{"id":"49154981-a8ec-4d51-ab74-207ed9251701","arxiv_id":"1908.09241","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A C*-algebra that admits a uniform approximate ideal structure over pairs satisfying the Künneth formula satisfies the Künneth formula itself.","lead":"This paper defines a flexible notion of decomposing a C*-algebra into two simpler subalgebras, called an approximate ideal structure, and proves that K-theoretic properties pass from the pieces to the whole. The main payoff is a new permanence theorem for the Künneth formula, a standard computational tool in noncommutative geometry.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.7 contains an unjustified reordering step: a D-near element conjugating C-near factors is asserted to be C-near, and the injectivity proof depends on it.","rationale":"The reader correctly identifies Proposition 5.7 as the delicate step and assigns a conditional verdict. My concern is more specific: even assuming f-uniformity, the proof of Proposition 5.7 contains a reordering of approximately-C and approximately-D factors whose validity is not demonstrated. The assertion that v^C is close to C is not a consequence of f-uniformity alone, because f-uniformity is a statement about common approximate intersections, not about commutators of C and D elements. The possible rescue is that Lemma 5.6's construction gives v^C ≈ 1 + hZ and v^D ≈ 1 + (1-h)Z, so the commutator [v^{D,a}, v^{C,a}] may indeed lie close to C∩D by the defining condition (iii) of an approximate ideal structure. But this computation is absent from the paper. Because Theorem 9.1 relies on Proposition 5.7, and Theorem 1.4 relies on Theorem 9.1, this is the most load-bearing gap. I do not recommend rejection, since the gap is localized and may be fillable; a detailed verification of Proposition 5.7 is the appropriate gate for acceptance.","tokens_in":41533,"tokens_out":26455,"duration_ms":253717,"concrete_test":"Compute the leading-order commutator in the proof of Proposition 5.7 using the explicit forms from Lemma 5.6: write v^{D,a} ≈ 1 + (1-h)Z_D and v^{C,a} ≈ 1 + hZ_C, and determine whether [v^{D,a}, v^{C,a}] is δ-in C via condition (iii) of Definition 3.1, which controls h(1-h) times monomials into C∩D. If the computation cannot be completed, Proposition 5.7 is unproved; if it succeeds, the concern is resolved. Cross-check the analogous reordering step against Oyono-Oyono-Yu [25, Lemma 2.9].","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The injectivity half of Theorem 1.4 rests on Proposition 5.7, whose proof contains a step that is not justified by the stated assumptions. After Lemma 5.6 yields v^{C,a}, v^{D,a}, v^{C,b}, v^{D,b}, the proof defines v^C := v^{D,a} v^{C,a} v^{C,b} (v^{D,a})^{-1} and v^D := v^{D,a} v^{D,b}, then asserts that v^C is δ2-in M_{2n}(rC). Lemma 5.6 only guarantees v^{D,a} is close to 1+D, while v^{C,a}, v^{C,b} are close to 1+C. For non-ideal subalgebras, conjugation of a C-near element by a D-near element need not be C-near: to first order one obtains 1+c+c' + [d,c+c'] + ..., and f-uniformity (Definition 5.1) does not control commutators [D,C]. This matters because the proof then uses v^C as a representative of the trivial class in K1(C); if v^C is not actually close to C, the conclusion [x] = [u_C] in K1(C) is unsupported. The step may be salvageable from the special structure v^C ≈ 1 + hZ, v^D ≈ 1 + (1-h)Z, using condition (iii) of Definition 3.1 to place h(1-h) terms in C∩D, but the paper does not supply this computation. Since Proposition 5.7 is used exactly in Theorem 9.1, this is the most load-bearing technical point in the central claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a notion of approximate ideal structure for C*-algebras: a quantitative partition of unity (h, 1-h) that approximately commutes with finite sets and sends h-times and (1-h)-times into prescribed subalgebras C and D, with an additional intersection condition. It develops boundary classes in K-theory and proves two main applications. The first is a vanishing theorem (Theorem 1.2): if A has an approximate ideal structure over pairs with trivial K-theory, then A has trivial K-theory. The second is a permanence theorem for the Künneth formula (Theorem 1.4): if A has a uniform approximate ideal structure over pairs (C,D) such that C, D, and C∩D satisfy the Künneth formula, then A satisfies the Künneth formula. The proof is elementary, avoiding controlled K-theory, and is modeled on the author's prior