Closed geodesics on compact Lorentzian solvmanifolds

Gabriela P. Ovando; Pablo Montenegro

REVIEW 3 major objections 6 minor 14 references

Closed geodesics on compact Lorentzian solvmanifolds

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read In compact quotients of oscillator groups by cocompact lattices, all lightlike geodesics are closed exactly when the lattice contains a nonzero element of the form $(0,0,t)$; otherwise exactly one lightlike direction, spanned by $Z$, has…

desk verdict The main dichotomy is likely correct and genuinely new; the paper is worth a serious referee, but the proof of Eq. (9) is underjustified and Theorem 3.6's key nonsingularity claim is false as stated. read the letter →

arxiv 2411.14237 v1 pith:3QD453XP submitted 2024-11-21 math.DG

Pablo Montenegro , Gabriela P. Ovando This is my paper · ORCID

classification math.DG MSC 53C5053C2222F3057S25

keywords Lorentziangeometrygeodesicscompactsolvmanifoldsoscillatorgroupscocompactlatticesbi-invariantmetricsisometrynormalizers

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 studies compact Lorentzian manifolds formed by dividing an oscillator Lie group by a discrete cocompact subgroup, with the metric invariant under both left and right translations. Its central result is a dichotomy: either the lattice contains a nonzero element of the form $(0,0,t)$, in which case every lightlike (null) geodesic of the quotient is closed, or it contains no such element, in which case at each point exactly one lightlike direction—spanned by the central vector $Z$—has only closed geodesics and every other lightlike direction has only non-closed ones. This matters because it shows that "all light rays close" is not a fixed geometric property of a compact Lorentzian solvmanifold but a property of the lattice used to build it, and both behaviours occur already in dimension four. The paper also establishes that timelike and spacelike geodesics always split into closed and open types, and it computes the isometry group of the quotients, including explicit normalizer tables in dimension six.

What carries the argument

The central object is the oscillator Lie group $\mathrm{Osc}_n(\lambda_1,\ldots,\lambda_n)$: a $(2n+2)$-dimensional solvable Lie group with Lie algebra spanned by $Z$, $X_i$, $Y_i$, $T$, brackets $[X_i,Y_i]=Z$, $[T,X_i]=\lambda_i Y_i$, $[T,Y_i]=-\lambda_i X_i$, and Lorentzian bi-invariant metric with $\lambda_i\langle X_i,X_i\rangle = \lambda_i\langle Y_i,Y_i\rangle = \langle Z,T\rangle = 1$. The mechanism is that geodesics starting at the identity are one-parameter subgroups, and for a lightlike initial velocity with $a\neq 0$ the explicit solution of the geodesic equations yields $\alpha(K_0t_0/a) = (0,0,K_0t_0)$. The lattice structure theorem supplies a positive generator $t_0$ of the $t$-coordinates of $\Gamma$ and a common integer $K_0$ such that $t_0 = 2\pi k_i/(K_0\lambda_i)$ for every $i$, which makes $\exp(K_0 t_0 N_\lambda)=\mathrm{Id}$. Thus the closedness of a lightlike geodesic reduces to the lattice-membership question "is $(0,0,K_0t_0)$ in $\Gamma$?", and the dichotomy follows from the answer. For $a=0$ the lightlike geodesics are $(ds,0,0)$, which meet $\Gamma$ because Lemma 3.2 guarantees an element $(w,0,0)\in\Gamma$.

What would settle it

Take a lattice $\Gamma$ in $\mathrm{Osc}_1(1)$ that contains no nonzero $(0,0,t)$—for instance one of the automorphic images $\varphi_p(\Lambda_{n,0})$ in Example 3.5—and solve the lightlike condition $2ad+(b^2+c^2)/\lambda=0$ together with the geodesic equations (4)-(7) for a positive time $s$ with $\gamma_X(s)\in\Gamma$ and $(b,c)\neq(0,0)$. Any such solution would refute the theorem's claim that only the $Z$-direction gives closed lightlike geodesics. Equivalently, checking membership of $(0,0,K_0t_0)$ in $\Gamma$ for a lattice with no nonzero $(0,0,t)$ would settle the dichotomy directly.

