betaFeedback

WorksMathematics

Positive scalar curvature forces rational inessentiality

Abstract

We prove that every closed connected oriented smooth manifold admitting strictly positive scalar curvature is rationally inessential: its rational fundamental class maps to zero under the classifying map. No spin or fundamental-group hypothesis is needed, and there is no upper dimension bound. This proves the Gromov–Lawson aspherical conjecture; moreover, every nonnegative-scalar-curvature metric on a closed aspherical manifold is flat. We also show that every closed oriented smooth manifold of positive dimension with nonnegative scalar curvature has zero real simplicial volume, proving the qualitative vanishing consequence of Gromov's conjectural comparison.

Introduction

Let MnM^n be a closed connected oriented smooth manifold, let Γ=π1(M)\Gamma= \pi_1(M), and let cM:M→BΓc_M : M \to B\Gamma be a classifying map for its universal cover. Here BΓB\Gamma is a CW K(Γ,1)K(\Gamma,1): a connected CW complex with fundamental group Γ\Gamma and contractible universal cover. The manifold is rationally essential if

(cM)∗[M]≠0in Hn(BΓ;Q),(c_M)_*[M] \ne0 \quad\text{in } H_n(B\Gamma;\mathbb{Q}),

and rationally inessential otherwise. The fundamental class in this definition has rational coefficients. The condition is independent of the choice of classifying map. It records whether the fundamental class remains visible after passing from the manifold to the topology of its fundamental group.

We prove the rational-essentiality obstruction to positive scalar curvature formulated in [18], Conjecture 2.9.

Theorem 1.1 (Rational inessentiality). Let MnM^n be a closed connected oriented smooth manifold of dimension n≥2n \ge2. If MM admits a Riemannian metric of strictly positive scalar curvature, then

(cM)∗[M]=0in Hn(Bπ1(M);Q).(c_M)_*[M] = 0 \quad\text{in } H_n(B\pi_1(M);\mathbb{Q}).

The theorem has no spin, universal-cover spin, or fundamental-group hypothesis, and no upper dimension bound. In particular, the group may have torsion and its classifying space need not have a finite-dimensional model.

A manifold is aspherical if its universal cover is contractible.

From a detected class to contracting maps

In the later geometric convention, a closed oriented Riemannian nn-manifold is enlargeable if, for every ε>0\varepsilon> 0, some cover with the lifted metric admits an ε\varepsilon-Lipschitz map of nonzero degree to the unit sphere SnS^n that is constant outside a compact set. This definition imposes no spin condition. Rational essentiality supplies a group-cohomology class whose pullback pairs nontrivially with [M][M], but it does not guarantee these contracting maps. Brunnbauer and Hanke construct rationally essential manifolds which are not enlargeable, as well as enlargeable manifolds whose universal covers alone do not admit such maps at every scale, that is, are not hyperspherical [4], Theorems 1.4–1.5. Their examples are not aspherical.

Guth and Braun–Sauer give a different comparison between topology and metric largeness: at every radius, they guarantee a ball in the universal cover whose volume is at least a dimensional multiple of the radius to the nnth power, for closed aspherical and rationally essential manifolds, respectively [16, 3]. These bounds do not have the sharp Euclidean leading constant, so they do not determine the scalar-curvature term in the small-ball volume expansion.

Our proof converts the cohomology pairing into a local degree by adding a parameter manifold, then changes the metric to contract the resulting map while retaining positive scalar curvature. The main estimate makes the scalar cost grow only logarithmically with the number of comparison coordinates. Suppose for contradiction that Scal⁡M≥κ>0\operatorname{Scal}_M \ge\kappa> 0 and that MM is rationally essential. The construction has four stages.

