A partitioned manifold index theorem for noncompact hypersurfaces

Peter Hochs; Thijs de Kok

REVIEW 2 major objections 4 minor 34 references

A partitioned manifold index theorem for noncompact hypersurfaces

T0 review · 2 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The partitioned manifold index theorem holds for noncompact hypersurfaces when the embedding is uniformly controlled.

desk verdict A solid, genuinely new generalization of Roe's partitioned index theorem, but the proof leans on a sign correction that deserves independent checking. read the letter →

arxiv 2507.16591 v2 pith:NQP6NIGZ submitted 2025-07-22 math.DG math.KTmath.OA

Peter Hochs , Thijs de Kok This is my paper · ORCID

classification math.DGmath.KTmath.OA MSC 58J2019K5653C2746L80

keywords partitionedmanifoldindextheoremnoncompacthypersurfaceRoealgebracoarseDiracoperatorpositivescalarcurvatureuniformequivalencecobordisminvariance

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper extends Roe's partitioned manifold index theorem, originally for compact hypersurfaces, to complete Riemannian manifolds cut along a possibly noncompact hypersurface. The central claim is that the index of the Dirac operator on the hypersurface $N$ equals, under a canonical map $\Phi$, the partitioned index built from the ambient Dirac operator, provided the embedding of $N$ into $M$ is uniformly controlled. This equality connects noncompact geometry to the $K$-theory of Roe algebras, yielding obstructions to uniformly positive scalar curvature and a generalized cobordism invariance of coarse indices. The proof establishes the equality first on product manifolds $\mathbb{R}\times N$ and then reduces the general case to that product case by a gluing construction.

What carries the argument

The central objects are the Roe algebra $C^*(M)$ and its localized version $C^*(N\subseteq M)$; a Roe algebra is the $C^*$-algebra generated by locally compact finite-propagation operators, and it is the receptacle for Dirac indices on noncompact manifolds. The theorem is an identity between two classes in the $K$-theory of these algebras: $\Phi(\mathrm{Ind}(D_N))$ and $\mathrm{Ind}(D;N)$, the latter defined through the boundary map applied to a unitary built from the Cayley transform of $D$ and a smooth step function across $N$. In the product case $M=\mathbb{R}\times N$, the proof runs through a commutative diagram that uses Paschke duality and suspension to identify the $K$-homology class of $D_N$ with the Cayley-transform class. The general case is carried back to the product by gluing a transition manifold $M'$ that looks like $M$ on one side and $\mathbb{R}\times N$ on the other, with uniform half-isomorphisms — isometries of the open halves that extend to uniform equivalences and conjugate the Dirac operators — transferring the equality; the three embedding assumptions are exactly what keeps these maps uniform and the glued metric complete.

What would settle it

A concrete test is a family of warped product metrics $g_M = dr^2 + \phi(r)^2 g_N$ on $M = \mathbb{R}\times N$, with $N$ a noncompact spin manifold whose Dirac index in $K_0(C^*(N))$ is nonzero and $\phi$ bounded between positive constants. The product argument gives an explicit prediction for both $\Phi(\mathrm{Ind}(D_N))$ and $\mathrm{Ind}(D;N)$; computing the partitioned index directly from the Cayley transform for such a metric and finding any mismatch between the two classes would refute Theorem 2.23.

Watch

Extended reading notes

Core claim

The paper's central claim is Theorem 2.23: for an oriented complete Riemannian manifold $M$ of odd dimension at least three, partitioned by a connected hypersurface $N$ that is uniformly coarsely equivalent to $N$ with its subspace metric, admits a uniform tubular neighbourhood, and has a metric on that tube uniformly comparable to the product metric, the equality $\Phi(\mathrm{Ind}(D_N)) = \mathrm{Ind}(D;N)$ holds in the $K$-theory of the Roe algebra of $N$. Here $D_N$ is the Dirac operator induced on the hypersurface, $\mathrm{Ind}(D;N)$ is the partitioned index constructed from the ambient Dirac operator $D$, and $\Phi$ is the canonical comparison map between the two $K$-theory groups. The theorem recovers Roe's original statement when $N$ is compact. The proof first verifies the equality for $M = \mathbb{R}\times N$ partitioned by $\{0\}\times N$, then transports the result to a general $M$ through a transition manifold that interpolates between $M$ and the product manifold.

