{"id":"be34f37c-91f2-40da-a267-f02125029b80","arxiv_id":"2509.03759","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An etale groupoid's invariant measures pair canonically with its zeroth Crainic-Moerdijk homology, yielding a groupoid Elliott invariant shown to match the C*-algebraic Elliott invariant for many integer actions and orbit-breaking constructions.","lead":"The paper defines a trace pairing between invariant probability measures and groupoid homology, giving an Elliott invariant for etale groupoids. It proves this refined invariant matches the C*-algebraic Elliott invariant for irrational rotations and orbit-breaking systems, refining Matui's HK conjecture.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 0.6 omits a freeness hypothesis: for non-free Z-actions the trace simplex of C*_r(Z⋉X) is larger than T(G), so HK-good fails.","rationale":"The reader's weakest_assumption correctly identifies the missing freeness condition. The paper's own machinery shows the issue: Lemma 2.8 proves τ is an affine homeomorphism only for principal groupoids, and HK-goodness (Definition 2.9) requires this homeomorphism. The trivial action example is decisive and simple, and it shows Theorem 0.6 is false verbatim. This is not an internal inconsistency in the main technical development—the computations for free actions are coherent, and the later sections explicitly assume freeness—so a conditional acceptance with the hypothesis added is the appropriate outcome. The reader already reached CONDITIONAL, and my concern strengthens that judgment without moving it to rejection. I found no separate load-bearing flaw in the pairing construction, the hyperhomology comparison, or the orbit-breaking long exact sequence; the central issue is the statement-level omission of freeness. A concrete check—computing trace simplices for a non-free action—settles the matter immediately.","tokens_in":49087,"tokens_out":5021,"duration_ms":54083,"concrete_test":"Let X=S¹ with the trivial Z-action. Compute the two trace simplices: T(Z⋉X)=Prob(S¹) (all invariant Borel probability measures on the base) and T(C*_r(Z⋉X))=Prob(S¹×S¹) ≅ T(C*(Z)⊗C(S¹)). Since Prob(S¹×S¹) is strictly larger than Prob(S¹), the map τ of Lemma 2.8 is not surjective, so condition (iii) of Definition 2.9 fails. This directly contradicts the unconditional statement of Theorem 0.6 for a d-sphere. Replacing the trivial action by any non-free action, e.g. a reflection of S¹ with fixed points, gives the same failure.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Definition 2.9 requires that the map τ:T(G)→T(C*_r(G)) of Lemma 2.8 be an affine homeomorphism. Lemma 2.8 guarantees this only when G is principal, i.e. for a transformation groupoid Z⋉X only when the Z-action is free. Theorem 0.6, however, states that Z⋉X is HK-good for any action on a sphere, torus, or space of covering dimension at most three, with no freeness condition. The proof route via Corollary 5.10 likewise does not state freeness. The omission is not cosmetic: if the action has isotropy, T(C*_r(G)) is strictly larger than T(G). For example, take the trivial action on S¹. Then G=Z×S¹, T(G)=Prob(S¹), but C*_r(G)≅C(S¹)⊗C(S¹), so T(C*_r(G))=Prob(S¹×S¹), which is strictly larger than Prob(S¹). The map τ is injective but not surjective, so Z⋉S¹ is not HK-good. Thus Theorem 0.6 as stated is false. The surrounding text suggests the intended statement is for free actions: Example 6.1 explicitly says 'take any free Z-action', and Section 8 fixes a free action. The fix is to add 'free' to Theorem 0.6 and to Corollary 5.10 if its hypotheses are meant to imply HK-good.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines a canonical pairing between the simplex of invariant Borel probability measures on the unit space of a locally compact, Hausdorff, etale groupoid with compact base space and its zeroth Crainic-Moerdijk groupoid homology. This pairing is used to define a 'groupoid Elliott invariant' (H_{**}(G), T(G), rho_H, [1]), and a groupoid is called HK-good if this invariant is isomorphic to the C*-algebraic Elliott invariant of C*_r(G). The main positive results are that transformation groupoids Z⋉X for actions on spheres, tori, or spaces of covering dimension at most three are HK-good (Theorem 0.6), and that certain orbit-breaking groupoids arising from point-like and Cantor-like systems are HK-good (Theorem 0.7). The