Finite comparison and local degree. At each sufficiently large finite word-metric comparison scale RR, an auxiliary closed oriented manifold CC with a principal left Γ\Gamma-cover C^→C\widehat{C} \to C gives the cover X=(M~×C^)/Γ→M×CX = (\widetilde{M} \times\widehat{C})/\Gamma\to M \times C. The detecting cocycle yields a compactly supported cohomology class on XX that evaluates nontrivially on its locally finite fundamental class. Distance profiles supply a preliminary map V:X→Rq\mathcal{V}: X \to\mathbb{R}^{q}, for a finite integer q>nq > n. Together with the parameter projection oˉ:X→C\bar{o}: X \to C, it defines the map x↦(oˉ(x),V(x))x \mapsto(\bar{o}(x), \mathcal{V}(x)) into C×RqC \times\mathbb{R}^{q}.

The construction transfers the compact pairing to a relative cohomology class on a compact region of this product, evaluated nontrivially on a relative cycle carried by the preceding map. Relative duality gives a complementary relative rational homology class, and rational realization represents a nonzero multiple by a compact oriented (q−n)(q-n)-manifold with boundary mapped to C×RqC \times\mathbb{R}^{q}. Its transverse intersections with the preceding map have nonzero signed count. Doubling the parameter manifold and reflecting its map gives a closed oriented manifold AA with a map (o0,y0):A→C×Rq(o_{0}, y_{0}): A \to C \times\mathbb{R}^{q}. Pulling back C^→C\widehat{C} \to C along o0o_{0} gives a principal cover A^→A\widehat{A} \to A and the fibre product

Z=X×CA=(M~×A^)/Γ⟶M×A.Z = X \times_{C} A = (\widetilde{M} \times\widehat{A})/\Gamma\longrightarrow M \times A.

This cover has dimension q=n+dim⁡Aq = n + \dim A, and both V\mathcal{V} and y0y_{0} pull back to Rq\mathbb{R}^{q}-valued maps on it. On a precompact open set Ω\Omega over one half of AA, the zeros of their scaled difference (V−y0)/R(\mathcal{V} - y_{0})/R are precisely the preceding intersections. A cutoff equal to one on Ω\Omega makes this difference a smooth compactly supported map G:Z→RqG: Z \to\mathbb{R}^{q} satisfying

deg⁡(G,Ω,0)≠0,∥G∥∞>3on ∂Ω,\deg(G,\Omega,0) \ne0,\qquad\lVert G\rVert_{\infty} > 3 \quad\text{on } \partial\Omega,

where ∥⋅∥∞\lVert\cdot\rVert_{\infty} is the maximum of the absolute target coordinates. The scalar coordinates satisfy

max⁡i∣dM~Gi∣≤J/R,log⁡q≤JR,(1.1)\max_{i}\lvert d_{\widetilde{M}}G_{i}\rvert\le J/R,\qquad\log q \le JR, \tag*{(1.1)}

with JJ independent of RR. Here dM~d_{\widetilde{M}} differentiates in the fibre directions, holding the parameter in AA fixed, and its norm uses the lifted metric of MM. Parameter derivatives are bounded at each fixed RR and are controlled later by metric enlargement. Section 3 constructs these data without assuming that M~\widetilde{M} is contractible or that the cocycle is bounded.

A graph and one warped circle per coordinate. A graph change increases the fibre metric in the direction of one function’s differential, and a reciprocal warped circle supplies favorable scalar terms. A nonlinear elliptic equation chooses a replacement coordinate close to the old one while shortening its differential. Processing all qq coordinates once is called a pass. Paying an independent worst-case scalar cost for every coordinate would introduce a factor qq and destroy the scalar lower bound. Instead, the decreases of the inverse metric telescope on the nn-dimensional fibre cotangent space. Their sum has trace at most nn in the metric at the start of the pass, independently of the numbers of coordinates and circles. The scalar identity in Section 4, the graph solution in Section 5, and the summation in Section 6 bound the fibre scalar-curvature loss by C(n+1)(2+log⁡q)L2/w2C(n+1)(2+\log q)L^{2}/w^{2}, for an absolute constant CC, when the input row norms are at most LL and each replacement is allowed coordinate error πw/2\pi w/2.