Load-bearing premise

The load-bearing premise is that distances between points measured along the noncompact hypersurface remain uniformly comparable to distances between the same points measured inside the ambient manifold.

Editorial extensions

If this is right

If the hypersurface Dirac operator has nonzero index in $K_0(C^*(N))$, then $M$ carries no metric of uniformly positive scalar curvature.
Two noncompact hypersurfaces that are cobordant in the paper's sense have coarse Dirac indices that vanish together.
When a discrete group acts properly and freely with $N/\Gamma$ compact, the equivariant equality lives in $K_0(C^*_r(\Gamma))$, giving higher-index obstructions for $M/\Gamma$.
For a compact hypersurface with injective $\pi_1(N)\to \pi_1(M)$, the higher index of the lifted spin-Dirac operator on the universal cover obstructs uniformly positive scalar curvature on $M$.
With finitely many connected components, a direct-sum version of the equality holds under an additional uniformity condition on the components of $M_+$.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The transition-manifold construction suggests the same gluing argument could prove partitioned-index equalities for other geometric operators, such as twisted or signature Dirac operators, once the product case is recomputed.
The paper gives sufficient conditions but does not address necessity; if condition (i) is also necessary, uniform comparability of intrinsic and ambient distances is the sharp dividing line for when a noncompact cut has a well-defined partitioned index.
In the equivariant setting with compact quotient, pairing the resulting classes with cyclic cocycles would convert the theorem into numerical positive-scalar-curvature obstructions for noncompact quotients, a step the paper signals but does not fully execute.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague, this paper proves a real extension of Roe's partitioned manifold index theorem to noncompact hypersurfaces. The key achievement is dropping the product-structure assumption near the hypersurface that Bunke-Ludewig and Engel-Wulff still need. The authors replace it with three geometric conditions on the embedding, and show that when those hold the partitioned index equality still works. That is a substantive result, not a repackaging. The proof is organized in two clear stages: a product case using coarse geometry and K-theory, then a reduction to that case via transition manifolds and half-isomorphisms. The comparison with the prior literature is honest, and the corollaries (cobordism invariance, PSC obstructions, equivariant version) are natural and useful.

The main soft spot is Lemma 1.34. The authors assert that the suspension of the Dirac class s(d) is -1, correcting [12, Lem. 9.5.7]. That sign is load-bearing: Proposition 3.8 uses it to get the sign of [D_N] in terms of T_+, and Proposition 3.14 then delivers the equality with the partitioned index. If the sign were +1, the theorem would assert the negative equality and contradict the compact case. The proof of Lemma 1.34 is a brief reference to two lemmas in [12] and a remark about a missing adjoint; it is not independently verified. I cannot check it off the top of my head, but this is exactly the kind of sign that a referee should be asked to verify carefully. It is not a manufactured flaw: the stress-test note is right that this is more load-bearing than the geometric assumptions, which only delimit the admissible hypersurfaces.

The geometric assumption (i) on uniform equivalence of the intrinsic and subspace metrics is nontrivial, but the authors use it precisely where the transition map needs controlled curves. That seems to be the cost of the generality, not a hidden circularity.

Overall, the paper is serious, the proof is structured, and the central argument holds up as far as I can tell, modulo the sign check. It deserves a careful referee. Send it to review, with a specific request to verify Lemma 1.34.

Referee Report

2 major / 4 minor

Summary. This paper proves a partitioned manifold index theorem for complete odd-dimensional Riemannian manifolds M partitioned by a connected, possibly noncompact hypersurface N. Under the three explicit conditions of Theorem 2.23—(i) the identity map from (N, d_N) to (N, d_M|_N) is a uniform equivalence, (ii) N admits a uniform tubular neighbourhood, and (iii) the metric on the collar is uniformly comparable to the product metric—the authors prove the equality Φ(Ind(D_N)) = Ind(D;N). Here Ind(D_N) is the coarse index of the Dirac operator on N and Ind(D;N) is the partitioned index constructed via the Cayley transform and the Roe homomorphism. The proof is organised in two stages: the product case in Section 3, and a reduction of the general case to the product case in Section 4 using half-isomorphisms and a transition manifold. Section 5 gives equivariant and disconnected generalizations.

