}(2)^\\omega$ whose dual has only finite bounded subsets yields a precompact reflexive abelian group of the first Baire category, showing that reflexivity for precompact groups does not force the Baire property."],"supporting_citations":[{"why":"Supplies the double-duality theorem used to identify $G$ with $(G_p^\\wedge)_p^\\wedge$ in Proposition 2.5 and Theorem 3.9.","marker":"[25]"},{"why":"The theorem in that paper is the result complemented by the example in Theorem 2.3 of a first-category, measure-zero subgroup whose dual has no nontrivial convergent sequences.","marker":"[18]"},{"why":"Establishes that the topology of a precompact group is the topology of pointwise convergence on its dual, which underlies the whole duality setup.","marker":"[6]"},{"why":"Shows that characters on products depend on finitely many coordinates, which is used to build the point $x^*$ in the example inside $\\mathbb{Z}(2)^\\omega$.","marker":"[21]"},{"why":"Gives a continuity result implying that duals of Baire groups have no nontrivial convergent sequences, motivating the restriction to first-category groups.","marker":"[13]"},{"why":"Provides the construction of a pseudocompact group with h-embedded countable subgroups used in Theorem 3.9 to produce the reflexive first-category group.","marker":"[4]"}],"fun_headline_variants":["Torsion groups: dual limits mean countable quotients","Convergent dual sequences are countable quotients","Dual convergence? Countable quotients!","Countable quotients: the key to dual convergence","In torsion groups, dual convergence is a countable quotient"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof leans on the cited theorem that every precompact abelian group is topologically isomorphic to its second dual; if that theorem only gives a dense embedding in general, the step converting a closed metrizable subgroup of the dual into a quotient of $G$ breaks.","fun_headline_variants_meta":{"raw":{"variants":["Torsion groups: dual limits mean countable quotients","Convergent dual sequences are countable quotients","Dual convergence? Countable quotients!","Countable quotients: the key to dual convergence","In torsion groups, dual convergence is a countable quotient"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001069,"raw_usage":{"total_tokens":4506,"prompt_tokens":998,"completion_tokens":3508,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":614,"completion_tokens_details":{"reasoning_tokens":3433}},"tokens_in":614,"tokens_out":3508,"duration_ms":27532,"temperature":1.0,"reasoning_tokens":3433,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:19:46.679688+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single counterexample would settle it: a precompact torsion abelian group $G$ such that $G_p^\\wedge$ has a nontrivial sequence converging to the identity while every infinite Hausdorff quotient of $G$ is uncountable. The paper's diagonal-product proof predicts such a sequence would produce a countably infinite quotient, so no such group exists.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the double-duality theorem used to identify $G$ with $(G_p^\\wedge)_p^\\wedge$ in Proposition 2.5 and Theorem 3.9."},{"cited_title":"Hart and K","cited_arxiv_id":null,"evidence_quote":"The theorem in that paper is the result complemented by the example in Theorem 2.3 of a first-category, measure-zero subgroup whose dual has no nontrivial convergent sequences."},{"cited_title":"Comfort, K.A","cited_arxiv_id":null,"evidence_quote":"Establishes that the topology of a precompact group is the topology of pointwise convergence on its dual, which underlies the whole duality setup."},{"cited_title":"Kaplan, Extension of the Pontrjagin duality I: Infini te products, Duke Math","cited_arxiv_id":null,"evidence_quote":"Shows that characters on products depend on finitely many coordinates, which is used to build the point $x^*$ in the example inside $\\mathbb{Z}(2)^\\omega$."},{"cited_title":"Fleischer, T","cited_arxiv_id":null,"evidence_quote":"Gives a continuity result implying that duals of Baire groups have no nontrivial convergent sequences, motivating the restriction to first-category groups."},{"cited_title":"Bruguera, M","cited_arxiv_id":null,"evidence_quote":"Provides the construction of a pseudocompact group with h-embedded countable subgroups used in Theorem 3.9 to produce the reflexive first-category group."}],"review_version":1}