Finite shortening with a summable cost. Successive passes divide LL by four and ww by two, so their coordinate errors and fibre scalar losses form summable series. The initial error allowance makes the uniform total error in each coordinate less than π/10\pi/10, preserving the maximum-norm boundary gap. Composing the shortened coordinate map on Ω\Omega with a degree-one map Rq→Sq\mathbb{R}^{q} \to S^{q} that is constant outside the unit ball, and extending constantly outside Ω\Omega, preserves this degree. By (1.1), the total fibre scalar loss is O((2+log⁡q)R−2)=O(R−1)O((2+\log q)R^{-2}) = O(R^{-1}). At each fixed finite stage, enlarging the metric on AA makes the parameter part of the differential small and makes the assembled scalar curvature uniformly close to the fibre scalar curvature. For RR sufficiently large and every finite integer k≥1k \ge1, there are a metric gkg_k on ZZ, positive warps mk,νm_{k,\nu}, and a smooth map Fk:Z→SqF_k: Z \to S^q, constant outside a compact set and of nonzero degree, such that

Scal⁡gk+∑ν=1kqmk,ν2dθν2≥κ/2,∣dFk∣gk,op⁡≤C(q)4−kJ/R.\operatorname{Scal}_{g_k+\sum_{\nu=1}^{kq}m_{k,\nu}^{2}\mathrm{d}\theta_{\nu}^{2}} \ge\kappa/2,\qquad\lvert\mathrm{d}F_k\rvert_{g_k,\operatorname{op}} \le C(q)4^{-k}J/R.

The scalar bound concerns Z×TkqZ \times\mathbb{T}^{kq}, while the differential bound uses all the base directions in ZZ. Figure 1 distinguishes these directions from the stabilizing circles.

Local direction splitting after $k$ finite passes

Figure 1. The local direction splitting after kk finite passes. The base ZZ is locally a product of original and parameter directions; no global product decomposition is asserted. The inverse-metric trace uses only the nn original directions. The sphere map uses the full base of dimension qq, which is fixed before the number kqkq of circles is chosen.

An obstruction independent of the circle count. Pulling back a torus tube in SqS^q gives a compact band W⊂ZW \subset Z carrying the degree. Doubling WW and retaining the circle factors gives a closed oriented manifold dbl⁡(W)×Tkq\operatorname{dbl}(W) \times\mathbb{T}^{kq} mapped with nonzero degree to S1×Tq−1×TkqS^1 \times\mathbb{T}^{q-1} \times\mathbb{T}^{kq}. The local torical theorem of Cecchini–Schick requires a scalar lower bound and a bound on the full differential only on the full preimage of one connected open target arc. The doubled metric therefore needs to agree with the constructed metric only there. Constant enlargement of the stabilizing circles controls their joint projection without scalar loss. Section 7 derives an obstruction threshold depending on qq and κ\kappa, but not on the number kqkq of circles. We first choose RR large enough to control the total fibre scalar loss, thereby fixing qq and all comparison data, and then choose a sufficiently large finite kk to contradict this obstruction.

Consequences

For a connected aspherical manifold, the classifying map is a homotopy equivalence. Thus every closed connected oriented aspherical manifold is rationally essential. The main theorem gives the following consequence, with the nonorientable case obtained from the orientation double cover.

Corollary 1.2 (The Gromov–Lawson aspherical conjecture). No closed smooth aspherical manifold, whether orientable or not, admits a Riemannian metric of strictly positive scalar curvature.

For nonnegative scalar curvature, the corresponding statement determines the given metric.

Corollary 1.3 (Aspherical nonnegative-scalar-curvature rigidity). Let gg be a Riemannian metric of nonnegative scalar curvature on a closed smooth aspherical manifold MM, not necessarily oriented. Then gg is flat.

Both aspherical statements apply componentwise and have no spin assumption. In dimensions zero and one scalar curvature vanishes identically, and every metric is flat. Gromov already noted the rigidity implication conditionally on the aspherical positive-scalar-curvature obstruction [13], Section 3, following Conjecture B. Here the proof combines Theorem 2.2 with the compact-manifold consequence of the Cheeger–Gromoll splitting theorem.