Significance. If the main theorem is correct, this is a substantial generalization of Roe's partitioned manifold index theorem to noncompact hypersurfaces, with natural applications to uniformly positive scalar curvature obstructions and to cobordism invariance of coarse indices. The paper is well structured and transparent about the hypotheses: the three geometric conditions are stated precisely, the product case is proven separately, and the reduction via half-isomorphisms is a coherent and workable strategy. The comparison with the recent work of Bunke–Ludewig and Engel–Wulff is also useful. The main caveat concerns the sign correction in Lemma 1.34, which is load-bearing for the equality as stated and needs independent verification.

major comments (2)

[§1.4, Lemma 1.34; §3.3–3.4] The sign of the suspension map s(d) is asserted to be -1 as a correction to [12, Lem. 9.5.7], but the argument is only two sentences appealing to a comparison of operators W_1 in [12]. This sign enters the proof of Theorem 3.1 at Proposition 3.8, where [D_N] is identified with -[q(T_+)], and then propagates through Proposition 3.14 to the equality Φ(Ind(D_N)) = Ind(D;N). If s(d) = +1, the signs in Propositions 3.8 and 3.14 would reverse, and Theorem 2.23 would assert the opposite equality, contradicting Corollary 2.25 and Roe's theorem in the compact case. The manuscript should include a complete, self-contained computation of s(d), or at minimum a detailed line-by-line verification of the claimed correction to [12, Lem. 9.5.7], before the main theorem can be accepted.
[§4.1, Proposition 4.17] Proposition 4.17, which asserts that the identity map on the product side is a uniform equivalence, is stated as 'analogous' to Proposition 4.16 but no proof is given. This result is used in Proposition 4.22(ii) to establish the second half-isomorphism and in Corollary 4.18 for completeness of the glued metric, so it is part of the load-bearing reduction. The proof should be supplied, or at least the differences from Proposition 4.16—where the curve modification uses assumption (i) and the metric comparison in assumption (iii)—should be explained explicitly.

minor comments (4)

[§2.1, Lemma 2.3] The proof that id_N is a uniform map is compressed: while uniform expansiveness follows from d_M|_N ≤ d_N, the metric-boundedness condition is not immediate from equality of topologies and uses closedness and completeness of N. This step should be spelled out.
[§2.2, Remark 2.12] The aside 'it may be true that every partitioning hypersurface is simple; we have not found any counterexamples' is not needed for the paper and is unsupported; it could be removed or replaced by a precise statement about which classes of hypersurfaces are known to be simple.
[§5.1, Theorem 5.2] The equivariant version is stated with the proof described only as 'completely analogous'. At minimum, the authors should list the equivariant versions of the key lemmas used, especially Lemma 1.34 and Paschke duality, so that the reader can check that the sign conventions and index maps behave equivariantly.
[§2.5, Corollary 2.25] The proof of Corollary 2.25 refers forward to Corollary 5.3. A direct proof that compactness of N implies the three assumptions of Theorem 2.23 would make the paper more self-contained and would avoid the detour through the equivariant theorem.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main equality is derived from independent coarse index theory and K-theory naturality, not from its own definition.