Watch

Extended reading notes

Core claim

The central claim, Theorem 3.4, is a dichotomy for any cocompact lattice $\Gamma$ in $\mathrm{Osc}_n(\lambda_1,\ldots,\lambda_n)$. Write $M = \mathrm{Osc}_n(\lambda_1,\ldots,\lambda_n)/\Gamma$. If $\Gamma$ contains an element $(0,0,t_0)$ with $t_0\neq 0$, then every lightlike geodesic of $M$ is closed. If $\Gamma$ contains no nonzero element of this form, then at every point of $M$ exactly one lightlike direction—the direction spanned by $Z$—has all its geodesics closed, and all other lightlike directions have non-closed geodesics. The proof shows that any lightlike geodesic with initial velocity $X = dZ + \sum_j(b_jX_j+c_jY_j) + aT$, $a\neq 0$, satisfies $\alpha(K_0 t_0/a) = (0,0,K_0 t_0)$, where $K_0$ is the common integer in the lattice relation $t_0 = 2\pi k_i/(K_0\lambda_i)$; hence the geodesic closes exactly when $(0,0,K_0 t_0)$ lies in $\Gamma$. The paper further claims (Theorem 3.6) that every such quotient contains both closed and open timelike geodesics and both closed and open spacelike geodesics, and that the isometries of $M$ are precisely the left translations together with inner automorphisms coming from the normalizer of $\Gamma$.

Load-bearing premise

The dichotomy rests on the structural fact, taken from the cited lattice classification, that every cocompact lattice has one common integer $K_0$ with $t_0 = 2\pi k_i/(K_0\lambda_i)$ for all frequencies; if that fact fails, the key return point $(0,0,K_0t_0)$ is not guaranteed and the proof's dichotomy no longer follows.

Editorial extensions

If this is right

If $\Gamma$ contains $(0,0,t)$ for some $t\neq 0$, then every lightlike geodesic of $M = \mathrm{Osc}_n(\lambda_1,\ldots,\lambda_n)/\Gamma$ is closed.
If $\Gamma$ contains no nonzero element of the form $(0,0,t)$, then at every point of $M$ exactly one lightlike direction—the line spanned by $Z$—has all its geodesics closed, and every other lightlike direction has non-closed geodesics.
For every lattice $\Gamma$, the quotient contains closed timelike geodesics, open timelike geodesics, closed spacelike geodesics, and open spacelike geodesics; the dichotomy is specific to the lightlike case.
Both sides of the dichotomy are realized by explicit lattices in dimension four, so compact Lorentzian solvmanifolds with the same oscillator group and metric can have different lightlike-geodesic behaviour.
The isometries of the quotient are exactly the left translations and the inner automorphisms induced by elements of the normalizer $N_G(\Gamma)$; the paper's explicit normalizer tables for $\Lambda_{k,q,M}$ in six dimensions determine the isometry group of those quotients.

Reading between the lines

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

Editorial extension: the dichotomy suggests that the generic lattice—one obtained by deforming a special lattice with an automorphism, as in the $\varphi_p(\Lambda_{n,0})$ examples—falls into the second, non-closed case, so the all-closed behaviour is exceptional rather than generic.
Editorial extension: because the proof only uses that geodesics are one-parameter subgroups and that a common period $K_0$ exists, the same dichotomy should hold for any solvable Lie group with bi-invariant Lorentzian metric whose cocompact lattices satisfy an analogous common-period condition; oscillator groups are the indecomposable case where the relevant lattices are explicitly classified.
Editorial extension: the openness argument for timelike and spacelike geodesics is essentially a discreteness argument—if all nearby initial data closed, their endpoints would converge inside the lattice—so a similar argument could give quantitative bounds on the density of closed timelike and spacelike geodesics in terms of lattice invariants.

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

Pablo and Gabriela's paper gives a complete lattice-dependent dichotomy for lightlike geodesics on oscillator solvmanifolds, in all dimensions, and computes normalizers in dimension six. That is a real advance over the dimension-four examples in del Barco–Ovando–Vittone, and the main theorem is very likely true. The structure of the argument — one-parameter subgroups of a bi-invariant Lorentzian group, projected to a compact quotient, with closedness controlled by lattice intersections — is sound and clean.