A nonzero-degree map to an aspherical target forces rational essentiality by naturality of classifying maps; no injectivity or surjectivity condition on the induced map of fundamental groups is needed. The following is the oriented form of the target-map conjecture recorded by Chodosh and Li [8], Conjecture 18. Their survey also states the connected-sum obstruction in dimensions four and five as a consequence of Gromov’s work [8], Theorem 17.

Corollary 1.4 (Oriented target-map obstruction). Let MnM^n and NnN^n be closed connected oriented smooth manifolds of dimension n≥2n \ge2, with NN aspherical. If a continuous map f:M→Nf : M \to N has nonzero ordinary oriented degree d∈Z∖{0}d \in\mathbb{Z} \setminus\{0\}, then MM admits no Riemannian metric whose scalar curvature is strictly positive everywhere.

In particular, if PnP^n is any closed connected oriented smooth manifold, the oriented connected sum N#PN\#P admits no such metric.

For a closed connected oriented nn-manifold, its real simplicial volume is

∥M∥=inf⁡{∑i∣ai∣:∑iaiσi is a real singular cycle representing [M]}.\lVert M \rVert= \inf\left\{ \sum_i \lvert a_i \rvert: \sum_i a_i \sigma_i \text{ is a real singular cycle representing } [M] \right\}.

Positive simplicial volume implies rational essentiality, by the bounded-cohomology mapping theorem and duality for this seminorm (Theorem 2.1). Therefore rational inessentiality implies zero simplicial volume. The converse fails: a positive-dimensional torus is rationally essential but has zero simplicial volume, as follows from its self-covers of degree greater than one. Its flat metric also shows why strict positivity in Theorem 1.1 cannot be replaced by nonnegative scalar curvature. For simplicial volume the nonnegative hypothesis does suffice.

Corollary 1.5 (Simplicial-volume vanishing). Every closed oriented smooth manifold of positive dimension admitting a Riemannian metric of nonnegative scalar curvature has zero real simplicial volume.

On each connected component, Section 2 reduces a nonnegative-scalar-curvature metric to either a positive-scalar-curvature metric or a Ricci-flat metric. The positive branch uses Theorem 1.1; the Ricci-flat branch uses volume comparison and amenability, without requiring rational inessentiality in that case.

The obstruction problem and earlier work

Two approaches have shaped scalar-curvature obstructions. Schoen and Yau used stable two-sided minimal hypersurfaces and conformal deformation to transfer positive scalar curvature to lower dimensions [25, 24]. Their original obstructions requiring no spin assumption apply in dimensions at most seven, including to manifolds mapping with nonzero degree to tori. Gromov and Lawson developed Dirac-operator obstructions using enlargeability and relative indices on complete manifolds [12]. Their 1983 enlargeability convention requires oriented spin coverings, and their compact obstruction has no upper dimension bound. The later geometric convention used above omits this spin requirement. The spin condition in geometric index applications and the regularity of minimizing hypersurfaces are thus different constraints.

More recently, Chodosh and Li established the aspherical obstruction in dimensions four and five using generalized soap bubbles [9]. Gromov independently treated the closed five-dimensional case, and also complete-manifold domination problems under properness and uniform-positivity or decay hypotheses [14]. Theorem 1.2 establishes the closed-aspherical statement in every dimension. The homological formulation in Theorem 1.1 concerns more manifolds than the aspherical class. Hanke explicitly formulates the conjecture that rationally essential manifolds do not admit positive scalar curvature [18]. It is also the classifying-map form, for closed oriented sources, of Gromov’s conjecture that a nontorsion homology class in the classifying space of a countable group cannot be the image of the fundamental class of a positive-scalar-curvature manifold [15]. The numbered sections cited from that work refer throughout to the author’s arXiv version 6. Yau also posed this homological obstruction in his survey [32].