full rationale

The derivation chain in Theorem 2.23 is self-contained and does not reduce to its own inputs. The partitioned index Ind(D;N) is defined in Definition 2.14 directly from the Cayley transform of D and the boundary map of an exact sequence of localized Roe algebras, with no fitted parameters and no prior appeal to equality with Ind(D_N). The index Ind(D_N) is independently defined via the coarse assembly map in Definition 1.39. The product case (Theorem 3.1) is proved by explicit Paschke duality computations, suspension, and naturality of boundary maps, leading to the identification of the relevant K-theory classes. The general case is reduced to the product case by constructing a transition manifold and proving naturality of the partitioned index under uniform half-isomorphisms; this is a genuine geometric argument rather than a restatement of the theorem. The hypotheses (i)-(iii) in Theorem 2.23 delimit the class of manifolds to which the theorem applies and are used, for example, in Proposition 4.16 to control distances after gluing; they do not encode the conclusion. The only self-citations are to the authors' earlier works [9,10] in Remark 5.1, where they note subtleties in generalizing equivariant Roe algebras; this remark is not load-bearing for the non-equivariant theorem or its proof. The sign correction in Lemma 1.34 is a correction to an external textbook result [12] and is derived by comparing operators in that book, not by assuming the desired partitioned index theorem; whether the correction is correct is a correctness risk, not a circularity. No fitted input is renamed as a prediction, no defining quantity is assumed equal to the target, and no load-bearing support is drawn from the authors' own prior claims. Therefore the circularity score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper introduces no free parameters or invented entities. It relies on standard mathematical machinery and on geometric assumptions about the embedding of the hypersurface, which the authors take as hypotheses rather than derive.

assumptions (3)

standard math Standard coarse index theory and K-theory of Roe algebras are valid.
The proof relies on established results: Paschke duality, coarse assembly, six-term exact sequences, and functoriality of K-theory, as cited from Higson-Roe.
domain assumption The manifold M is a complete, oriented, odd-dimensional Riemannian manifold with a Dirac bundle satisfying the chiral condition (2.8).
These are explicit standing assumptions in Section 2.4 and are required for the definitions of the indices and the Clifford action on the hypersurface.
domain assumption The hypersurface N satisfies the three geometric conditions in Theorem 2.23: uniform coarse equivalence, uniform tubular neighborhood, and uniform metric comparison.
These conditions are used in the transition manifold construction (Section 4.3) and are the load-bearing premises that allow the reduction to the product case. They are not automatically true for arbitrary noncompact hypersurfaces.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A partitioned manifold index theorem for noncompact hypersurfaces." pith.science (2026). https://pith.science/paper/NQP6NIGZ

@misc{pith2026250716591,
  author       = {Pith},
  title        = {Pith review of: A partitioned manifold index theorem for noncompact hypersurfaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NQP6NIGZ}},
  note         = {Machine review of arXiv:2507.16591}
}

abstract

Roe's partitioned manifold index theorem applies when a complete Riemannian manifold $M$ is cut into two pieces along a compact hypersurface $N$. It states that a version of the index of a Dirac operator on $M$ localized to $N$ equals the index of the corresponding Dirac operator on $N$. This yields obstructions to positive scalar curvature, and implies cobordism invariance of the index of Dirac operators on compact manifolds. We generalize this result to cases where $N$ may be noncompact, under assumptions on the way it is embedded into $M$. This results in an equality between two classes in the $K$-theory of the Roe algebra of $N$. Bunke and Ludewig, and Engel and Wulff, have recently obtained related results based on different approaches.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 34 canonical work pages

[12]

Higson and J

N. Higson and J. Roe.AnalyticK-homology. Oxford Mathematical Monographs. Oxford University Press, Oxford, 2000. Oxford Science Publications

work page 2000
[1]

Atiyah and I.M

M.F. Atiyah and I.M. Singer. The index of elliptic operators. I.Ann. of Math. (2), 87:484–530, 1968

work page 1968
[2]

Atiyah and I.M

M.F. Atiyah and I.M. Singer. The index of elliptic operators. III.Ann. of Math. (2), 87:546– 604, 1968

work page 1968
[3]

P. Baum, A. Connes, and N. Higson. Classifying space for proper actions andK-theory of groupC ∗-algebras. InC ∗-algebras: 1943–1993 (San Antonio, TX, 1993), volume 167 of Contemp. Math., pages 240–291. Amer. Math. Soc., Providence, RI, 1994

work page 1943
[4]

