Dynamical and Hamiltonian formulation of General Relativity
Domenico Giulini Institute for Theoretical Physics Riemann Center for Geometry and Physics Leibniz University Hannover, Appelstrasse 2, D-30167 Hannover, Germany and ZARM Bremen, Am Fallturm, D-28359 Bremen, Germany
Abstract This is a substantially expanded version of a chapter-contribution to The Springer Handbook of Spacetime, edited by Abhay Ashtekar and Vesselin Petkov, published by Springer Verlag in 2014. It introduces the reader to the reformulation of Einstein’s field equations of General Relativity as a constrained evolutionary system of Hamiltonian type and discusses some of its uses, together with some technical and conceptual aspects. Attempts were made to keep the presentation self contained and accessible to first-year graduate students. This implies a certain degree of explicitness and occasional reviews of background material.
Contents 1 Introduction
1 Introduction 1 The purpose of this contribution is to explain how 2 Notation and conventions 2 the field equations of General Relativity—often simply referred to as Einstein’s equations—can be 3 Einstein’s equations 4 understood as dynamical system; more precisely, 4 Spacetime decomposition 9 as a constrained Hamiltonian system.
5 Curvature tensors 15 In General Relativity, it is often said, spacetime becomes dynamical. This is meant to say that 6 Decomposing Einstein’s equations 22 the geometric structure of spacetime is encoded in 7 Constrained Hamiltonian systems 27 a field that, in turn, is subject to local laws of prop- agation and coupling, just as, e.g., the electromag- 8 Hamiltonian GR 36 netic field. It is not meant to say that spacetime as a whole evolves. Spacetime does not evolve, space- 9 Asymptotic flatness and global charges 48 time just is. But a given spacetime (four dimen- 10 Black-Hole data 54 sional) can be viewed as the evolution, or history, of space (three dimensional). There is a huge re- 11 Further developments, problems, and outlook 60 dundancy in this representation, in the sense that 12 Appendix: Group actions on manifolds 61 apparently very different evolutions of space rep- resent the same spacetime. However, if the result- References 66 ing spacetime is to satisfy Einstein’s equations, the Index 72 evolution of space must also obey certain well de-
1 fined restrictions. Hence the task is to give pre- Dirac’s [54] from 1958. He also noticed its con- cise mathematical expression to the redundancies strained nature and started to develop the cor- in representation as well as the restrictions of evolu- responding generalization of constrained Hamilto- tion for this picture of spacetime as space’s history. nian systems in [53] and their quantization [55]. This will be our main task. On the classical side, this developed into the more This dynamical picture will be important for geometric Dirac-Bergmann theory of constraints posing and solving time-dependent problems in [76] and on the quantum side into an elaborate General Relativity, like the scattering of black holes theory of quantization of systems with gauge re- with its subsequent generation and radiation of dundancies; see [80] for a comprehensive account. gravitational waves. Quite generally, it is a key Dirac’s attempts were soon complemented by an technology to extensive joint work of Richard Arnowitt, Stanley Deser, and Charles Misner - usually and henceforth • formulate and solve initial value problems; abbreviated by ADM. Their work started in 1959 • integrate Einstein’s equations by numerical by a paper [3] of the first two of these authors and codes; continued in the series [4] [6] [5] [9] [8] [7] [10] [11] [12] [14] [15] [13] of 12 more papers by all three. A • characterize dynamical degrees of freedom; comprehensive summary of their work was given in • characterize isolated systems and the associ- 1962 in [16], which was republished in 2008 in [17]; ation of asymptotic symmetry groups, which see also the editorial note [104] with short biogra- will give rise to globally conserved ‘charges’, phies of ADM. like energy and linear as well as angular mo- A geometric discussion of Einstein’s evolution mentum (Poincar´echarges). equations in terms of infinite-dimensional symplec- tic geometry has been worked out by Fischer and Moreover, it is also the starting point for the canon- Marsden in [57]; see also their beautiful summaries ical quantization program, which constitutes one and extended discussions in [59] and [58]. More main approach to the yet unsolved problem of on the mathematical aspects of the initial-value Quantum Gravity. In this approach one tries to problem, including the global behavior of gravita- make essential use of the Hamiltonian structure of tional fields in General Relativity, can be found the classical theory in formulating the correspond- in [39], [45], and [106]. Modern text-books on the ing quantum theory. This strategy has been ap- 3+1 formalism and its application to physical prob- plied successfully in the transition from classical lems and their numerical solution-techniques are to quantum mechanics and also in the transition [27, 77]. The Hamiltonian structure and its use in from classical to quantum electrodynamics. Hence the canonical quantization program