Gromov introduced simplicial volume and its bounded-cohomology description in [11]. His quantitative scalar-curvature/simplicial-volume conjecture asks for a dimension-dependent upper bound on simplicial volume in terms of Riemannian volume under a fixed scalar-curvature lower bound [15]. Rescaling a nonnegative-scalar-curvature metric in such a comparison would force zero simplicial volume. Theorem 1.5 proves this qualitative vanishing consequence. The quantitative comparison remains a separate question.

Recent work of Ma, Wang, Xie, Yu and Zhu proves that nonnegative scalar curvature forces zero real simplicial volume for closed oriented manifolds with spin universal cover [20]. Their stated version has no further group-decay assumption. This conclusion controls the simplicial norm, rather than the rational classifying image. Bi and Zhu develop singular hypersurface descent and rule out positive scalar curvature on closed oriented enlargeable and overtorical manifolds, where an nn-manifold is overtorical if it admits a nonzero-degree map to Tn\mathbb{T}^{n} [1]. Their theorem for complete oriented sources and complete oriented enlargeable targets retains properness, nonzero-degree and Lipschitz hypotheses and rules out a uniform positive scalar lower bound. Under strict positive scalar curvature, Theorem 1.1 gives rational classifying-image vanishing without these geometric-largeness or spin-cover hypotheses.

The terminal geometric input here is the local torical theorem of Cecchini and Schick [6]. It has no spin assumption or upper dimension bound, and requires both the scalar lower bound and the full differential bound only on the full preimage of one connected open target arc. Its higher-dimensional proof uses the minimal-slicing theory of Schoen and Yau [27].

Section 2 records the degree conventions and the nonnegative scalar-curvature reduction. Section 3 constructs the comparison data, and Sections 4 to 6 establish the graph stabilization and shortening estimates. Section 7 supplies the stabilized obstruction. Finally, Section 8 combines these results and proves the four corollaries.

Conventions and the nonnegative scalar-curvature reduction

We record the degree conventions for the comparison construction and reduce nonnegative scalar curvature to the positive and Ricci-flat cases used in the corollaries. All manifolds and metrics are smooth. A closed manifold is compact and has no boundary. Unless stated otherwise, manifolds are oriented. Fundamental classes are taken with the indicated coefficients; we use the same symbol [M][M] for an integral fundamental class and its images under extension of coefficients. The Laplacian is Δ=div⁡∇\Delta= \operatorname{div} \nabla, and a round sphere has positive scalar curvature. For a hypersurface with unit normal ν\nu, our second fundamental form is II⁡(X,Y)=⟨∇Xν,Y⟩\operatorname{II}(X,Y) = \langle\nabla_X \nu,Y\rangle and its mean curvature is its trace. These conventions fix the signs in the graph calculations below. We give SdS^{d} its unit round metric and write Tb=(R/2πZ)b\mathbb{T}^{b} = (\mathbb{R}/2\pi\mathbb{Z})^{b} for the standard flat torus.

Fundamental classes and degree

Throughout the proof, Γ=π1(M)\Gamma= \pi_1(M) and cM:M→BΓc_M : M \to B\Gamma have the meaning fixed in Section 1. No finite-dimensional model for BΓB\Gamma is assumed. Disconnected closed manifolds are treated one component at a time; compactness leaves only finitely many components.

For a possibly noncompact oriented dd-manifold YY, a smooth map F:Y→SdF : Y \to S^d that is equal to one value ∗∈Sd\ast\in S^d outside a compact set has a well-defined degree. One may compute it as the signed number of preimages of any regular value different from ∗\ast. Those preimages are contained in a compact set. Equivalently, pull back a compactly supported top-degree class on Sd∖{∗}S^d \setminus\{\ast\} and evaluate it on the locally finite fundamental class of YY. This definition also permits disconnected YY: only finitely many components can meet the compact region on which FF is nonconstant. We use constant outside a compact set in this sense, rather than assigning a linear meaning to compact support for a sphere-valued map.

