On holonomy of Weyl connections in Lorentzian signature

https://doi.org/10.1016/j.difgeo.2021.101759Get rights and content

Abstract

Holonomy algebras of Weyl connections in Lorentzian signature are classified. In particular, examples of Weyl connections with all possible holonomy algebras are constructed.

Introduction

The holonomy group of a connection is an important invariant. This motivates the classification problem for holonomy groups. There are classification results for some cases of linear connections. There is a classification of irreducible connected holonomy groups of linear torsion-free connections [27]. Important result is the classification of connected holonomy groups of Riemannian manifolds [6], [7], [10], [24]. Lorentzian holonomy groups are classified [4], [26], [13], [3], [21], [16], [17]. There are partial results for holonomy groups of pseudo-Riemannian manifolds of other signatures [8], [9], [5], [11], [20], [19], [18], [15], [23], [30], [31].

Of certain interests are Weyl manifolds (M,c,), where c is a conformal class of pseudo-Riemannian metrics and ∇ is a torsion-free linear connection preserving c. In the Riemannian signature the connected holonomy groups of such connection are classified [2], [22].

The result of this paper is a complete classification of connected holonomy groups (equivalently, of holonomy algebras) of Weyl connections in the Lorentzian signature.

In Sections 2 we give necessary background on Weyl manifolds and Berger algebras. Berger algebras have the same algebraic properties as the holonomy algebras and they are candidates to the holonomy algebras. In Section 3 the main results of the paper are stated. Theorem 3 gives the classification of the Berger algebras that preserve proper non-degenerate subspaces of the Minkowski space, these algebras correspond to conformal products in the sense of [2]. Then we assume that the algebras do not preserve any proper non-degenerate subspace of the Minkowski space, these algebras are called weakly irreducible. If such an algebra is different from co(1,n+1), then it preserves an isotropic line. Theorem 4 gives the classification of such Berger algebras. Theorem 5 states that each obtained Berger algebra is the holonomy algebra of a Weyl connection. The rest of the paper is dedicated to the proofs of the main results. In Section 4 we describe weakly irreducible subalgebras of co(1,n+1). In Section 5 we provide the auxiliary results, namely, we found some spaces of algebraic curvature tensors using the first prolongation of Lie algebras representations. In Section 6 we describe the structure of the spaces of curvature tensors for subalgebras of co(1,n+1). The proofs of Theorem 3, Theorem 4 are given in Sections 7 and 8 correspondingly. Finally, in Section 9, for each Berger algebra gco(1,n+1), we construct a Weyl connection ∇ such that the holonomy algebra of ∇ is g, this proves Theorem 5.

The author would like to express his sincere gratitude to A. Galaev for useful discussions and suggestions. The author is grateful to the anonymous referee for careful reading of the paper and valuable comments that greatly affected the final appearance of the paper. The work was supported by the project MUNI/A/1160/2020.

Section snippets

Preliminaries

Denote by (M,c) a conformal manifold, where M is a smooth manifold, and c is a conformal class of pseudo-Riemannian metrics on M. Recall that two metrics g and h are conformally equivalent if and only if h=e2fg, for some fC(M).

Definition 1

A Weyl connection ∇ on a conformal manifold (M,c) is a torsion-free linear connection that preserves the conformal class c. The triple (M,c,) is called a Weyl manifold.

By preserving a conformal class, we understand that if gc, then there exists a 1-form ωg such that

Main results

Here we present the classification of the holonomy algebras gco(1,n+1) of the Weyl connections such that gso(1,n+1), n0.

First suppose that gco(1,n+1) is irreducible. Then it is obvious that prso(1,n+1)gso(1,n+1) is irreducible as well. Therefore, prso(1,n+1)g=so(1,n+1), since so(1,n+1) does not have any proper irreducible algebra [29]. Thus, g=co(1,n+1).

Next, let us suppose that g preserves a non-degenerate subspace of R1,n+1.

Theorem 3

Let gco(1,n+1), gso(1,n+1), be a Berger algebra which admits

Weakly irreducible subalgebras of co(1,n+1)Rp

We will consider the obvious projectionprso(1,n+1):co(1,n+1)=Rso(1,n+1)so(1,n+1), which is a homomorphism of the Lie algebras. For a subalgebra gco(1,n+1), denote by g its projection to so(1,n+1). It is obvious that g is weakly irreducible if and only if g is weakly irreducible. It is clear that if g is weakly irreducible and is not contained in so(1,n+1), then only the following two situations are possible:

  • (a)

    g contains RidR1,n+1, in this caseg=RidR1,n+1g;

  • (b)

    g does not contain RidR1,n+1 and

Auxiliary results

Let Rr,s be a pseudo-Euclidean space. We will use the standard isomorphism 2Rr,sso(r,s) of the so(r,s)-modules:(XY)Z=(X,Z)Y(Y,Z)X.

