{"id":"cde9a156-6391-41f8-b293-3609bac3c998","arxiv_id":"2506.01161","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Lomonosov-type theorem holds for compact operators on non-finitely-generated Hilbert C*-modules over finite-dimensional C*-algebras: each such operator has a proper nonzero hyperinvariant submodule.","lead":"This paper introduces invariant submodules for Hilbert C*-modules and proves a Lomonosov-type theorem: every nonzero compact operator on a non-finitely-generated Hilbert module over a finite-dimensional C*-algebra has a proper nonzero hyperinvariant submodule. It also characterizes when operators have nontrivial invariant complemented submodules via a quadratic operator equation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central theorem is likely correct, but its proof hinges on Lemma 4.1, whose proof sketch in the paper fails to produce a single subsequence valid for all v; the lemma is true via reflexivity, so this is a proof gap rather than a false claim.","rationale":"After checking the central argument, I find no fatal flaw in Theorem 4.8. The eigenspace case is standard, and the quasinilpotent case follows the Lomonosov covering argument, with the inequality ∥Kf_n - Kf_m∥ > 1/2 valid because Lf_n - Lf_m + f_m ∈ F_{n+1}. Lemma 4.6(1) follows from the ideal property of K(E). The main weakness is Lemma 4.1, as the reader noted; I agree it is the load-bearing compactness premise. However, I believe the lemma is true: Hilbert modules over finite-dimensional C*-algebras are reflexive, so bounded sequences have weakly convergent subsequences, which gives (4.1), and density of finite-rank operators gives (4.2). The paper's proof, though, does not establish this, since it does not produce a single subsequence. Thus the correct verdict is CONDITIONAL: accept if the lemma is verified from [1,5] or the proof is rewritten. The other issues (Lemma 2.2(2), missing justification for strict decrease in Lemma 4.6, proximate nearest points in Lemma 4.3) are minor and repairable. No ad hominem; the critique is on the argument.","tokens_in":13115,"tokens_out":32138,"duration_ms":307050,"concrete_test":"Retrieve [1] and [5] and verify that they prove Lemma 4.1 exactly in the single-subsequence form used here. If not, write a complete proof of Lemma 4.1 using the Eberlein–Šmulian theorem for the reflexive Banach space E and check that the resulting subsequence satisfies (4.1) for every v∈E and (4.2) for every T∈K(E). If neither the references nor such a proof can confirm the lemma, the proof of Theorem 4.8 is incomplete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Lemma 4.1 the paper claims that a bounded sequence (ζ_n) in a Hilbert module over a finite-dimensional C*-algebra has a subsequence satisfying (4.1) for every v∈E. The proof chooses, for each v, a convergent subsequence of the A-valued sequence ⟨v, ζ_n⟩, but these subsequences need not be the same, so the limit map g(v) is not well-defined and need not be A-linear. A single subsequence valid for all v is not obtained in the text. This is not merely cosmetic: Lemma 4.1 is used in Lemma 4.3 to bound the sequence ζ_n and to extract a convergent subsequence of Kζ_n, and in Lemma 4.6 to extract a convergent subsequence of {Kf_n}; without (4.1)–(4.2) for a common subsequence, the Lomonosov covering argument in Theorem 4.8 loses its key compactness input. The lemma is attributed to [1, Theorem 2.3] and [5, Proposition 2.1]; if those references supply the missing argument, the central theorem still stands. But as written, the proof of the paper is not self-contained and relies on an imported result whose proof sketch is invalid. A correct proof can likely be supplied by noting that E is a reflexive Banach space (being a Hilbert module over a finite-dimensional C*-algebra), so by the Eberlein–Šmulian theorem every bounded sequence has a weakly convergent subsequence; this yields (4.1) for all v and then (4.2) by density of finite-rank operators.