The soft spots are all in the proofs, not in the statements. The derivation of Eq. (9) is too quick: finiteness of the conjugate orbit of a single lattice vector does not imply e^{t0 N} has finite order. The fix is almost immediate, because the t=0 part of a cocompact lattice projects to a full-rank lattice in R^{2n}, and finiteness on a spanning set gives a common power. But the text does not say that, and the authors' own marginal note suggests they noticed. A referee should ask for that step to be written properly.

Theorem 3.6 has a concrete error: the matrix in Eq. (15) is claimed nonsingular at λ_j(K0−1)t0, but it is singular whenever K0 divides k_j. That happens for a natural class of lattices, so the construction of closed timelike and spacelike geodesics in the K0>1 case does not go through as written. The theorem may be salvageable with a different choice of lattice element, but the present proof fails.

The w=e example in Section 3.1 is also a slip. It relies on the assertion that e^2 is not a rational multiple of π, which is not known; choosing w=1 avoids the problem entirely since π is irrational. That's a one-line fix.

Section 4's normalizer tables look careful and are a useful addition, though I did not re-derive all the cases.

Bottom line: the main dichotomy is probably correct and worth publishing, but the manuscript needs a proper revision before it is refereed cleanly. The flaws are specific and repairable. I would send it to a good referee, with a request to check the lightlike proof and the closed timelike/spacelike construction carefully.

Referee Report

3 major / 6 minor

Summary. The paper studies geodesics on compact Lorentzian solvmanifolds obtained as quotients of oscillator groups Osc_n(λ_1,...,λ_n) by cocompact lattices. The main result, Theorem 3.4, asserts a dichotomy for lightlike geodesics: either the lattice contains a nonzero element of the form (0,0,t), in which case every lightlike geodesic is closed, or it does not, in which case exactly the Z-direction gives closed lightlike geodesics and all other lightlike directions give non-closed geodesics. Theorem 3.6 claims that for every lattice there exist both closed and open timelike and spacelike geodesics. The final section computes normalizers of certain lattices and describes induced isometries, with explicit tables in dimension six. The paper also contains an example (Osc_1(1)×R) where no lightlike geodesic is closed.

Significance. If the main dichotomy is correct, it is a meaningful advance: it shows that the property 'all lightlike geodesics are closed' is not an invariant of the oscillator group but depends sensitively on the lattice, thereby sharpening earlier results of del Barco, Ovando, and Vittone. The explicit normalizer computations in Proposition 4.14 are concrete and checkable, and the use of Fischer's lattice classification is appropriate. The paper does not contain machine-checked proofs or reproducible code; its value lies in the geometric constructions and the classification arguments. However, the central proofs contain a gap in the derivation of the common integer K0 and a false nonsingularity claim in Theorem 3.6, both of which are load-bearing. The underlying statements may well be true with additional argument, but the manuscript in its current form does not establish them rigorously.

major comments (3)

[§3.4, Theorem 3.6, Eq. (15)] The proof that there exists a common integer K0 with e^{K0 t0 N}=Id is incomplete as written. The text observes that for a single lattice element (w,b,0), the conjugate orbit {(w, e^{n t0 N} b, 0)} is finite, and concludes that e^{K0 t0 N}=Id on all of R^{2n}. Finiteness of the orbit of one vector under an orthogonal transformation does not imply that the transformation has finite order. The argument can be repaired by noting that the projection of Γ to R^{2n} is a lattice and that e^{t0 N} preserves it, so e^{t0 N} lies in the finite group O(2n) ∩ GL(2n,Z) in a lattice basis; hence some power is the identity. Since this step supports Eq. (9), Observations 3.3, and the dichotomy in Theorem 3.4, it must be rewritten rigorously. The marginal note 'esta bien?...' in the proof of Theorem 3.4 signals the authors' own uncertainty about this step.
[Theorem 3.6] The claim that the matrix in Eq. (15) is non-singular for t=(K0−1)t0 is false in general. From Eq. (9), λ_j(K0−1)t0 = 2π(K0−1)k_j/K0, which is an integer multiple of 2π whenever K0 divides k_j; this does not force K0=1. For example, K0=2 and k_j=2 gives (K0−1)k_j/K0=1, so the 2×2 block has determinant zero. Therefore the construction of closed timelike and spacelike geodesics in the case K0>1 is not justified as written. The theorem may still be true, but a different choice of lattice element (for instance t=t0, where not all blocks are singular because the order of e^{t0 N} is exactly K0) or a compatibility argument is needed.
[Theorem 3.4] The final sentence of the proof states that when an element of the form (0,0,kt0) is in the lattice, every lightlike geodesic of M is closed, whereas the theorem's first case is phrased with an element (0,0,t0). The proof correctly reduces to the existence of a nonzero (0,0,rt0), but the statement should be made uniform and the implication 'there exists (0,0,rt0) in Γ' should be explicitly aligned with the dichotomy. More importantly, the proof that a closed lightlike geodesic with direction independent of Z forces the lattice to contain such an element relies on Eq. (9); without the repaired derivation of K0, the dichotomy is not established. This is a load-bearing point and must be fixed before the theorem can be accepted.