We also use the ordinary oriented local degree deg⁡(G,Ω,0)\operatorname{deg}(G,\Omega,0) for a smooth map from an oriented qq-manifold to Rq\mathbb{R}^q, where Ω\Omega is a precompact open set and 0∉G(∂Ω)0 \notin G(\partial\Omega). Smoothness of ∂Ω\partial\Omega is unnecessary. A homotopy whose boundary values avoid zero preserves this degree. More generally, if they avoid a fixed ball, composition with a degree-one map Rq→Sq\mathbb{R}^q \to S^q constant outside that ball gives a sphere-valued map of the same degree, extending constantly across the boundary of Ω\Omega.

Lemma 2.1 (Simplicial volume detects rational essentiality). If a closed connected oriented manifold MnM^n has positive real simplicial volume, then (cM)∗[M]≠0(c_M)_\ast[M] \ne0 in Hn(BΓ;Q)H_n(B\Gamma;\mathbb{Q}).

Proof. The bounded-cohomology mapping theorem identifies real bounded cohomology of MM with that of BΓB\Gamma by pullback along cMc_M. The duality between bounded cohomology and the ℓ1\ell^1-seminorm implies that ∥M∥>0\lVert M\rVert> 0 supplies a bounded class whose ordinary cohomology image evaluates nontrivially on [M][M]. These are the mapping and duality results of [11]; neither requires a spin structure.

If (cM)∗[M](c_M)_\ast[M] vanished rationally, its real image would vanish. Every bounded class pulled back from BΓB\Gamma would then evaluate trivially on [M][M], a contradiction. In dimension one, real bounded cohomology already vanishes and the assertion is vacuous; the direct circle argument below also suffices.

