The metric from energy-momentum non-conservation: Generalizing Noether and completing spectral geometry

Achim Kempf

REVIEW 4 major objections 5 minor 46 references

The metric from energy-momentum non-conservation: Generalizing Noether and completing spectral geometry

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Nonlinear resonances reveal the shape of spacetime, because interactions supply the position basis that the spectrum alone cannot.

desk verdict The paper's central step—identifying a position basis by the singularity structure of G(n>2)—is asserted, not proven, and as stated it conflicts with standard OPE and lightcone behavior in QFT. read the letter →

arxiv 2506.11186 v1 pith:3OOC3247 submitted 2025-06-12 gr-qc math-phmath.MPquant-ph

Achim Kempf This is my paper · ORCID

classification gr-qcmath-phmath.MPquant-ph

keywords spectralgeometrymetricreconstructionfromcorrelatorsnonconservationofenergy-momentumgeneralizedNoethertheorempositionbasislocalityabstractquantumfieldtheoryoncurvedspacetimeemergence

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

The paper aims to show that the full metric of a spacetime—or the shape of any vibrating manifold—can be recovered from the quantum or classical correlators of a field living on it, provided one uses not only the two-point function's spectrum but also the higher-order, nonlinear correlators. The missing ingredient in classical spectral geometry is a position basis; the paper argues that interactions supply it, because a local interaction vertex must be singular exactly when all its arguments coincide, and only a position basis has that property. Once that basis is found, the two-point function expressed in it determines the metric through a known short-distance formula. The result is framed as a completion of spectral geometry and as a generalization of Noether's theorem: on a curved spacetime, the specific pattern of energy-momentum non-conservation encodes the metric that causes it.

What carries the argument

The load-bearing object is the abstract $n$-point correlator $G^{(n)}$, regarded as a multilinear operator on tensor products of a Hilbert space, together with the locality criterion: a basis is a position basis exactly when $G^{(n>2)}$ is singular if and only if all its arguments coincide. At tree level this reduces to genuine diagonalization of the vertex; beyond tree level it means locating the coincidence singularity of the renormalized vertex. The argument then uses this criterion to fix the unitary $U$ from the eigenbasis of $G^{(2)}$ to a position basis, transforms $G^{(2)}$ with that same $U$, and feeds the resulting propagator's short-distance behaviour into equation (2) to obtain $g_{\mu\nu}(x)$.

What would settle it

Take two compact Riemannian manifolds that are isospectral but not isometric, place a $\lambda\phi^4$ scalar field on each, and compute the connected 4-point function in the common eigenbasis of the Laplacian. If the matrix elements of $G^{(4)}$ coincide for the two manifolds and yet the locality-criterion unitaries lead to different reconstructed metrics, the claim is false; if the matrix elements already differ, the spectrum-plus-vertex data do distinguish the geometries, supporting the claim.

Watch

Extended reading notes

Core claim

The central claim is that for a local interacting quantum field theory, the set of abstract $n$-point correlators $G^{(n)}$, given in an arbitrary basis such as the eigenbasis of the wave operator, determines the spacetime metric. The two-point correlator $G^{(2)}$ carries the metric only when written in a position basis, and its spectrum alone is invariant under the full unitary group, so it cannot fix that basis. The paper's proposal is to use a higher-order correlator $G^{(n>2)}$ as the basis-finder: a unitary transformation $U$ that brings the vertex to a form that is singular if and only if all arguments coincide defines a position basis. Applying the same $U$ to $G^{(2)}$ yields the position-space propagator, from which the metric $g_{\mu\nu}(x)$ follows via the Hadamard-based reconstruction formula (2). In this sense the geometry is fully encoded in the abstract correlators, and nonlinearities are not a nuisance but the key to spectral geometry.

Load-bearing premise

The reconstruction relies on the premise that in a position basis the interaction vertex $G^{(n>2)}$ is singular exactly when all its arguments coincide, and that this property picks out the position basis uniquely; if other singularities, such as partial-coincidence or light-cone singularities, also appear, the criterion may not identify a unique basis and the procedure has no well-defined output.