minor comments (6)

[Abstract] The abstract contains the typo 'lightlight' in 'every lightlight geodesic' and 'closeness' should be 'closedness'.
[Section 2, Eq. (12)] The notation '~t0' in Eq. (12) is undefined. The text says it is either 1/t0 or −1/t0, but this should be stated precisely, and the expression 'SNλS−1 1' appears to contain a typo.
[Example 3.5] The phrase 'most likely do not contain' for the lattices φ(Λ_{n,•}) is imprecise. The example with φ_p(Λ_{n,0}) gives a rigorous proof, but for the other families a definitive statement or a proof sketch should be given.
[Section 4.1] In the dimension-four normalizer list, the third bullet repeats 'For Λ_{k,π}' but should refer to Λ_{k,π/2}.
[Proposition 4.14] The table header 'N(Λ_{k,1,q})' appears to be a typo; it should be 'N(Λ_{k,q,M})'.
[References] The reference list for [4] shows duplicated page numbers and inconsistent formatting; please check the bibliographic details.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation found: the lightlike dichotomy follows from lattice-structural input, and the only self-citation is a non-load-bearing baseline.

full rationale

The central result (Theorem 3.4) is not circular: its dichotomy is derived from the lattice structural fact (Eq. (9), t0 = 2π k_i/(K0 λ_i)) imported from the Medina-Revoy/Fischer classification, plus discreteness of Γ, rather than from the closed-geodesic conclusion. The geodesic formulas (4)-(8) are derived independently from the bi-invariant metric, and the implication from a closed non-Z lightlike geodesic to (0,0,K0 r t0) ∈ Γ is a forward argument, not an assumed conclusion. No fitted parameter is renamed as a prediction. The only author-overlapping citation is [4] (Ovando), used for Lemma 2.1 and Theorem 4.10 in the isometry section; these are prior published baselines and do not force the main geodesic dichotomy, so under the review rules they do not raise the circularity score. I flag as non-circular correctness concerns: the marginal note 'esta bien?...' in the proof of Theorem 3.4 and the invalid nonsingularity claim in the K0 > 1 case of Theorem 3.6 (λ_j(K0−1)t0 = 2πs_j only implies K0 divides k_j, not K0 = 1). These affect validity, not circularity; hence score 1 reflects only a minor self-citation with no load-bearing circular step.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters or invented entities appear. The derivation is deductive; all nontrivial input comes from cited theorems about lattices, isometries, and Lie group geometry, which are treated as axioms for this paper.

assumptions (5)