Berline, E

N. Berline, E. Getzler, and M. Vergne.Heat kernels and Dirac operators, volume 298 of Grundlehren der mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences]. Springer-Verlag, Berlin, 1992

work page 1992
[5]

Coronas and Callias type operators in coarse geometry

U. Bunke and M. Ludewig. Coronas and Callias type operators in coarse geometry.arXiv preprint arXiv:2411.01646, 2024

work page Pith review arXiv 2024
[6]

Connes and H

A. Connes and H. Moscovici. Cyclic cohomology, the Novikov conjecture and hyperbolic groups.Topology, 29(3):345–388, 1990

work page 1990
[7]

Conway.A course in functional analysis, volume 96 ofGraduate Texts in Mathematics

J.B. Conway.A course in functional analysis, volume 96 ofGraduate Texts in Mathematics. Springer-Verlag, New York, second edition, 1990

work page 1990

Show all 34 references

[8]

Engel and C

A. Engel and C. Wulff. The relative index in coarse index theory and submanifold obstructions to uniform positive scalar curvature.arXiv preprint arXiv:2506.14301, 2025

2025 arXiv
[9]

H. Guo, P. Hochs, and V. Mathai. Coarse geometry and Callias quantisation.Trans. Amer. Math. Soc., 374(4):2479–2520, 2021

2021
[10]

H. Guo, P. Hochs, and V. Mathai. Equivariant Callias index theory via coarse geometry.Ann. Inst. Fourier (Grenoble), 71(6):2387–2430, 2021

2021
[11]

N. Higson. A note on the cobordism invariance of the index.Topology, 30(3):439–443, 1991

1991
[13]

Higson and J

N. Higson and J. Roe. Mapping surgery to analysis. I. Analytic signatures.K-Theory, 33(4):277–299, 2005

2005
[14]

Higson and J

N. Higson and J. Roe. Mapping surgery to analysis. II. Geometric signatures.K-Theory, 33(4):301–324, 2005

2005
[15]

Higson and J

N. Higson and J. Roe. Mapping surgery to analysis. III. Exact sequences.K-Theory, 33(4):325–346, 2005

2005
[16]

Karami, M.E

M. Karami, M.E. Zadeh, and A. Sadegh. A coarse relative-partitioned index theorem.Bull. Sci. Math., 153:57–71, 2019

2019
[17]

Lee.Introduction to Smooth Manifolds, volume 218 ofGraduate Texts in Mathematics

J.M. Lee.Introduction to Smooth Manifolds, volume 218 ofGraduate Texts in Mathematics. Springer New York, NY, second edition, 2012

2012
[18]

Ludewig and G.C

M. Ludewig and G.C. Thiang. Large-scale quantization of trace I: Finite propagation opera- tors.arXiv preprint arXiv:2506.10957, 2025

2025 arXiv
[19]

Paschke.K-theory for commutants in the Calkin algebra.Pacific J

W.L. Paschke.K-theory for commutants in the Calkin algebra.Pacific J. Math., 95(2):427– 434, 1981

1981
[20]

Pflaum, H

M.J. Pflaum, H. Posthuma, and X. Tang. The localized longitudinal index theorem for Lie groupoids and the van Est map.Adv. Math., 270:223–262, 2015

2015
[21]

Pflaum, H

M.J. Pflaum, H. Posthuma, and X. Tang. The transverse index theorem for proper cocompact actions of Lie groupoids.J. Differential Geom., 99(3):443–472, 2015. A PARTITIONED INDEX THEOREM FOR NONCOMPACT HYPERSURF ACES 63

2015
[22]

J. Roe. Partitioning noncompact manifolds and the dual Toeplitz problem. InOperator al- gebras and applications, Vol. 1, volume 135 ofLondon Math. Soc. Lecture Note Ser., pages 187–228. Cambridge Univ. Press, Cambridge, 1988

1988
[23]

Roe.Index theory, coarse geometry, and topology of manifolds, volume 90 ofCBMS Re- gional Conference Series in Mathematics