Editorial extensions

If this is right

The metric of a spacetime can be computed from the spectrum of the wave operator plus the matrix elements of an interaction vertex, with no coordinate system assumed in advance.
Spectral geometry is thereby completed for interacting systems: nonlinear resonances of a body contain the position information that the linear spectrum lacks.
On a generic curved spacetime, the detailed failure of energy-momentum conservation—encoded in the interaction vertex in the eigenbasis—becomes, in principle, a measurement of the metric.
If a set of abstract correlators admits a basis with the locality singularity, spacetime and matter emerge as a derived, approximate representation; if no such basis exists, as the paper suggests may happen near the Planck scale, the spacetime picture breaks down.

Reading between the lines

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

A testable consequence the paper leaves implicit: in any concrete model, one could scan over unitaries numerically to find bases with the coincidence-only singularity, and check whether the resulting metric is independent of the chosen unitary; this would sharpen the claim that the position basis is unique up to diffeomorphism.
The criterion may need refinement in realistic theories, since renormalized correlators are singular not only at full coincidence but also at partial coincidences, and on light cones; a refined fingerprint might be needed to make the search for $U$ well-posed.
If the reconstruction works, the same logic could be applied to laboratory vibrating systems: measuring mode-coupling amplitudes in the nonlinear regime would yield the shape of an object without any prior knowledge of its geometry, connecting the argument to inverse problems in acoustics.
The paper's Planck-scale speculation suggests a sharp target: find an abstract correlator set that provably admits no local representation, and characterize the generic condition for non-representability in terms of spectral statistics.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Here's my take. The paper's central step—using the singularity structure of the interaction correlator G(n>2) to identify a position basis—is asserted, not derived, and as stated it does not match how correlators actually behave in local QFT. That is the load-bearing gap. But the proposal itself is genuinely interesting: add nonlinear data to the spectrum, use locality of interactions to fix the missing unitary, then read the metric off the position-space propagator. The paper is clearly written and honest that it is a program; it points to earlier work for the metric formula (2) and does not claim the functional analysis is done. The generalized Noether framing is an interpretive gloss, not a theorem, but it communicates the intuition well.

The soft spot is in Section 4. The criterion that a position basis is one where the renormalized connected correlator G(n>2) is singular if and only if all arguments coincide is stated without proof. In ordinary interacting QFT, operator product expansions produce singularities at partial coincidences, and massless fields produce lightcone singularities at null separation. So the proposed fingerprint does not single out a unique basis even when a position basis exists. The search over U is also not well-posed unless the abstract labels carry a topology or distributional structure; the paper provides none. The paper itself hedges: it says existence of U is contingent on representability as a local QFT, and admits that at higher orders the correlator is finite away from coincidence—but that undercuts the 'exactly at coincidence' test used to define the position basis.

So the abstract's claim of 'completing spectral geometry' is stronger than what is shown. The paper is a program, not a derivation, and it is advertised as more. There is no worked example, no explicit U construction even in a toy model, and no discussion of how the abstract eigenbasis labels acquire points on a manifold. Still, the core question—can interactions provide the missing basis information?—deserves an airing. It is the kind of idea that could either fizzle or open a real route, and the paper gives enough structure for a referee to engage with the central conjecture. I would send it to peer review with the expectation of major revision, not because it is obviously dead, but because the reconstruction criterion needs proof in a restricted setting or replacement with something that respects renormalization and OPE.

For a reading group: maybe, if it enjoys speculative foundations. I would not cite it in my own work yet, but I would keep an eye on follow-ups.

Referee Report

4 major / 5 minor

Summary. The paper proposes that the metric of a spacetime is fully reconstructible from the spectrum of the two-point correlator G(2) together with the matrix elements, in that eigenbasis, of a higher connected correlator G(n>2). The reconstruction proceeds by finding a unitary U that transforms G(n>2) into a form singular exactly when all arguments coincide; U is then applied to G(2), and Eq. (2) yields g_{\mu\nu}(x). The paper presents this as completing spectral geometry, as a generalized Noether theorem relating the pattern of energy-momentum non-conservation to the metric, and as evidence that spacetime and matter can emerge from abstract correlators.