standard math Existence of geodesics as one-parameter subgroups for bi-invariant metrics (O'Neill, Chapter 11).
Used throughout Section 2 to reduce geodesics on the quotient to one-parameter subgroups of G that intersect the lattice.
standard math Medina-Revoy classification: oscillator algebras and sl(2,R) are the only indecomposable Lie algebras admitting a Lorentzian ad-invariant metric.
Invoked in Remark 1 to frame the class of groups studied; not used directly in the main proofs.
domain assumption Fischer's lattice structure results: every cocompact lattice of Osc_n is isomorphic to a lattice L(xi0) in Osc_n(omega_r,B_r); T(Gamma)=Z t0; exp(K0 t0 N_lambda)=Id; Lemma 3.2 gives (w,0,0) in Gamma and, for K0=1, (z,0,t) in Gamma.
Used centrally in Observations 3.3 and in the proofs of Theorems 3.4 and 3.6. If any of these structural facts fails, the dichotomy does not follow.
domain assumption Muller's theorem characterizing differentials of local isometries of bi-invariant metrics (Theorem 4.1) and Bourseau's description of F(Osc_n) (Theorem 4.6).
Basis for Section 4 isometry computations; accepted as prior theorems without reproof.
domain assumption Discreteness of Gamma, local isometry lifting, and the normalizer criterion for fiber-preserving isometries (Theorem 4.10 from reference [4]).
Used to compute fiber-preserving isometries and the normalizer tables in Propositions 4.11 and 4.14.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Closed geodesics on compact Lorentzian solvmanifolds." pith.science (2026). https://pith.science/paper/3QD453XP

@misc{pith2026241114237,
  author       = {Pith},
  title        = {Pith review of: Closed geodesics on compact Lorentzian solvmanifolds},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3QD453XP}},
  note         = {Machine review of arXiv:2411.14237}
}

abstract

The aim of this work is the study of geodesics on Lorentzian homogeneous spaces of the form $M=G/\Lambda$, where $G$ is a solvable Lie group endowed with a bi-invariant Lorentzian metric and $\Lambda < G$ is a cocompact lattice. Conditions to assert closedness of light, time or spacelike geodesics on the compact quotient spaces are given. This study implicitly requires additional information about the lattices in each case. We found conditions for which every lightlight geodesic on the quotient space is closed. And more important, this situation depends on the lattice. Moreover, even in dimension four, there are examples of compact solvmanifolds for which not every lightlike geodesic is closed. For time and spacelike geodesics, the conclusion are different. Finally, we study isometry groups of those compact spaces and show some computations in dimension six.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[1]

Baues, W

O. Baues, W. Globke , Rigidity of compact pseudo-Riemannian homogeneous spaces for solvable Lie groups . Int. Math. Res. Not. 2018 (10), 3199–3223 (2018)

work page 2018
[2]

Beardon , The geometry of discrete groups

A. Beardon , The geometry of discrete groups . Springer (1983). First Edition. 2

work page 1983
[3]

Bourseau , Die Isometrien der Oszillatorgruppe und einige ergebnisse ¨ uber Pr¨ amorphismen liescher Algebren

F. Bourseau , Die Isometrien der Oszillatorgruppe und einige ergebnisse ¨ uber Pr¨ amorphismen liescher Algebren. Fakult¨ at f¨ ur Mathematik der Universit¨ at Biele- feld (1989)

work page 1989
[4]

del Barco, G

V. del Barco, G. Ovando, F. Vittone , Lorentzian compact manifolds: Isome- tries and geodesics , J. Geom. Phys. 78, 48–58 (2014). 2, 14, 16 2, 6, 9, 18, 19, 20

work page 2014
[5]

Fischer , Lattices of oscillator groups , J

M. Fischer , Lattices of oscillator groups , J. Lie Theory 27 (1), 85–110 (2017). 6, 7, 9, 10

work page 2017
[6]

Galloway , Compact Lorentzian manifolds without closed nonspacelike geodesics, Proc

G. Galloway , Compact Lorentzian manifolds without closed nonspacelike geodesics, Proc. Amer. Math. Soc. 98, 119–123 (1986). 1

work page 1986
[7]

Helgasson , Diﬀerential Geometry, Lie Groups, and Symmetric Spaces , Grad- uate Studies in Mathematics, vol

S. Helgasson , Diﬀerential Geometry, Lie Groups, and Symmetric Spaces , Grad- uate Studies in Mathematics, vol. 34, American Math. Soc. (1999). 5

work page 1999
[8]

Medina , Groupes de Lie munis de m´ etriques bi-invariantes

A. Medina , Groupes de Lie munis de m´ etriques bi-invariantes . (Lie groups ad- mitting bi-invariant metrics), Tˆ ohoku Math. J., II. Ser. 37, 405–421 (1985). 5

work page 1985

Show all 14 references

[9]