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper develops a theory of invariant and reducing submodules for adjointable operators on Hilbert C*-modules. The early sections define invariant submodules, characterize them via projections satisfying PTP=TP and via nontrivial solutions of STS=TS, derive block-matrix decompositions relative to complemented submodules, and study the Moore-Penrose inverse under a numerical-range condition. The later sections specialize to Hilbert modules over C*-algebras of compact operators and over finite-dimensional C*-algebras. The main result, Theorem 4.8, asserts that if A is finite-dimensional and E is a non-finitely generated Hilbert A-module, then every nonzero compact operator on E has a proper nonzero hyperinvariant submodule; Corollary 4.9 draws the usual Lomonosov consequence for operators commuting with compact operators. The proof follows Lomonosov's spectral and covering argument, using a compactness lemma (Lemma 4.1) imported from [1,5] and a spectral lemma (Lemma 4.6).","tokens_in":13396,"tokens_out":12075,"duration_ms":115753,"significance":"Should Theorem 4.8 hold, it is a substantive extension of Lomonosov's theorem to a class of Hilbert C*-modules, and the framework of invariant submodules introduced here may be useful for further module operator theory. The paper gives explicit formulas for solution sets of operator equations, which are concrete and potentially applicable. The main proof is a recognizable Lomonosov argument and, modulo the gaps identified below, is likely correct. However, the two load-bearing lemmas in Section 4 are not fully proved as written: Lemma 4.1 has an invalid proof sketch, and Lemma 4.6 contains an asserted strict descent that is not justified. These are repairable, but until then the central theorem is not established by the manuscript. The paper does not ship machine-checked proofs or code; its value rests on the mathematical argument and on the correctness of the imported compactness results.","major_comments":[{"comment":"The proof as written is not a valid proof of the statement. For each v in E, the argument produces a subsequence (zeta_{n_k(v)}) for which <v, zeta_{n_k(v)}> converges, but these subsequences need not be compatible, so the limit map g(v) is not defined on a common subsequence and need not be A-linear. Since (4.1) and (4.2) are subsequently used with a single subsequence in Lemma 4.3, in Lemma 4.6, and in the covering argument of Theorem 4.8, this is a load-bearing gap. The lemma may be correct (for example, via reflexivity of E and the Eberlein-Smulian theorem), and [1, Theorem 2.3] and [5, Proposition 2.1] may supply it, but the paper must either prove it fully or import it with a correct self-contained argument.","section":"Section 4, Lemma 4.1"},{"comment":"The assertion that the sequence F_1 contains F_2 contains ... is strictly decreasing is not proved. From injectivity of L and Lemma 4.3 one knows that each F_n is closed, but strictness requires an additional cancellation argument: if Ran(L^n)=Ran(L^{n-1}), then for every y in E there is z with L^{n-1}y = L^n z = L^{n-1}(Lz), and injectivity of L^{n-1} gives y=Lz, contradicting properness of Ran(L). The proof should also state that L^n = I - K_n with K_n compact, so that Lemma 4.3 applies to each F_n. Without strictness, the selection of f_n with ||f_n - F_{n+1}|| > 1/2 has no basis, and the displayed contradiction in the proof of part (2) collapses.","section":"Section 4, Lemma 4.6"},{"comment":"The proof asserts the existence of u_n in Ker(L) achieving the distance from x_n to Ker(L). This existence is true because Hilbert modules over finite-dimensional C*-algebras are reflexive Banach spaces, but the text does not justify it. Since this proximinality is used to form the sequence zeta_n, a short proof or explicit citation should be supplied. This is a smaller gap than the two above, but it is still a step that a reader cannot verify from the text as written.","section":"Section 4, Lemma 4.3"}],"minor_comments":[{"comment":"The word 'underling' should be 'underlying'.","section":"Abstract"},{"comment":"The phrase 'In the early thirties' should read 'In the early 1930s'; the historical comments on the invariant subspace problem would also benefit from a clear statement that the Enflo and Neville preprints have not yet been peer-reviewed.","section":"Section 1, Introduction"},{"comment":"In the second part of the proof, 'suppose T S=0' is a typographical slip for 'suppose TS=0'; the surrounding argument does not otherwise define the first case clearly.","section":"Section 2, proof of Lemma 2.3"},{"comment":"The notation 'KB=KB is a compact subset' is confusing; the closure of K(B) should be denoted overline{K(B)} or another explicit symbol to distinguish the set from its closure.","section":"Section 4, proof of Theorem 