proof strategy combines a Pimsner-Voiculescu-type long exact sequence for groupoid homology, a Chern character from K-theory to homology for integer actions, a long exact sequence for open inclusions of etale groupoids, and detailed computations for orbit-breaking systems. The paper also contains appendices on groupoid homology for infinite-dimensional base spaces, group hyperhomology, and integral Chern characters.","tokens_in":49499,"tokens_out":5096,"duration_ms":50765,"significance":"If the stated results are correct, the paper makes a substantial contribution to the program relating groupoid homology to C*-algebraic K-theory and traces. The definition of the trace pairing for non-ample etale groupoids is new and natural, and the explicit computations—especially the irrational rotation example and the three different HK-good models for the irrational rotation algebra—are instructive and valuable. The paper also provides useful technical machinery, including a treatment of groupoid homology for infinite-dimensional base spaces and integral Chern characters in low dimension. However, the central theorem as stated is false because it omits a freeness hypothesis, and several load-bearing steps are left as 'inspection' or as unverified assumptions. These issues need to be fixed before the main claims are reliable.","major_comments":[{"comment":"Theorem 0.6 states that Z⋉X is HK-good for any action on a d-sphere, d-torus, or space of covering dimension at most three, without any freeness hypothesis. Definition 2.9(iii) requires the canonical map tau: T(G) -> T(C*_r(G)) of Lemma 2.8 to be an affine homeomorphism, but Lemma 2.8 establishes this only when G is principal. For transformation groupoids, principality is equivalent to freeness of the action. The omission is not cosmetic: for the trivial action on S^1, G = Z×S^1, T(G) = Prob(S^1), but C*_r(G) ≅ C(S^1)⊗C(S^1), so T(C*_r(G)) ≅ Prob(S^1×S^1), which is strictly larger. The map tau is injective but not surjective, so Z⋉S^1 is not HK-good. Thus Theorem 0.6 is false as stated. The surrounding text indicates the intended hypothesis is free actions (see Section 6 and Example 6.1). The fix is to add 'free' to Theorem 0.6 and Corollary 5.10, or to prove an alternative surjectivity","section":"Theorem 0.6; Definition 2.9; Lemma 2.8"},{"comment":"Theorem 7.5 asserts that Putnam's and Matui's results on the groupoids H and H' for an open inclusion extend from ample, second countable groupoids to general etale groupoids with the single sentence 'Inspection of the arguments reveals that those assumptions are not necessary.' This is a load-bearing step: it underpins Proposition 7.6 and Corollary 8.3, which are used in the orbit-breaking computations of Sections 8–10. The extension is not obvious and should be proved in detail, or a precise reference should be provided for the non-ample case. If the existing arguments genuinely apply, the authors should say how the use of total disconnectedness is avoided; if not, the orbit-breaking theorems are not established.","section":"Theorem 7.5"},{"comment":"Proposition 8.8 assumes in part (iii) that the quotient map K0(C*_r(RY)) -> K0(C*_r(R_phi)) splits, with the parenthetical admission 'We do not know if this last condition is necessary: it is possible it follows from the other assumptions.' This is an explicit unproved assumption. More importantly, Theorem 9.4 claims that the point-like orbit-breaking groupoids are HK-good as an 'immediate' consequence of Proposition 8.8, but the splitting hypothesis is never verified for the point-like construction. The proof of HK-goodness for Section 9 is therefore incomplete. Either the splitting must be proved, or the statement of Theorem 9.4 must be conditional on it.","section":"Proposition 8.8(iii)"},{"comment":"The proof of Proposition 5.2(iii) says that the commutative diagram of short exact sequences follows because the Pimsner-Voiculescu sequence is 'exactly defined by forcing the given diagram to commute.' This is not a rigorous argument. Part (iii) is essential for the construction of the Chern character in Theorem 5.9 and for the K-theory/homology comparisons in Section 6. The authors should give a direct proof, or at least provide a precise commutative-diagram chase from the known six-term exact sequences, so that the claimed vertical isomorphisms are justified.","section":"Proposition 