for gravity is the canonical approach to Quantum Gravity may discussed in [34, 92, 107, 113]. be regarded as conservative, insofar as it tries to apply otherwise established rules to a classical the- ory that is experimentally and observationally ex- 2 Notation and conventions tremely well tested. The underlying hypothesis here is that we may quantize interaction-wise. This From now on “General Relativity” will be abbrevi- distinguishes this approach from string theory, the ated by “GR”. Spacetime is a differentiable man- underlying credo of which is that Quantum Gravity ifold M of dimension n, endowed with a metric only makes sense on the basis of a unified descrip- g of signature (ε, +, ··· , +). In GR n = 4 and tions of all interactions. ε = −1 and it is implicitly understood that these Historically the first paper to address the prob- are the “right” values. However, either for the lem of how to put Einstein’s equations into the sake of generality and/or particular interest, we form of a Hamiltonian dynamical system was will sometimes state formulae for general n and
2 ε, where usually n ≥ 2 (sometimes n ≥ 3) and of f. If f is a diffeomorphism we can not only push- either ε = −1 (Lorentzian metric) or ε = +1 (Rie- forward vectors and pull back co-vectors, but also mannian metric; also called Euclidean metric). vice versa. Indeed, if Y is a vector field on N one ∗ −1 The case ε = 1 has been extensively considered can write f Y := (f )∗Y and call it the pull back in path-integral approaches to Quantum Gravity, of Y by the diffeomorphism f. Likewise, if β is a −1 ∗ then referred to as Euclidean Quantum Gravity. co-vector field on M, one can write f∗β := (f ) β The tangent space of M at point p ∈ M will be and call it the push forward of β. In this fashion ∗ denoted by TpM, the cotangent space by Tp M, and we can define both, push-forwards and pull-backs, u the tensor product of u factors of TpM with d fac- of general tensor fields T ∈ ΓTd M by linearity and ∗ u ∗ tors of Tp M by Tpd M. (Mnemonic in components: applying f∗ or f tensor-factor wise. u = number of indices “upstairs”, d = number of Note that the general definition of metric is as u 0 indices “downstairs”.) An element T in Tpd M is follows: g ∈ ΓT2 M, such that gp is a symmet- called a tensor of contravariant rank u and covari- ric non-degenerate bilinear form on TpM. Such ant rank d at point p, or simply a tensors of rank a metric provides isomorphisms (sometimes called (u,d) at p. T is called contravariant if d = 0 and the musical isomorphisms)
u > 0, and covariant if u = 0 and d > 0. A tensor ∗ with u > 0 and d > 0 is then referred to as of mixed [ : TpM → Tp M 1 ∗ 0 [ type. Note that TpM = Tp0M and Tp M = Tp1M. X 7→ X := g(X, · ) , (1a) The set of tensor fields, i.e. smooth assignments of ∗ u ] : Tp M → TpM an element in Tpd M for each p ∈ M, are denoted u ω 7→ ω] := [−1(ω) . (1b) by ΓTd M. Unless stated otherwise, smooth means C∞, i.e. continuously differentiable to any order. Using ] we obtain a metric g−1 on T ∗M from the For t ∈ ΓT uM we denote by t ∈ T uM the eval- p p d p pd metric g on T M as follows: uation of t at p ∈ M. C∞(M) denotes the set of p p all C∞ real-valued functions on M, which we often −1 ] ] ] gp (ω1, ω2) := gp(ω1, ω2) = ω1(ω2) . (2) simply call smooth functions. 1 If f : M → N is a diffeomorphism between We also recall that the tensor space Tp1M is nat- manifolds M and N, then f∗p : TpM → Tf(p)N urally isomorphic to the linear space End(TpM) denotes the differential at p. The transposed (or of all endomorphisms (linear self maps) of TpM. dual) of the latter map is, as usual, denoted by Hence it carries a natural structure as associative ∗ ∗ ∗ fp : Tf(p)N → Tp M. If X is a vector field algebra, the product being composition of maps on M then f∗X is a vector field on N, called denoted by ◦. As usual, the trace, denoted Tr, the push forward of X by f. It is defined by and the determinant, denoted det, are the natu- (f∗X)q := f∗f −1(q)Xf −1(q), for all q ∈ N. If α rally defined real-valued functions on the space of is a co-vector field on N then f ∗α is a co-vector endomorphisms. For purely co- or contravariant field on M, called the pull back of α by f. It is tensors the trace can be defined by first applying ∗ defined by (f α)p := αf(p) ◦ f∗p, for all p ∈ M. For one of the isomorphisms (1). In this case we write these definitions to make sense we see that f need Trg to indicate the dependence on the metric g. generally not be a diffeomorphism; M and N need Geometric representatives of curvature are of- not even be of the same dimension. More precisely, ten denoted by bold-faced abbreviations of their if α is a smooth field of co-vectors that is defined at names, like Riem and Weyl for the (covariant, i.e. least on the image of f in N, then f ∗α, as defined all indices down) Riemann and Weyl tensors, Sec above, is always a smooth field of co-vectors on M. for the sectional curvature, Ric and Ein for the However, for the push forward f∗ of a general vec- Ricci and Einstein tensors, Scal for the scalar cur- tor field on M to result in a well defined vector field vature, and Wein for the Weingarten map (which on the image of f in N we certainly need injectivity is essentially equivalent to the extrinsic curvature).