The scalar-flat alternative

References

  1. [1]Yuchen Bi and Jintian Zhu. Positive scalar curvature obstructions via singular dimension descent, 2026. arXiv:2606.20528v3, 11 September 2026.arxiv.org/abs/2606.20528v3
  2. [2]R. L. Bishop, & B. O’Neill. (1969). Manifolds of negative curvature. https://doi.org/10.1090/S0002-9947-1969-0251664-4
  3. [3]Sabine Braun and Roman Sauer. Volume and macroscopic scalar curvature. Geometric and Functional Analysis, 31:1321–1376, 2021. DOI.DOI
  4. [4]Michael Brunnbauer and Bernhard Hanke. Large and small group homology. Journal of Topology, 3(2):463–486, 2010. Theorem numbers refer to arXiv:0902.0869v2.DOI
  5. [5]Stewart S. Cairns. On the triangulation of regular loci. Annals of Mathematics, 35(3):579–587, 1934. DOI.DOI
  6. [6]Simone Cecchini and Thomas Schick. Enlargeable metrics on nonspin manifolds. Proceedings of the American Mathematical Society, 149(5):2199–2211, 2021. DOI; author’s version: arXiv:1810.02116v4.DOI
  7. [7]Jeff Cheeger and Detlef Gromoll. The splitting theorem for manifolds of nonnegative Ricci curvature. Journal of Differential Geometry, 6(1):119–128, 1971. DOI.DOI
  8. [8]Otis Chodosh, & Chao Li. (2023). Recent results concerning topological obstructions to positive scalar curvature. https://doi.org/10.1142/9789811273230_0006
  9. [9]Otis Chodosh, & Chao Li. (2024). Generalized soap bubbles and the topology of manifolds with positive scalar curvature.
  10. [10]David Gilbarg, & Neil S. Trudinger. (2001). Elliptic Partial Differential Equations of Second Order. https://doi.org/10.1007/978-3-642-61798-0
  11. [11]Mikhael Gromov. Volume and bounded cohomology. Publications Mathématiques de l’IHÉS, 56:5–99, 1982. DOI.DOI
  12. [12]Mikhael Gromov and H. Blaine Lawson, Jr. Positive scalar curvature and the Dirac operator on complete Riemannian manifolds. Publications Mathématiques de l’IHÉS, 58:83–196, 1983. DOI.DOI
  13. [13]Misha Gromov. A dozen problems, questions and conjectures about positive scalar curvature. In Joseph Kouneiher, editor, Foundations of Mathematics and Physics One Century After Hilbert, pages 135–158. Springer, Cham, 2018. Section locators refer to the author’s version of 19 September 2017. DOI; author’s PDF.DOI
  14. [14]Misha Gromov. No metrics with positive scalar curvatures on aspherical 5-manifolds, 2020. Version 1, 11 September 2020; arXiv:2009.05332v1.arxiv.org/abs/2009.05332v1
  15. [15]Misha Gromov. (2023). Four lectures on scalar curvature. https://doi.org/10.1142/9789811273223_0001
  16. [16]Larry Guth. (2011). Volumes of balls in large Riemannian manifolds. https://doi.org/10.4007/annals.2011.173.1.2
  17. [17]Richard S. Hamilton. Three-manifolds with positive Ricci curvature. Journal of Differential Geometry, 17(2):255–306, 1982. DOI.DOI
  18. [18]Bernhard Hanke. Positive scalar curvature, K-area and essentialness. In Christian Bär, Joachim Lohkamp, and Matthias Schwarz, editors, Global Differential Geometry, volume 17 of Springer Proceedings in Mathematics, pages 275–302. Springer, Berlin, 2012. Definition and conjecture numbers refer to arXiv:1011.3987v2. DOI.DOI
  19. [19]Allen Hatcher. (2002). Algebraic Topology. https://pi.math.cornell.edu/~hatcher/AT/AT.pdfpi.math.cornell.edu/~hatcher/AT/AT.pdf
  20. [20]Qiaochu Ma, Jinmin Wang, Zhizhang Xie, Guoliang Yu, and Bo Zhu. Gromov’s simplicial volume vanishing conjecture for positive scalar curvature, 2026. arXiv:2606.29135v4, 28 August 2026.arxiv.org/abs/2606.29135v4
  21. [21]John Milnor. A note on curvature and fundamental group. Journal of Differential Geometry, 2:1–7, 1968. DOI.DOI
  22. [22]James R. Munkres. (1966). Elementary Differential Topology.
  23. [23]Arthur Sard. The measure of the critical values of differentiable maps. Bulletin of the American Mathematical Society, 48(12):883–890, 1942. DOI.DOI
  24. [24]Richard Schoen, & Shing-Tung Yau. (1979). Existence of incompressible minimal surfaces and the topology of three dimensional manifolds with non-negative scalar curvature. https://doi.org/10.2307/1971247
  25. [25]Richard Schoen, & Shing-Tung Yau. (1979). On the structure of manifolds with positive scalar curvature. https://doi.org/10.1007/BF01647970
  26. [26]Richard Schoen and Shing-Tung Yau. Proof of the positive mass theorem. II. Communications in Mathematical Physics, 79:231–260, 1981. Journal article.DOI
  27. [27]Richard Schoen and Shing-Tung Yau. Positive scalar curvature and minimal hypersurface singularities, 2017. arXiv:1704.05490v1.arxiv.org/abs/1704.05490v1
  28. [28]René Thom. Quelques propriétés globales des variétés différentiables. Commentarii Mathematici Helvetica, 28:17–86, 1954. DOI.DOI
  29. [29]Peter Topping. Lectures on the Ricci Flow, volume 325 of London Mathematical Society Lecture Note Series. Cambridge University Press, Cambridge, 2006. Theorem locators refer to the author’s notes of 9 March 2006. DOI.DOI
  30. [30]Guofang Wei. (2007). Manifolds with a lower Ricci curvature bound. https://doi.org/10.4310/SDG.2006.v11.n1.a7
  31. [31]J. H. C. Whitehead. (1940). On C¹-complexes. https://doi.org/10.2307/1968861
  32. [32]Shing-Tung Yau. Review of geometry and analysis. Asian Journal of Mathematics, 4(1):235–278, 2000. DOI.DOI

Paper details

vlnc:W_jy5mgbhf
Version
v1 · latest
Published
License
LicenseRef-OpenAI-public
Usage

views · 10

downloads · 2

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

unreleased internal OpenAI model

View statement
Looking for related papers…
Contents