4.8"},{"comment":"The statement contains a formatting error: 'T=I. a' should be typeset as 'T = I \\cdot a'.","section":"Problem 2.16"}],"recommendation":"major_revision","confidential_remarks":"The paper sits within the scope of math.OA and the main result, if repaired, is likely publishable in a specialist operator theory or Hilbert module theory journal. The key issue is that the proof of Theorem 4.8 relies on Lemma 4.1 and Lemma 4.6, both of which contain genuine gaps as written; the lemma attributed to [1,5] is essential and the proof sketch in the paper is invalid. I would advise the editor to require a corrected proof of Lemma 4.1 and a justification of the strict descent in Lemma 4.6 before final acceptance. The self-citations to the author's previous work are contextually appropriate, but the dependence on [1,5] for the central compactness input should be made explicit and checked."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read on Sharifi's paper. The main claim, Theorem 4.8, is a Lomonosov-type theorem for compact operators on Hilbert C*-modules over finite-dimensional C*-algebras, when the module is not finitely generated. That is a real extension of Lomonosov's 1973 result to a setting where it hadn't been done, and the proof strategy is a faithful adaptation of Lomonosov's covering argument. The reader checked the literature: the result is not in the cited references, there's no circularity, and no fitted parameters. If the theorem is correct—and I think it is—it's a solid contribution to Hilbert C*-module theory.\n\nWhat the paper does well: it gives a clean statement and proof of the extension, and the surrounding sections on invariant submodules, operator equations, and generalized inverses are mostly standard but competently handled. The proof of Theorem 4.8 itself, once you accept Lemma 4.1 and Lemma 4.6, is a direct and readable version of the classical argument.\n\nNow the soft spots, in proportion. Lemma 4.1, which is imported from [1,5] and is the compactness engine, has a proof sketch in the paper that doesn't work as written. The text just picks, for each v, a subsequence of ⟨v,·⟩; there's no diagonalization to get a single subsequence valid for all v, and the limit map g(v) isn't shown to be well-defined or A-linear. The stress-test note is right about this: the lemma is true (E is reflexive, so Eberlein–Smulian would do it, or the cited references likely have the full proof), but the proof gap is real. Lemma 4.6 also asserts without argument that the ranges of L^n form a strictly decreasing sequence when L is injective but not surjective; that needs a proof, and the contradiction argument depends on it. Minor issues: Lemma 2.2(2) is unused and its hypothesis is vague, and Lemma 4.3 silently assumes proximinality of closed submodules for the distance-minimizing sequence. None of this undermines the truth of Theorem 4.8; it's all repairable.\n\nWho this is for: researchers in Hilbert C*-module theory and operator theory who care about invariant subspaces in module settings. A referee should be sent in, and should require a correct proof of Lemma 4.1 and a justification for the strict decrease in Lemma 4.6. I would accept this for review, not desk-reject.","headline":"Natural extension of Lomonosov to Hilbert C*-modules over finite-dimensional C*-algebras; the main theorem is likely correct but the proof has repairable gaps in two key lemmas.","tokens_in":13967,"tokens_out":3116,"would_cite":true,"duration_ms":28428,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L08","47A05","46C50","46L05"],"pacs":[],"model":"deepseek-v4-flash","headline":"On Hilbert C*-modules over finite-dimensional C*-algebras, every nonzero compact operator on a non-finitely-generated module has a proper nonzero hyperinvariant submodule.","keywords":["Hilbert C*-module","invariant submodule","hyperinvariant submodule","compact operator","Lomonosov theorem","finite-dimensional C*-algebra","generalized inverse","operator equation"],"falsifier":"Concrete check: try to construct a non-finitely-generated Hilbert module over a finite-dimensional $C^*$-algebra and a nonzero compact operator on it with either a nonzero spectral point that is not an eigenvalue, or a vector whose orbit under the commutant is dense while the spectral-radius estimate (4.3) still holds; either observation would falsify Theorem 4.8 or one of the lemmas that feed