Significance. The question whether nonlinear or interaction data can remove the unitary ambiguity left by the spectrum is a natural and potentially important one, and the paper states a concrete data set and a clear four-step procedure. Eq. (2) is grounded in earlier work, and the connection to spectral geometry is appropriately framed. However, the central step—identifying a position basis from singularity structure—is asserted, not proved, and the stated criterion is both ill-defined for abstract bases and false for renormalized correlators in standard local QFT. The manuscript therefore does not establish its principal claims; at present it is a speculative proposal whose key mechanism is unsupported.

major comments (4)

[Sec. 4, Step 2] The criterion for a position basis is not well-defined in the abstract setting. Starting from the eigenbasis of G(2), the labels |k> are merely eigenvalues; there is no manifold, coordinate system, or topology with respect to which the phrase 'coincidence of its arguments' has meaning. A singularity of a kernel is defined relative to such a structure, and none is given for the abstract eigenbasis. Consequently, the instruction to 'find a unitary U that reveals the local nature of the interaction vertex' is not a well-posed inverse problem, and Steps 3 and 4 inherit this gap.
[Sec. 4, after Eq. (3)] The claim that renormalized G(n>2) 'remains singular if and only if all arguments coincide' is false for standard local QFT. Operator product expansions produce singularities at partial coincidences (e.g., x1 approaching x2 while x3 and x4 remain separated), and massless fields produce lightcone singularities at null separations. Thus the singularity set is not a single coincidence point, and the proposed fingerprint does not select a unique position basis. Moreover, the tree-level delta-function form of Eq. (3) does not survive renormalization, yet the text states that the procedure covers weak and strong nonlinearity.
[Sec. 4, Step 2] No proof or construction is given that such a unitary U exists for any nontrivial abstract data. The statement that existence is 'contingent on the underlying abstract interaction being representable as local' restates the desired conclusion rather than establishing it. Without a characterization of diagonalizability in the sense of the singularity criterion, the central conditional claim—'if abstract higher-order correlators G(n>2) can be diagonalized, these correlators can be represented as local quantum field theoretic vertices on a curved spacetime'—is unsupported. The paper also provides no explicit worked example, even in a fixed background, demonstrating the procedure.
[Sec. 5] The generalized Noether claim—that the specific pattern of energy-momentum non-conservation encoded in (G(n))_{k1...kn} is sufficient to compute the metric—is presented as a conclusion of Section 4. Since Section 4's identification of U is the unproved step, the one-to-one correspondence asserted in Section 5 is not established; at most it is conditional on the existence of a position basis of the kind the paper postulates.

minor comments (5)

[Sec. 3] 'Spectreal geometry' should be 'spectral geometry'; also 'simultaneopusly' in Section 4 is a typo.
[Eq. (2)] The special handling for D=2 is mentioned but never given; since Eq. (2) is the basis for the metric reconstruction, the D=2 case should either be specified or explicitly declared irrelevant to the paper's argument.
[Sec. 5] The notation 'P pin = P pout' and 'δ(D)(P pi)' is undefined; the momentum-conservation delta should be written as δ^{(D)}(∑ p_i) with summation limits specified.
[Sec. 4] Several sentences in the description of the procedure are grammatically unclear, e.g., 'We can then pick any such a unitary transformations U'; a careful proofreading pass is needed.
[References] Reference [42] is listed as '(in preparation)' and is not accessible; if it is needed for a claim, it should be replaced by a preprint or the claim should be supported by an accessible reference.

Circularity Check

1 steps flagged · score 6.0 of 10

The position-basis identification is definitional: 'diagonalizing' is stipulated to mean exhibiting the locality singularity, so the claimed representation as local QFT vertices on a spacetime is built into the definition; the metric formula Eq. (2) is independent.

self definitional [Section 4, definition of 'diagonalizing' (paragraph after Eq. (4)) and Step 2; Abstract; Section 5]
"To cover the cases of weak and strong nonlinearity simultaneopusly, we can here define the term 'diagonalizing' to possess a specific technical meaning: diagonalizing is here to seek a basis in which the interaction vertex G(n>2) exhibits the unique signature of locality — namely, it is singular if and only if all its arguments coincide. [...] Abstract: 'if abstract higher-order correlators G(n>2) can be diagonalized, these correlators can be represented as local quantum field theoretic vertices on a curved spacetime whose metric gμν(x) can be explicitly calculated.'"