5.2(iii)"}],"minor_comments":[{"comment":"The title/abstract contain a typo: 'P AIRING' should be 'PAIRING'.","section":"Abstract"},{"comment":"The text refers to 'Analogously to Lemma 5.3' when the preceding result is Proposition 5.3; the cross-reference should be corrected.","section":"Section 5, after Lemma 5.4"},{"comment":"Example 6.1 explicitly says 'take any free Z-action', which is consistent with the needed hypothesis. The contrast with the statement of Theorem 0.6 should be noted and reconciled in the introduction.","section":"Example 6.1"},{"comment":"The notation S^1_B and S^1 for the sheaf of Borel/continuous circle-valued functions is confusing because S^1 is also used for the unit circle as a space. Please use distinct notation.","section":"Remark 4.2"},{"comment":"Conjecture 2.10 repeats Conjecture 0.5; consider keeping only one statement to avoid redundancy.","section":"Conjecture 2.10"}],"recommendation":"major_revision","confidential_remarks":"The paper is clearly a serious contribution and the central framework appears sound, but the missing freeness hypothesis in Theorem 0.6 is a false statement as written, and the unproved 'inspection' arguments in Theorem 7.5 and Proposition 8.8(iii) leave important gaps. These are fixable within the scope of the paper, so I recommend major revision rather than rejection. The authors should be encouraged to add the freeness hypothesis to all relevant statements, supply proofs for the claimed extensions, and verify the splitting condition in the point-like case."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth reading: it defines a genuinely new trace pairing between invariant measures and Crainic–Moerdijk homology for non-ample étale groupoids, packages it into a groupoid Elliott invariant, and proves HK-goodness for a substantial class of examples (irrational rotations, orbit-breaking systems) that go beyond the usual totally disconnected setting. Proposition 2.6 is proved cleanly, and the explicit computation for irrational rotation groupoids is a nice touch.\n\nThe soft spot is a missing hypothesis that is not cosmetic. Theorem 0.6 and Corollary 5.10 claim HK-goodness for any integer action on spheres, tori, or dimension-at-most-three spaces. But Definition 2.9 requires the map τ: T(G) → T(C*_r(G)) to be an affine homeomorphism, and Lemma 2.8 only gives this when G is principal. For a transformation groupoid that means the action is free. The trivial action on S^1 is a counterexample: T(G) is Prob(S^1), while C*_r(Z⋉S^1) ≅ C(S^1)⊗C(S^1) has trace space Prob(S^1×S^1). So the results as stated are false without freeness. The fix is simple – add 'free' to the hypotheses – and the paper's own examples (Section 8, Example 6.1) already assume it. This should be caught and corrected.\n\nTwo smaller issues: Theorem 7.5 extends Matui/Putnam by 'inspection' to non-ample, non-second-countable groupoids without a detailed argument; that step deserves more justification. And Proposition 8.8(iii) rests on a splitting assumption the authors do not know is necessary. These are minor relative to the main construction, and the authors flag them.\n\nOverall: the central construction and the positive examples hold up. The paper is a serious contribution to the HK conjecture program, aimed at operator algebraists who care about classification and groupoid models. It deserves rigorous peer review, provided the freeness hypothesis is fixed.","headline":"New trace pairing and groupoid Elliott invariant with a fixable but real freeness omission in the main theorem.","tokens_in":49908,"tokens_out":3631,"would_cite":true,"duration_ms":36449,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L80","22A22"],"pacs":[],"model":"deepseek-v4-flash","headline":"Invariant Borel measures pair canonically with groupoid homology, yielding a groupoid Elliott invariant that agrees with the C*-algebraic invariant in many integer-action and orbit-breaking examples.","keywords":["HK-conjecture","groupoid homology","K-theory","Elliott invariant","invariant measures","etale groupoids","Chern character","orbit breaking"],"falsifier":"Take a minimal but non-free integer action on a sphere or torus, for example a minimal homeomorphism with a periodic point, and compute the tracial state space of its reduced crossed product. If there is a trace that does not arise from integration against an invariant Borel probability measure on the base