it.","tokens_in":12870,"feed_emoji":"🧩","tokens_out":14049,"duration_ms":102533,"temperature":0.7,"pith_summary":"Invariant submodules are the Hilbert $C^*$-module analogue of invariant subspaces: closed submodules preserved by an adjointable operator. The paper develops their basic theory, connects them to projections, reducing submodules, and the operator equation $XTX = TX$, and then proves a Lomonosov-type theorem. The theorem states that if $A$ is a finite-dimensional $C^*$-algebra and $E$ is a Hilbert $A$-module that is not finitely generated, then every nonzero compact operator on $E$ has a proper nonzero hyperinvariant submodule. This extends a classical result from Hilbert and Banach spaces to a module setting where self-duality and orthogonal complementation fail in general, so the proof has to build the needed compactness by hand.","feed_headline":"Compact operators gain hyperinvariant submodules","feed_subtitle":"Lomonosov's invariant-subspace theorem extends to Hilbert C*-modules over finite-dimensional C*-algebras.","key_machinery":"The mechanism that carries the argument is a weak compactness property special to Hilbert modules over finite-dimensional $C^*$-algebras (Lemma 4.1, taken from [1, Theorem 2.3] and [5, Proposition 2.1]). It says that every bounded sequence has a subsequence whose images under every compact operator converge in norm, and it follows from the self-duality of such modules proved by Frank [10, Proposition 4.4]. That property makes the range of $I - K$ an orthogonal summand (Lemma 4.3), turns nonzero spectral points into eigenvalues (Lemma 4.6), and finally makes the closed image $KB$ of the unit ball compact enough for the Lomonosov covering argument. The covering argument itself, with the constant $c = \\max \\|S_i\\|$ and the vanishing of $\\|(cK)^m\\|$, is the same combinatorial skeleton as in the classical proof.","core_discovery":"The central claim is Theorem 4.8: for a finite-dimensional $C^*$-algebra $A$ and a Hilbert $A$-module $E$ that is not finitely generated, every nonzero compact operator $K$ on $E$ admits a proper nonzero hyperinvariant submodule---a closed submodule invariant under every $S \\in L(E)$ that commutes with $K$. If $K$ has a nonzero spectral value $\\lambda$, then $\\mathrm{Ker}(\\lambda I - K)$ is such a submodule by Lemma 4.6, and it is proper because $K$ is compact and $E$ is not finitely generated. If the spectrum of $K$ is just $\\{0\\}$, the paper runs the Lomonosov covering argument: the sets $O(S) = \\{u : \\|Su - x_0\\| < 1\\}$, for $S$ commuting with $K$, cover the compact set $KB$, and the spectral-radius condition (4.3) forces a contradiction with $0$ lying outside the closed unit ball $B$ centered at $x_0$. The upshot is that the closed submodule generated by the orbit of some vector under the commutant of $K$ is neither zero nor all of $E$.","pith_inferences":["Inference: the same hyperinvariant-submodule conclusion plausibly holds for any adjointable operator whose commutant contains a nonzero compact operator on these modules, mirroring Lomonosov's stronger formulation in Banach spaces.","Inference: because the proof keys off self-duality, a natural test case is Hilbert modules over finite-dimensional $W^*$-algebras, where a version of the weak compactness property might persist under extra hypotheses.","Inference: a constructive version of the theorem would specify the vector whose commutant orbit yields the hyperinvariant submodule in the $\\sigma(K) = \\{0\\}$ case; working it out for $E = \\ell^2 \\otimes M_n(\\mathbb{C})$ would give an explicit, checkable example."],"forward_implications":["Corollary 4.9: any adjointable operator $S$ commuting with a nonzero compact operator on such a module has a proper nonzero invariant submodule.","Over $C^*$-algebras of compact operators, the equivalence of having a nontrivial invariant submodule, having a nontrivial projection $P$ with $PTP = TP$, and having a nontrivial solution of $STS = TS$ becomes unconditional because every closed submodule is complemented (Corollary 3.1).","For a complemented invariant submodule $W$ with closed range of $TP_W$, the solutions of $XTX = TX$ are exactly $P_{\\mathrm{Ran}(TP_W)} + P_W Z(I - P_{\\mathrm{Ran}(TP_W)})$ with arbitrary $Z \\in L(E)$ (Corollaries 2.6 and 3.2).","A complemented submodule that reduces a Moore-Penrose invertible operator $T$ also reduces its Moore-Penrose inverse $T^\\dagger$ whenever $0$ is outside the $C^*$-numerical range of $T$ (Theorem 2.12)."],"supporting_citations":[{"why":"Supplies the weak-compactness statement (Theorem 2.3) imported into Lemma 4.1.","marker":"[1]"},{"why":"Supplies the same weak-compactness property (Proposition 2.1) used in Lemma 4.1.","marker":"[5]"},{"why":"Proves self-duality of Hilbert $C^*$-modules over finite-dimensional $C^*$-algebras (Proposition 4.4), which powers Lemma 4.1.","marker":"[10]"},{"why":"Provides the closed-range/orthogonal-summand equivalence used in Lemma 4.3 and throughout.","marker":"[20]"},{"why":"The original Lomonosov theorem whose covering argument Theorem 4.8 adapts.","marker":"[22]"},{"why":"Gives the finitely-generated/unital-$K(E)$ equivalence used in Lemma 4.6(1).","marker":"[33]"},{"why":"Shows every closed submodule of a Hilbert module over a $C^*$-algebra of compact operators is complemented, used in Section 3.","marker":"[4]"},{"why":"Characterizes $C^*$-algebras of compact operators by complementedness of closed submodules, used in Section 3.","marker":"[29]"}],"fun_headline_variants":["Lomonosov-type theorem for compact operators on Hilbert C*-modules","Compact operators on C*-modules have hyperinvariant submodules","Non-finitely generated C*-modules force Lomonosov-type submodules","Finite-dimensional C*-algebras: compact operators get hyperinvariant submodules"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is Lemma 4.1: for Hilbert modules over finite-dimensional $C^*$-algebras, every bounded sequence has a subsequence whose images under every compact operator converge in norm; if that weak compactness ever failed, the range-closedness, spectral-eigenvalue, and covering arguments in Theorem 4.8 would all break down.","fun_headline_variants_meta":{"raw":{"variants":["Lomonosov-type theorem for compact operators on Hilbert C*-modules","Compact operators on C*-modules have hyperinvariant submodules","Non-finitely generated C*-modules force Lomonosov-type submodules","Finite-dimensional C*-algebras: compact operators get hyperinvariant submodules"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000822,"raw_usage":{"total_tokens":3577,"prompt_tokens":904,"completion_tokens":2673,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":520,"completion_tokens_details":{"reasoning_tokens":2605}},"tokens_in":520,"tokens_out":2673,"duration_ms":19667,"temperature":1.0,"reasoning_tokens":2605,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:52:37.395369+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Concrete check: try to construct a non-finitely-generated Hilbert module over a finite-dimensional $C^*$-algebra and a nonzero compact operator on it with either a nonzero spectral point that is not an eigenvalue, or a vector whose orbit under the commutant is dense while the spectral-radius estimate (4.3) still holds; either observation would falsify Theorem 4.8 or one of the lemmas that feed it.","supporting_citations":[{"cited_title":"Arambaˇ si´ c, D","cited_arxiv_id":null,"evidence_quote":"Supplies the weak-compactness statement (Theorem 2.3) imported into Lemma 4.1."},{"cited_title":"Chmieli´ nski, D","cited_arxiv_id":null,"evidence_quote":"Supplies the same weak-compactness property (Proposition 2.1) used in Lemma 4.1."},{"cited_title":"Frank, Self-duality and C*-reflexivity of Hilbert C*-modules,Z","cited_arxiv_id":null,"evidence_quote":"Proves self-duality of Hilbert $C^*$-modules over finite-dimensional $C^*$-algebras (Proposition 4.4), which powers Lemma 4.1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the closed-range/orthogonal-summand equivalence used in Lemma 4.3 and throughout."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The original Lomonosov theorem whose covering argument Theorem 4.8 adapts."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the finitely-generated/unital-$K(E)$ equivalence used in Lemma 4.6(1)."},{"cited_title":"Baki´ c and B","cited_arxiv_id":null,"evidence_quote":"Shows every closed submodule of a Hilbert module over a $C^*$-algebra of compact operators is complemented, used in Section 3."},{"cited_title":"Schweizer, A description of Hilbert C*-modules in which all closed submodules are orthogonally closed,Proc","cited_arxiv_id":null,"evidence_quote":"Characterizes $C^*$-algebras of compact operators by complementedness of closed submodules, used in Section 3."}],"review_version":1}