The paper's own definition of 'diagonalizing' is 'to seek a basis in which the interaction vertex G(n>2) exhibits the unique signature of locality — namely, it is singular if and only if all its arguments coincide.' The headline consequent, 'can be represented as local quantum field theoretic vertices on a curved spacetime,' is that same property restated: Step 2 stipulates that such a basis 'can then be identified as a position basis,' and Step 4 computes the metric from G(2) in that basis — at tree level this is just 'regular diagonalization,' so the metric is whatever Eq. (2) yields from an arbitrarily chosen diagonalizing basis.

full rationale

Non-circular core: Eq. (2), cited to [1] (whose author overlaps with the present paper), is displayed in full and is a parameter-free mathematical formula converting a Hadamard-scaled position-space propagator into a metric; the self-citations [1]-[3] are restated in the text, so they function as real evidence rather than opaque authority. The circular step is Section 4's resolution of the missing-basis problem: the paper defines 'diagonalizable' (for G^(n>2)) as 'admitting a basis in which the vertex exhibits the unique signature of locality (singular iff all arguments coincide),' stipulates that any such basis is a position basis, and reads off the metric from G^(2) in it. The abstract's headline — 'if abstract higher-order correlators G^(n>2) can be diagonalized, these correlators can be represented as local quantum field theoretic vertices on a curved spacetime whose metric gμν(x) can be explicitly calculated' — therefore has its representability-as-local-vertices portion built into the definition of the antecedent; the genuinely derived residue is the independent formula Eq. (2). Existence and uniqueness of the position basis are not proved: Section 4 concedes 'The existence of such a basis U is contingent on the underlying abstract interaction being representable as local,' and one may 'pick any such a unitary transformations U,' with only an assertion that the others differ by diffeomorphisms. The Generalized Noether 'one-to-one correspondence' (Sec. 5) inherits this unproven uniqueness. A separate correctness risk, not itself circularity: the 'singular iff all arguments coincide' fingerprint is false for renormalized interacting QFT (partial-coincidence OPE and lightcone singularities) and undefined for abstract basis labels ('coincidence of arguments' has no meaning until a manifold is selected), so the search over U is not a well-posed inverse problem; Section 7 also lists 'characterizing their diagonalizability properties' as future work. Verdict: score 6 — the metric formula is independent, but the load-bearing step that is said to complete spectral geometry (identifying the position basis from the interaction structure) is definitional rather than derived.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The central claim rests on four assumptions: the locality singularity structure, the existence of U, Hadamard scaling of the transformed propagator, and the availability of a discrete eigenbasis. None of the last three is proved, and the first is questionable for renormalized correlators. There are no fitted numerical parameters.

assumptions (4)

ad hoc to paper Local interacting QFT correlators on curved spacetime are singular if and only if all arguments coincide in a position basis.
Section 4 asserts this as the key insight; it is not proved and is not generally true beyond tree level.
ad hoc to paper There exists a unitary U from the eigenbasis of G(2) to a position basis, and it can be identified from the singularity structure of G(n).
Section 4 Step 2 assumes existence and identifiability; no construction or uniqueness proof is given.
domain assumption The transformed G(2) obeys the Hadamard scaling needed for Eq. (2).
Eq. (2) requires the propagator to have Hadamard form; for arbitrary abstract G(2) this is assumed, not shown.
domain assumption An abstract eigenbasis for the wave operator exists despite continuous Lorentzian spectra.
Section 3 notes continuous spectra for Lorentzian signature, but the procedure assumes a discrete abstract eigenbasis; the role of cutoffs is not resolved.

invented entities (1)

Abstract correlators G(n) as primary fundamental objects
purpose: To replace spacetime and matter with a pre-geometric information-theoretic structure from which the metric and fields emerge.
Section 6 postulates {G(n)} as fundamental. No falsifiable prediction is offered, and in the pre-geometric regime the objects are said to lose even their correlator interpretation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of The metric from energy-momentum non-conservation: Generalizing Noether and completing spectral geometry." pith.science (2026). https://pith.science/paper/3OOC3247