Definition 5

For a subalgebra ggl(n,R) its first prolongation is defined as:g(1):={φHom(Rn,g)|φ(X)Y=φ(Y)X}.

We are interested in the first prolongations, since below we will see that they are tightly related to algebraic curvature tensors. It is well known that (so(r,s))(1)=0 andco(r,s)(1)={Z+(Z,)idRr,s|ZRr,s}Rr,s.

Lemma 1

Let h be a subalgebra of so(r,s) satisfying one of the

Algebraic curvature tensors

For a subalgebra hso(n) consider the following space:P(h)={PHom(Rn,h)|g(P(X)Y,Z)+g(P(Y)Z,X)+g(P(Z)X,Y)=0,X,Y,ZRn} defined in [26], see also [14]. The space P(h) is called the space of weak curvature tensors of type h.

Theorem 7

Every algebraic curvature tensor RR(co(1,n+1)Rp) is uniquely determined by the elements:μ,λR,X0,Z0Rn,γHom(Rn,R),PP(so(n)),K2Rn,S+τidRnR(co(n)),whereSHom(2Rn,so(n)),τHom(2Rn,R) by the equalitiesR(p,q)=(μ,λ,A0,X0),R(p,V)=((Z0,V),(Z0,V),Z0V,(A0+μEn)V),R(U,V)=((A0U,V),(

Proof of Theorem 3

By the assumption of the theorem we have a g-invariant decompositionR1,n+1=R1,k+1Rnk, i.e.,gso(1,k+1)so(nk)RidR1,n+1. We will consider two cases.

Case 1. First suppose, that prso(1,k+1)g is irreducible. Then, prso(1,k+1)g=so(1,k+1) and k1. Consider the ideal a=gso(nk)prso(nk)g. Since each subalgebra of so(nk) is reductive, there exists an ideal bprso(nk)g such thatprso(nk)g=ab. By the definition of the ideal a, the projection prso(1,n+1)g may be described in the following way.

Proof of Theorem 4

For convenience, we will use the notation of Theorem 7 for the components of an algebraic curvature tensor R. Consider the cases as in Section 4.

Case a. Let g=RidR1,n+1g, where g of the type 1, 2 or 3 from Theorem 2. If g is a Berger algebra, then according to Theorem 9, h is the holonomy algebra of a Riemannian manifold. Consider the algebraic curvature tensor R, which is defined by the condition μ=1 and all other elements are zeros. Since RR(g), it holds RidR1,n+1L(R(g)). Since g is a

Realization of Berger algebras

In this section we prove Theorem 5. To do this we show that for every Berger algebra gco(1,n+1) obtained above there exists a Weyl connection ∇ such that the holonomy algebra of ∇ is isomorphic to g.

Let (M,c,) be a Weyl manifold. Fix a metric gc. There exists a 1-form ω such thatg=2ωg. Let be the Levi-Civita connection corresponding to g. It is well-known that the connection ∇ is uniquely defined by g and ω through the formulas=+K,g(KX(Y),Z)=g(Y,Z)ω(X)+g(X,Z)ω(Y)g(X,Y)ω(Z). In what

References (32)

  • N.I. Bezvitnaya

    Holonomy groups of pseudo-quaternionic-Kählerian manifolds of non-zero scalar curvature

    Ann. Glob. Anal. Geom.

    (2011)
  • R.L. Bryant

    Metrics with exceptional holonomy

    Ann. Math.

    (1987)
  • A. Fino et al.

    Holonomy groups of G2-manifolds

    Trans. Am. Math. Soc.

    (2019)
  • A.S. Galaev

    Metrics that realize all Lorentzian holonomy algebras

    Int. J. Geom. Methods Mod. Phys.

    (2006)
  • A.S. Galaev

    Note on the holonomy groups of pseudo-Riemannian manifolds

    Math. Notes

    (2013)
  • A.S. Galaev

    Holonomy of Einstein Lorentzian manifolds

    Class. Quantum Gravity

    (2010)
  • View full text