space, the trace-space homeomorphism fails, the groupoid Elliott invariant differs from the C*-algebraic one, and the unrestricted statement of the transformation-groupoid theorem would be false.","tokens_in":48985,"feed_emoji":"🎯","tokens_out":9657,"duration_ms":93345,"temperature":0.7,"pith_summary":"This paper introduces a canonical pairing between the invariant Borel probability measures on the unit space of an etale groupoid and the zeroth groupoid homology group, and packages it into a 'groupoid Elliott invariant' modeled on the C*-algebraic Elliott invariant. The point is to sharpen the conjecture that K-theory of a groupoid C*-algebra is isomorphic to groupoid homology: instead of only asking whether those groups agree, the paper asks whether the whole invariant—homology, invariant measures, trace pairing, and unit class—matches the C*-algebraic Elliott invariant. The authors prove this 'HK-good' property for many transformation groupoids, including irrational rotations and actions on spheres, tori, and spaces of covering dimension at most three, and for orbit-breaking groupoids built from point-like and Floyd-type systems. If correct, this gives a homological route to the data that classifies a large family of C*-algebras.","feed_headline":"Invariant measures tie groupoid homology to Elliott invariants","feed_subtitle":"For integer actions and orbit-breaking groupoids, the groupoid Elliott invariant matches the C*-algebraic one.","key_machinery":"The load-bearing mechanism is the pairing between invariant measures and H_0, together with a Chern character from the K-theory of an integer crossed product to the rational groupoid homology of Z⋉X, built through the mapping torus of the action. The pairing is defined using a canonical Borel resolution of the constant sheaf, and the Chern character is shown to be compatible with both the measure pairing and the C*-algebraic trace pairing, which is what makes the full Elliott invariants match.","core_discovery":"For a locally compact Hausdorff etale groupoid with compact base space, the paper defines a pairing rho_H from the simplex of invariant probability measures on the unit space to Hom(H_0(G), R), using integration against a canonical Borel resolution; invariance of the measure is exactly what makes the pairing descend to homology. With this pairing it defines the groupoid Elliott invariant (H_{**}(G), T(G), rho_H, [1]) and calls G HK-good if this invariant is isomorphic to the Elliott invariant of the reduced groupoid C*-algebra, meaning compatible K-theory/homology isomorphisms, an affine homeomorphism between invariant measures and tracial states, and matching pairings. The main results esta","pith_inferences":["The pairing should extend to higher homology via cyclic cocycles and higher-dimensional invariant currents, connecting to noncommutative geometry; the paper hints at this but does not develop it.","If the freeness assumption is truly required for the trace-space homeomorphism, the main theorem about transformation groupoids is best read as a theorem about free actions; testing non-free minimal actions on spheres would settle whether the statement can be relaxed.","The mapping-torus Chern character for Z-actions suggests a template for higher-rank actions: a Z^d-version via higher mapping tori would likely yield HK-goodness for many Z^d-transformation groupoids.","The existence of multiple HK-good models for the irrational rotation algebra suggests that the groupoid Elliott invariant is a property of a model, not of the C*-algebra; a natural next step is to ask which classifiable algebras admit at least one such model."],"forward_implications":["For any etale groupoid with compact base, invariant measures now produce explicit numerical invariants of H_0, so trace data can be read directly from groupoid homology.","Integer actions on spheres, tori, and dimension-at-most-three spaces have a complete groupoid Elliott invariant, matching the C*-algebraic invariant used in classification.","Orbit-breaking groupoids with low-dimensional break loci provide HK-good models for many classifiable C*-algebras, including the point-like and Cantor-like systems from earlier constructions.","The same C*-algebra can admit HK-good models with different Z-graded groupoid homology; only after passing to Z/2-graded homology do the invariants agree, so the Z-grading carries extra model-dependent