@misc{pith2026250611186,
  author       = {Pith},
  title        = {Pith review of: The metric from energy-momentum non-conservation: Generalizing Noether and completing spectral geometry},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3OOC3247}},
  note         = {Machine review of arXiv:2506.11186}
}

abstract

We complete the program of spectral geometry, in the sense that we show that a manifold's shape, i.e., its metric, can be reconstructed from its resonant sound when tapped lightly, i.e., from its spectrum, -- if in addition we also record the resonances' mutual excitation pattern when the driving is strong enough to reach the nonlinear regime. Applied to spacetime, this finding yields a generalization of Noether's theorem: the specific pattern of energy-momentum non-conservation on a generic curved spacetime, encoded within the quantum field theoretic scattering matrices, is sufficient to calculate the metric. Applied to quantum gravity, this shows that the conventional dichotomy of spacetime versus matter can emerge from an underlying information-theoretic framework of only one type of quantity: abstract correlators, $G^{(n)}$, that are, a priori, merely operators on $n$ tensor factors of a Hilbert space. This is because, on one hand, if abstract higher-order correlators $G^{(n>2)}$ can be diagonalized, these correlators can be represented as local quantum field theoretic vertices on a curved spacetime whose metric $g_{\mu\nu}(x)$ can be explicitly calculated. On the other hand, at sufficiently high energies, such as the Planck scale, the $G^{(n)}$ may not be even approximately representable as correlators of a local QFT on a spacetime, indicating a regime that is mathematically controlled but transcends the concepts of spacetime and matter.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 43 canonical work pages

[1]

Saravani M, Aslanbeigi S and Kempf A 2016 Phys. Rev. D 93 045026

work page 2016
[2]

Kempf A 2021 Front. Phys. 9 655857

work page 2021
[3]

Reitz M, ˇSoda B and Kempf A 2023 Phys. Rev. Lett. 131 211501

work page 2023
[4]

Kac M 1966 Am. Math. Mon. 73 1

work page 1966
[5]

Weyl H 1911 Nachr. d. K”onigl. Ges. d. Wiss. zu G”ottingen 110

work page 1911
[6]

Datchev K and Hezari H 2012 Inverse Problems and Applications: Inside Out II Mathematical Sciences Research Institute Publications vol 60 (Cambridge: Cambridge University Press) p 12

work page 2012
[7]

Aasen D, Bhamre T and Kempf A 2013 Phys. Rev. Lett. 110 121301

work page 2013
[8]

Synge J L 1960 Relativity: The General Theory (Amsterdam: North Holland Publishing) p 56

work page 1960

Show all 46 references

[9]

Fulling S A 1989 Aspects of Quantum Field Theory in Curved Space-time (Cambridge: Cambridge University Press)

1989
[10]

Birrell N D and Davies P C W 1982 Quantum Fields in Curved Space (Cambridge: Cambridge University Press)

1982
[11]

Einstein A 1981 Grundz¨ uge der Relativit¨ atstheorie5th ed (Braunschweig: Vieweg)

1981
[12]

Valentini A 1991 Phys. Lett. A 153 321

1991
[13]

Reznik B 2003 Found. Phys. 33 167

2003
[14]

Reznik B, Retzker A and Silman J 2005 Phys. Rev. A 71 042104

2005
[15]

Sorkin R D 2014 On the Entropy of the Vacuum outside a Horizon Preprint arXiv:1402.3589

2014 arXiv
[16]

Srednicki M 1993 Phys. Rev. Lett. 71 666

1993
[17]

Papageorgiou M and Pye J 2019 J. Phys. A: Math. Theor. 52 375304

2019
[18]

Hawking S W and Ellis G F R 1973The Large Scale Structure of Space-Time(Cambridge: Cambridge University Press)
[19]

Bachor H-A and Ralph T C 2004 A Guide to Experiments in Quantum Optics 2nd ed (Weinheim: Wiley-VCH)

2004
[20]

Walls D F and Milburn G J 2008 Quantum Optics 2nd ed (Heidelberg: Springer)