work with Guentner and Yu and on Oyono-Oyono-Yu. Appendices give a weak approximate ideal structure for nuclear dimension one and uniform approximate ideal structures from finite dynamical complexity of ample groupoids, with consequences for the Baum-Connes conjecture and the Künneth formula.","tokens_in":41834,"tokens_out":5770,"duration_ms":60455,"significance":"If the main theorem is correct, it provides a genuinely new permanence principle for the Künneth formula, reducing the problem to checking simpler subalgebras. The paper is also valuable as a self-contained exposition of an approximate Mayer-Vietoris method that avoids controlled K-theory. The strengths of the manuscript are its explicit quantitative lemmas, its honest discussion of the weak quantifiers in Proposition 5.7, and its non-vacuous examples from nuclear dimension and groupoid decompositions. The main theorem is not circular: it assumes Künneth for C, D, and C∩D and derives it for A, using newly constructed boundary classes. However, the proof contains one load-bearing gap in Proposition 5.7 that affects the injectivity half of Theorem 1.4, and the scope of Theorem 1.4 is narrower than the word 'approximate' in the title and abstract suggests because of the f-uniformity hypothesis.","major_comments":[{"comment":"The display after 'Note that v^C and v^D are δ2-in M_{2n}(rC) and M_{2n}(rD) respectively' asserts that v^C := v^{D,a} v^{C,a} v^{C,b} (v^{D,a})^{-1} is δ2-in M_{2n}(rC). Lemma 5.6 only guarantees that v^{D,a} is close to 1 + D while v^{C,a} and v^{C,b} are close to 1 + C. For non-ideal subalgebras, conjugation of a C-near element by a D-near element need not be C-near: to first order the error contains a commutator [d, c + c'], and f-uniformity (Definition 5.1) controls distances between elements of C⊗B and D⊗B, not commutators [D,C]. This step is load-bearing because the proof then uses v^C as a representative of the trivial class in K1(C) to conclude [x] = [u_C], and Proposition 5.7 is used exactly in the proof of Theorem 9.1. The step may be salvageable from the special structure v^C ≈ 1 + hZ, v^D ≈ 1 + (1-h)Z: condition (iii) of Definition 3.1 places h(1-h) terms in C∩D, which would control the conjugation error. But the paper does not supply this computation, so the proof as written is incomplete.","section":"§5, Proposition 5.7"}],"minor_comments":[{"comment":"There is a leftover editorial marker 'HERE' immediately after Lemma 6.4; it should be removed.","section":"§6, after Lemma 6.4"},{"comment":"The word 'quatifiers' should be 'quantifiers', and the statement that the quantifiers are 'in the wrong order' is accurate but should be followed by a precise explanation of why the order still suffices in the applications.","section":"§5, introductory paragraph"},{"comment":"The reference to 'Lemma 9' in the proof should be to Lemma 5.6; this cross-reference error will confuse readers.","section":"§5, proof of Proposition 5.7"},{"comment":"The abstract says 'if A can be decomposed into a pair (C,D) such that ...', omitting the word 'uniform' from the hypothesis of Theorem 1.4. Since the uniform version is substantially stronger and excludes natural examples (see Example 5.3), the abstract and introduction should state the hypothesis precisely.","section":"§1, Theorem 1.4 and abstract"},{"comment":"The phrase 'commutates' should be 'commutes', and the diagram commutativity argument would benefit from naming the maps explicitly rather than relying only on 'naturality'.","section":"§8, proof of Theorem 8.1"}],"recommendation":"major_revision","confidential_remarks":"The main gap in Proposition 5.7 is real but appears fixable by a direct computation using the partition-of-unity structure and condition (iii) of Definition 3.1. Because the gap is load-bearing for the central permanence theorem, I recommend major revision rather than rejection. The paper is otherwise careful and the results, if repaired, would be a solid contribution suitable for a general operator-algebra journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read the paper. The new notion of approximate ideal structure is the real thing: it gives a uniform reason for several K-theoretic permanence results, and Theorem 1.4 (Künneth property is permanent under uniform approximate ideal structure) is new as stated. The paper is also honest about its debt to Oyono-Oyono–Yu and avoids controlled K-theory, which is a genuine simplification. The vanishing theorem and the groupoid examples are useful.