Medina, P

A. Medina, P. Revoy , Les groupes oscillateurs et leurs r´ eseaux. (Oscillator groups and their lattices). , Manuscr. Math. 52, 81–95 (1985). 5, 6

1985
[10]

M ¨uller, Isometries of bi-invariant pseudo-Riemannian metrics on L ie groups, Geom

D. M ¨uller, Isometries of bi-invariant pseudo-Riemannian metrics on L ie groups, Geom. Dedicata 29 (1), 65–96 (1989). 13

1989
[11]

O’Neill, Semi-Riemannian geometry with applications to relativity , Academic Press (1983)

B. O’Neill, Semi-Riemannian geometry with applications to relativity , Academic Press (1983). 1, 3, 4, 5

1983
[12]

Ovando, Lie algebras with ad-invariant metrics- A survey , In Memorian Sergio Console, Rendiconti del Seminario Matematico di Torino

G. Ovando, Lie algebras with ad-invariant metrics- A survey , In Memorian Sergio Console, Rendiconti del Seminario Matematico di Torino. 74, 1-2, 241 – 266 (2016)

2016
[13]

Raghunathan , Discrete Subgroups of Lie Groups, Springer Verlag, 1972

M.S. Raghunathan , Discrete Subgroups of Lie Groups, Springer Verlag, 1972. 18 GEODESICS AND ISOMETRIES ON COMPACT LORENTZIAN SOL VMANIFO LDS 25

1972
[14]

Suhr , Closed geodesics in Lorentzian surfaces , Trans

S. Suhr , Closed geodesics in Lorentzian surfaces , Trans. Amer. Math. Soc. 365, 1469–1486 (2013). 2 Departamento de Matem ´atica, ECEN - FCEIA, Universidad Nacional de Rosario, Pellegrini 250, 2000 Rosario, Santa Fe, Argentina . Email address : gabriela@fceia.unr.edu.ar

2013

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Baues, W

O. Baues, W. Globke , Rigidity of compact pseudo-Riemannian homogeneous spaces for solvable Lie groups . Int. Math. Res. Not. 2018 (10), 3199–3223 (2018)

work page 2018

[2] [2]

Beardon , The geometry of discrete groups

A. Beardon , The geometry of discrete groups . Springer (1983). First Edition. 2

work page 1983

[3] [3]

Bourseau , Die Isometrien der Oszillatorgruppe und einige ergebnisse ¨ uber Pr¨ amorphismen liescher Algebren

F. Bourseau , Die Isometrien der Oszillatorgruppe und einige ergebnisse ¨ uber Pr¨ amorphismen liescher Algebren. Fakult¨ at f¨ ur Mathematik der Universit¨ at Biele- feld (1989)

work page 1989

[4] [4]

del Barco, G

V. del Barco, G. Ovando, F. Vittone , Lorentzian compact manifolds: Isome- tries and geodesics , J. Geom. Phys. 78, 48–58 (2014). 2, 14, 16 2, 6, 9, 18, 19, 20

work page 2014

[5] [5]

Fischer , Lattices of oscillator groups , J

M. Fischer , Lattices of oscillator groups , J. Lie Theory 27 (1), 85–110 (2017). 6, 7, 9, 10

work page 2017

[6] [6]

Galloway , Compact Lorentzian manifolds without closed nonspacelike geodesics, Proc

G. Galloway , Compact Lorentzian manifolds without closed nonspacelike geodesics, Proc. Amer. Math. Soc. 98, 119–123 (1986). 1

work page 1986

[7] [7]

Helgasson , Diﬀerential Geometry, Lie Groups, and Symmetric Spaces , Grad- uate Studies in Mathematics, vol

S. Helgasson , Diﬀerential Geometry, Lie Groups, and Symmetric Spaces , Grad- uate Studies in Mathematics, vol. 34, American Math. Soc. (1999). 5

work page 1999

[8] [8]

Medina , Groupes de Lie munis de m´ etriques bi-invariantes

A. Medina , Groupes de Lie munis de m´ etriques bi-invariantes . (Lie groups ad- mitting bi-invariant metrics), Tˆ ohoku Math. J., II. Ser. 37, 405–421 (1985). 5

work page 1985