J. Roe.Index theory, coarse geometry, and topology of manifolds, volume 90 ofCBMS Re- gional Conference Series in Mathematics. Conference Board of the Mathematical Sciences, Washington, DC; by the American Mathematical Society, Providence, RI, 1996

1996
[24]

J. Roe. Comparing analytic assembly maps.Q. J. Math., 53(2):241–248, 2002

2002
[25]

J. Roe. Positive curvature, partial vanishing theorems and coarse indices.Proc. Edinb. Math. Soc. (2), 59(1):223–233, 2016

2016
[26]

Rørdam, F

M. Rørdam, F. Larsen, and N. Laustsen.An introduction toK-theory forC ∗-algebras, vol- ume 49 ofLondon Mathematical Society Student Texts. Cambridge University Press, Cam- bridge, 2000

2000
[27]

Schick and M.E

T. Schick and M.E. Zadeh. Large scale index of multi-partitioned manifolds.J. Noncommut. Geom., 12(2):439–456, 2018

2018
[28]

T. Seto. Toeplitz operators and the Roe-Higson type index theorem.J. Noncommut. Geom., 12(2):637–670, 2018

2018
[29]

Willett and G

R. Willett and G. Yu.Higher index theory, volume 189 ofCambridge Studies in Advanced Mathematics. Cambridge University Press, Cambridge, 2020

2020
[30]

C. Wulff. Bordism invariance of the coarse index.Proc. Amer. Math. Soc., 140(8):2693–2697, 2012

2012
[31]

C. Wulff. Coarse indices of twisted operators.J. Topol. Anal., 11(4):823–873, 2019

2019
[32]

G. Yu. The Novikov conjecture for groups with finite asymptotic dimension.Ann. of Math. (2), 147(2):325–355, 1998

1998
[33]

G. Yu. The coarse Baum-Connes conjecture for spaces which admit a uniform embedding into Hilbert space.Invent. Math., 139(1):201–240, 2000

2000
[34]

M.E. Zadeh. Index theory and partitioning by enlargeable hypersurfaces.J. Noncommut. Geom., 4(3):459–473, 2010. Institute for Mathematics, Astrophysics and Particle Physics, Radboud University, p.hochs@math.ru.nl Institute for Mathematics, Astrophysics and Particle Physics, Ra...

2010

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [12]

Higson and J

N. Higson and J. Roe.AnalyticK-homology. Oxford Mathematical Monographs. Oxford University Press, Oxford, 2000. Oxford Science Publications

work page 2000

[2] [1]

Atiyah and I.M

M.F. Atiyah and I.M. Singer. The index of elliptic operators. I.Ann. of Math. (2), 87:484–530, 1968

work page 1968

[3] [2]

Atiyah and I.M

M.F. Atiyah and I.M. Singer. The index of elliptic operators. III.Ann. of Math. (2), 87:546– 604, 1968

work page 1968

[4] [3]

P. Baum, A. Connes, and N. Higson. Classifying space for proper actions andK-theory of groupC ∗-algebras. InC ∗-algebras: 1943–1993 (San Antonio, TX, 1993), volume 167 of Contemp. Math., pages 240–291. Amer. Math. Soc., Providence, RI, 1994

work page 1943

[5] [4]

Berline, E

N. Berline, E. Getzler, and M. Vergne.Heat kernels and Dirac operators, volume 298 of Grundlehren der mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences]. Springer-Verlag, Berlin, 1992

work page 1992

[6] [5]

Coronas and Callias type operators in coarse geometry

U. Bunke and M. Ludewig. Coronas and Callias type operators in coarse geometry.arXiv preprint arXiv:2411.01646, 2024

work page Pith review arXiv 2024

[7] [6]

Connes and H

A. Connes and H. Moscovici. Cyclic cohomology, the Novikov conjecture and hyperbolic groups.Topology, 29(3):345–388, 1990

work page 1990

[8] [7]

Conway.A course in functional analysis, volume 96 ofGraduate Texts in Mathematics

J.B. Conway.A course in functional analysis, volume 96 ofGraduate Texts in Mathematics. Springer-Verlag, New York, second edition, 1990

work page 1990