\n\nThe soft spot is Proposition 5.7, the exactness property at position (III) that carries the injectivity half of Theorem 1.4. In the proof, after Lemma 5.6 produces v^{C,a}, v^{D,a}, v^{C,b}, v^{D,b}, the author defines v^C = v^{D,a} v^{C,a} v^{C,b} (v^{D,a})^{-1} and asserts it is C-near. Lemma 5.6 only says v^{D,a} is close to 1+D and v^{C,a}, v^{C,b} to 1+C. For non-ideal C and D, conjugating a C-near element by a D-near element is not automatically C-near; the first-order cross terms involve [D,C], which f-uniformity does not control. The special form of the elements (roughly 1+hZ and 1+(1-h)Z) suggests the cross terms land in C∩D because of condition (iii), but the paper does not supply that computation, and it is not immediate that the relevant polynomials belong to the finite-dimensional subspace Y built earlier. Since Proposition 5.7 is used exactly in Theorem 9.1, this is load-bearing.\n\nI would not call this fatal. The framework is coherent and the gap looks fixable by enlarging Y and doing the commutator estimate carefully. But an editor should send this to a careful referee rather than accept on faith.\n\nMinor blemishes: a stray 'HERE' after Lemma 6.4, a 'Lemma 9' reference that should be 'Lemma 5.6', and a missing citation for the earlier Oyono-Oyono result mentioned in footnote 5. The citation pattern is otherwise fine: the paper credits [25] heavily, and self-citations to [16] are for examples, not assumptions of the main theorem.\n\nWho it is for: anyone working on K-theoretic permanence, Baum-Connes, or controlled K-theory. It deserves a serious referee; I'd recommend conditional acceptance after the Proposition 5.7 gap is closed.","headline":"A genuinely new framework with an important Künneth permanence theorem, but the proof of the key exactness proposition has an unjustified conjugation step that needs fixing.","tokens_in":42415,"tokens_out":7519,"would_cite":true,"duration_ms":69551,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L80","19K35","46L05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The Künneth formula passes from the pieces of a uniform approximate ideal decomposition to the whole C*-algebra.","keywords":["K-theory","C*-algebras","Künneth formula","approximate ideal structure","Mayer-Vietoris sequence","Baum-Connes conjecture","nuclear dimension","groupoid C*-algebras"],"falsifier":"To refute Theorem 1.4 one would need a C*-algebra A with a uniform approximate ideal structure over pairs (C,D) for which C, D, and C∩D satisfy the Künneth formula, but for which the product map $K_*(A)\\otimes K_*(B) \\to K_*(A\\otimes B)$ fails to be an isomorphism for some B with free abelian K-theory; the theorem asserts no such example exists. A concrete place to look is any pair violating f-uniformity, such as the hereditary subalgebras of the compact operators described in Example 5.3, checking whether a decomposition built from such pairs can still satisfy the other hypotheses and break Künneth.","tokens_in":41279,"feed_emoji":"🧩","tokens_out":10627,"duration_ms":89973,"temperature":0.7,"pith_summary":"This paper introduces an 'approximate ideal structure' for a C*-algebra: a positive contraction h that almost commutes with any finite set and sends the algebra, up to small errors, into two subalgebras C and D, with certain products landing near their intersection. The headline result is a permanence theorem: if A admits a uniform approximate ideal structure over pairs (C,D) for which C, D, and C∩D all satisfy the Künneth formula, then A satisfies the Künneth formula. A second theorem shows that if the three pieces all have trivial K-theory, then A has trivial K-theory, and this reproves a Baum-Connes-style vanishing result without controlled K-theory. The paper also exhibits examples: groupoid decompositions give uniform structures, while nuclear-dimension-one algebras give only a weak version. If the main theorem is right, the Künneth property is stable under a much more flexible class of decompositions than the classical Mayer-Vietoris ideal setting.","feed_headline":"Künneth formula survives approximate ideal decompositions","feed_subtitle":"A decomposition into subalgebras satisfying the Künneth formula forces the whole C*-algebra to satisfy it.","key_machinery":"The object carrying the argument is an approximate ideal structure (Definition 1.1): for every finite set and tolerance there is a positive contraction h in the multiplier algebra such that h almost commutes with the set, h moves