information.","Minimal integer actions on spaces such as S^3 × RP^4 are not HK-good, showing the property is a genuine refinement and not automatic once K-theory and homology agree."],"supporting_citations":[{"why":"defines the groupoid homology used throughout.","marker":"[9]"},{"why":"constructs the reduced groupoid C*-algebra whose Elliott invariant is the comparison target.","marker":"[54]"},{"why":"supplies the six-term exact sequence for integer crossed products used to compute K-theory.","marker":"[44]"},{"why":"supplies the Thom-isomorphism and dual-trace machinery used to make the Chern character compatible with trace pairings.","marker":"[8]"},{"why":"states the conjecture that K-theory of a groupoid C*-algebra is isomorphic to groupoid homology, which the paper refines with pairings.","marker":"[35, 36]"},{"why":"gives the long exact sequence for open inclusions that the paper generalizes beyond the ample case.","marker":"[37]"},{"why":"supplies the orbit-breaking excision results in K-theory that the new homology long exact sequence mirrors.","marker":"[51]"},{"why":"constructs the point-like and Cantor-like dynamical systems whose orbit-breaking groupoids are shown to be HK-good.","marker":"[17]"}],"fun_headline_variants":["Pairing links groupoid homology and invariant measures to Elliott","New pairing yields groupoid Elliott invariant, matches C*-algebraic","Groupoid Elliott invariant defined via homology and traces","Invariant measure pairing refines groupoid homology to Elliott","Canonical pairing gives groupoid Elliott invariant, verified for key cases"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The claim that the invariant-measure space of a groupoid is homeomorphic to the tracial state space of its reduced C*-algebra is proved only for principal groupoids; the HK-good theorems for transformation groupoids rely on it, so their statements implicitly require the integer action to be free.","fun_headline_variants_meta":{"raw":{"variants":["Pairing links groupoid homology and invariant measures to Elliott","New pairing yields groupoid Elliott invariant, matches C*-algebraic","Groupoid Elliott invariant defined via homology and traces","Invariant measure pairing refines groupoid homology to Elliott","Canonical pairing gives groupoid Elliott invariant, verified for key cases"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00019,"raw_usage":{"total_tokens":1173,"prompt_tokens":736,"completion_tokens":437,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":480,"completion_tokens_details":{"reasoning_tokens":353}},"tokens_in":480,"tokens_out":437,"duration_ms":4721,"temperature":1.0,"reasoning_tokens":353,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T10:41:40.266384+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a minimal but non-free integer action on a sphere or torus, for example a minimal homeomorphism with a periodic point, and compute the tracial state space of its reduced crossed product. If there is a trace that does not arise from integration against an invariant Borel probability measure on the base space, the trace-space homeomorphism fails, the groupoid Elliott invariant differs from the C*-algebraic one, and the unrestricted statement of the transformation-groupoid theorem would be false.","supporting_citations":[{"cited_title":"Crainic and I","cited_arxiv_id":null,"evidence_quote":"defines the groupoid homology used throughout."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"constructs the reduced groupoid C*-algebra whose Elliott invariant is the comparison target."},{"cited_title":"Pimsner and D.-V","cited_arxiv_id":null,"evidence_quote":"supplies the six-term exact sequence for integer crossed products used to compute K-theory."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the Thom-isomorphism and dual-trace machinery used to make the Chern character compatible with trace pairings."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the long exact sequence for open inclusions that the paper generalizes beyond the ample case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the orbit-breaking excision results in K-theory that the new homology long exact sequence mirrors."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"constructs the point-like and Cantor-like dynamical systems whose orbit-breaking groupoids are shown to be HK-good."}],"review_version":1}