2008
[21]

Banks T 2020 Front. Phys. 8 111

2020
[22]

Shannon C E and Weaver W 1963 The Mathematical Theory of Communication (Urbana and Chicago: University of Illinois Press)

1963
[23]

Kempf A 1997 J. Math. Phys. 38 1347 (Preprint hep-th/9405067)

1997 arXiv
[24]

Mermin N D 1998 Am. J. Phys. 66 753 (Preprint quant-ph/9801057)

1998 arXiv
[25]

Jackson, D M, Kempf A, Morales A H 2021 Towards a more algebraic footing for quantum field theory. J. of Physics A: Mathematical and Theoretical, 54 (28), 285201

2021
[26]

Vanrietvelde A, Hoehn P A, Giacomini F and Castro-Ruiz E 2020 Quantum 4 225

2020
[27]

Giacomini F, Castro-Ruiz E and Brukner C 2019 Nat. Commun. 10 494

2019
[28]

Castro-Ruiz E, Giacomini F, Belenchia A and Brukner C 2020 Nat. Commun. 11 2672

2020
[29]

Giacomini F, Castro-Ruiz E and Brukner ˇC 2016 New J. Phys. 18 113026

2016
[30]

Rovelli C 2008 Living Rev. Relativ. 11 5

2008
[31]

Quantum Grav

Loll R 2020 Class. Quantum Grav. 37 013002

2020
[32]

Finster F, Kindermann S, Treude J H, 2024 Causal Fermion Systems: An Introduction to Funda- mental Structures, Methods and Applications arXiv:2411.06450

2024
[33]

Kempf A 2000 Phys. Rev. Lett. 85 2873

2000
[34]

Kempf A 2004 Phys. Rev. Lett. 92 221301

2004
[35]

Kempf A 2009 Phys. Rev. Lett. 103 231301

2009
[36]

Kempf A 2010 New J. Phys. 12 115001

2010
[37]

Kempf A 2014 Found. Phys. 44 472

2014
[38]

Pye J, Donnelly W and Kempf A 2015 Phys. Rev. D 92 105022

2015
[39]

Chatwin-Davies A, Kempf A and Martin R T W 2017 Phys. Rev. Lett. 119 031301

2017
[40]

Kempf A 2018 Found. Phys. 48 1191

2018
[41]

Quantum Grav

Yazdi Y K, Letizia M and Kempf A 2021 Class. Quantum Grav. 38 015011

2021
[42]

ˇSoda B and Kempf A (in preparation)
[43]

XXII DGM Conf

Kempf A 1994 Proc. XXII DGM Conf. (Ixtapa, Mexico) Adv. Appl. Cliff. Alg (Proc. Suppl.) S1

1994
[44]

Kempf A, Mangano G and Mann R B 1995 Phys. Rev. D 52 1108

1995
[45]

Scardigli F 1999 Phys. Lett. B 452 39

1999
[46]

Chatwin-Davies A, Kempf A, Simidzija P 2023 Covariant predictions for Planck-scale features in primordial power spectra Physical Review D, 107(10), 103527

2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Saravani M, Aslanbeigi S and Kempf A 2016 Phys. Rev. D 93 045026

work page 2016

[2] [2]

Kempf A 2021 Front. Phys. 9 655857

work page 2021

[3] [3]

Reitz M, ˇSoda B and Kempf A 2023 Phys. Rev. Lett. 131 211501

work page 2023

[4] [4]

Kac M 1966 Am. Math. Mon. 73 1

work page 1966

[5] [5]

Weyl H 1911 Nachr. d. K”onigl. Ges. d. Wiss. zu G”ottingen 110

work page 1911

[6] [6]

Datchev K and Hezari H 2012 Inverse Problems and Applications: Inside Out II Mathematical Sciences Research Institute Publications vol 60 (Cambridge: Cambridge University Press) p 12

work page 2012

[7] [7]

Aasen D, Bhamre T and Kempf A 2013 Phys. Rev. Lett. 110 121301

work page 2013

[8] [8]

Synge J L 1960 Relativity: The General Theory (Amsterdam: North Holland Publishing) p 56

work page 1960