elements into C, $1-h$ moves them into D, and $h(1-h)$ and $h^2(1-h)$ move them into $C\\cap D$. The uniform version (Definition 5.1) adds a decay function f: if $c\\in C\\otimes B$ and $d\\in D\\otimes B$ satisfy $\\|c-d\\|\\le \\delta$, some $x\\in (C\\cap D)\\otimes B$ lies within $f(\\delta)$ of both; this makes the intersection of tensor products behave correctly (Lemma 6.4) and drives exactness at the summation position (Proposition 5.7). Boundary classes $B_v(u)\\in K_0(C\\cap D)$, constructed from 'lifts' of unitaries, supply the missing connecting map in the approximate Mayer-Vietoris sequence. The inverse Bott map, implemented by an asymptotic family, is used to move between $K_1$ and $K_0$ statements in the Künneth product comparisons.","core_discovery":"The central claim is Theorem 1.4: a C*-algebra A that admits a uniform approximate ideal structure over a set of pairs (C,D) of C*-subalgebras satisfies the Künneth formula whenever each C, each D, and each C∩D does. The proof develops partial exactness for the sequence $K_1(C\\cap D) \\to K_1(C)\\oplus K_1(D) \\to K_1(A) \\to K_0(C\\cap D) \\to \\cdots$ at three positions, using boundary classes built from approximate lifts. Exactness at the two harder positions requires the uniformity assumption (Definition 5.1), which guarantees that for every B the natural inclusion $(C\\cap D)\\otimes B \\subseteq C\\otimes B \\cap D\\otimes B$ is an equality with controlled error. Surjectivity and injectivity of the Künneth product $K_*(A)\\otimes K_*(B) \\to K_*(A\\otimes B)$ are then transferred from the corresponding products for C, D, and C∩D. The paper also proves Theorem 1.2/3.8, that an approximate ideal structure over pieces with trivial K-theory forces A to have trivial K-theory.","pith_inferences":["A natural test is whether the f-uniformity condition can be relaxed to a B-dependent decay; the paper's Example 5.3 shows ordinary approximate ideal structures need not satisfy it, since hereditary subalgebras of the compact operators can fail the property badly.","Remark A.3, if checked, would distinguish the weak and ordinary notions: weak structures from nuclear dimension one would not imply approximate ideal structures, because a Kirchberg algebra with torsion in $K_1$ would have no boundary-class construction over finite-dimensional pieces.","The paper explicitly leaves open whether its range of validity coincides with controlled K-theory; a reader could test this by seeking a uniform approximate ideal structure not arising from a controlled filtration.","Because the proof uses ordinary K-theory only, the same boundary-class technology may be adaptable to crossed products or exact groupoid algebras where controlled filtrations are unavailable, though the paper does not claim this."],"forward_implications":["Theorem 1.4 implies that the Künneth property is closed under this kind of decomposition: once the three subalgebras in each pair satisfy Künneth, the ambient algebra does, for every B with free abelian K-theory.","Theorem 1.2 yields a proof, without controlled K-theory, that an algebra with an approximate ideal structure over pieces with trivial K-theory has trivial K-theory, recovering a Baum-Connes-style vanishing statement.","Appendix B shows that strong finite dynamical complexity for a principal ample groupoid implies its reduced C*-algebra satisfies Künneth, and that uniform Roe algebras of bounded geometry spaces with finite decomposition complexity satisfy Künneth.","Appendix A shows every separable nuclear-dimension-one C*-algebra has a weak approximate ideal structure over subhomogeneous pieces, evidence that the defining conditions are natural even though weak structures are not yet enough for the K-theory conclusions."],"supporting_citations":[{"why":"Supplies the controlled K-theory method and the lemma on which the paper's exactness at position (III), Proposition 5.7, is closely based, together with the approximate Mayer-Vietoris strategy.","marker":"[25]"},{"why":"Introduces the groupoid decomposition notion used in Appendix B to produce (uniform) approximate ideal structures and the Baum-Connes vanishing application.","marker":"[16]"},{"why":"Defines nuclear dimension and supplies the structure theory from which Appendix A derives weak approximate ideal structures.","marker":"[35]"},{"why":"Provides the K-theory product formulas and the $K_1\\times K_0$ product used to compare boundary classes with the Künneth product.","marker":"[19]"},{"why":"Establishes the Künneth-formula framework for C*-algebras, including the equivalence with the short exact sequence involving Tor that the paper targets.","marker":"[30]"},{"why":"Shows commutative C*-algebras satisfy the Künneth formula, the base case from which the paper's permanence statements are meant to extend.","marker":"[1]"},{"why":"Relates the Künneth formula to the universal coefficient theorem, framing why the formula is a desirable proxy for the UCT.","marker":"[29]"},{"why":"Provides counterexamples showing the inclusion $(C\\cap D)\\otimes B \\subseteq (C\\otimes B)\\cap (D\\otimes B)$ can be proper, motivating the f-uniformity assumption and Lemma 6.4.","marker":"[20]"}],"fun_headline_variants":["Künneth formula holds under approximate ideal splits","Approximate ideals guarantee Künneth formula","When subalgebras satisfy Künneth, so does the whole","Uniform approximate ideals force Künneth","K-theory rigidity from approximate decompositions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the f-uniformity of every pair in the class: whenever an element of $C\\otimes B$ is close to an element of $D\\otimes B$, there must be one element of $(C\\cap D)\\otimes B$ close to both, with the closeness controlled by a fixed decay function. This condition is what makes the tensor-product intersection identification and the exactness-at-position-III argument work; the paper's own Example 5.3 shows it fails for natural hereditary subalgebras of the compact operators, so the Künneth theorem does not follow from the existence of an approximate ideal structure alone.","fun_headline_variants_meta":{"raw":{"variants":["Künneth formula holds under approximate ideal splits","Approximate ideals guarantee Künneth formula","When subalgebras satisfy Künneth, so does the whole","Uniform approximate ideals force Künneth","K-theory rigidity from approximate decompositions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000204,"raw_usage":{"total_tokens":1422,"prompt_tokens":1011,"completion_tokens":411,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":627,"completion_tokens_details":{"reasoning_tokens":336}},"tokens_in":627,"tokens_out":411,"duration_ms":4392,"temperature":1.0,"reasoning_tokens":336,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:17:52.622998+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"To refute Theorem 1.4 one would need a C*-algebra A with a uniform approximate ideal structure over pairs (C,D) for which C, D, and C∩D satisfy the Künneth formula, but for which the product map $K_*(A)\\otimes K_*(B) \\to K_*(A\\otimes B)$ fails to be an isomorphism for some B with free abelian K-theory; the theorem asserts no such example exists. A concrete place to look is any pair violating f-uniformity, such as the hereditary subalgebras of the compact operators described in Example 5.3, checking whether a decomposition built from such pairs can still satisfy the other hypotheses and break Künneth.","supporting_citations":[{"cited_title":"Oyono-Oyono and G","cited_arxiv_id":null,"evidence_quote":"Supplies the controlled K-theory method and the lemma on which the paper's exactness at position (III), Proposition 5.7, is closely based, together with the approximate Mayer-Vietoris strategy."},{"cited_title":"Winter and J","cited_arxiv_id":null,"evidence_quote":"Defines nuclear dimension and supplies the structure theory from which Appendix A derives weak approximate ideal structures."},{"cited_title":"Higson and J","cited_arxiv_id":null,"evidence_quote":"Provides the K-theory product formulas and the $K_1\\times K_0$ product used to compare boundary classes with the Künneth product."},{"cited_title":"Schochet","cited_arxiv_id":null,"evidence_quote":"Establishes the Künneth-formula framework for C*-algebras, including the equivalence with the short exact sequence involving Tor that the paper targets."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows commutative C*-algebras satisfy the Künneth formula, the base case from which the paper's permanence statements are meant to extend."},{"cited_title":"Rosenberg and C","cited_arxiv_id":null,"evidence_quote":"Relates the Künneth formula to the universal coefficient theorem, framing why the formula is a desirable proxy for the UCT."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides counterexamples showing the inclusion $(C\\cap D)\\otimes B \\subseteq (C\\otimes B)\\cap (D\\otimes B)$ can be proper, motivating the f-uniformity assumption and Lemma 6.4."}],"review_version":1}