3 This is done in order to highlight the geometric in mind when explicit models for T are used and meaning behind some basic formulae, at least the when we speak of “vacuum”, which now means: simpler ones. Later, as algebraic expressions be- −1 come more involved, we will also employ the stan- Tvacuum = TΛ := −κ gΛ . (4) dard component notation for computational ease. The signs are chosen such that a positive Λ ac- counts for a positive energy density and a negative 3 Einstein’s equations pressure if the spacetime is Lorentzian (ε = −1). There is another form of Einstein’s equations In n-dimensional spacetime Einstein’s equations which is sometimes advantageous to use and in 1 which n explicitly enters: form a set of 2 n(n + 1) quasi-linear partial dif- ferential equations of second order for 1 n(n + 1) 2 1 functions (the components of the metric tensor) Ric = κ T − n−2 g Trg(T) . (5) depending on n independent variables (the coor- dinates in spacetime). At each point of spacetime These two forms are easily seen to be mathemati- (event) they equate a purely geometric quantity to cally equivalent by the identities the distribution of energy and momentum carried 1 by the matter. More precisely, this distribution Ein = Ric − 2 g Trg(Ric) , (6a) comprises the local densities (quantity per unit vol- 1 Ric = Ein − g Trg(Ein) . (6b) ume) and current densities (quantity per unit area n−2 and unit time) of energy and momentum. The ge- With respect to a local field of basis vectors ometric quantity in Einstein’s equations is the Ein- {e0, e1, ··· , en−1} we write Ein(eµ, eν ) =: Gµν , stein tensor Ein, the matter quantity is the energy- T(eµ, eν ) =: Tµν , and Ric(eµ, eν ) =: Rµν . Then momentum tensor T. Both tensors are of second (3) and (5) take on the component forms rank, symmetric, and here taken to be covariant (in components: “all indices down”). Their num- Gµν = κ Tµν (7) ber of independent components in n spacetime di- 1 and mensions is 2 n(n + 1) 1 λ Einstein’s equations (actually a single tensor Rµν = κ Tµν − n−2 gµν Tλ (8) equation, but throughout we use the plural to emphasize that it comprises several component respectively. Next we explain the meanings of the equations) state the simple proportionality of Ein symbols in Einstein’s equations from left to right. with T Ein = κ T , (3) 3.1 What aspects of geometry? where κ denotes the dimensionful constant of pro- The left-hand side of Einstein’s equations com- portionality. Note that no explicit reference to the prises certain measures of curvature. As will be dimension n of spacetime enters (3), so that even explained in detail in Section 5, all curvature in- if n 6= 4 it is usually referred to as Einstein’s equa- formation in dimensions higher than two can be tions. We could have explicitly added a cosmolog- reduced to that of sectional curvature. The sec- ical constant term gΛ on the left-hand side, where tional curvature at a point p ∈ M tangent to Λ is a constant the physical dimension of which is span{X,Y } ⊂ TpM is the Gaussian curvature at the square of an inverse length. However, as long p of the submanifold spanned by the geodesics in as we write down our formulae for general T we M emanating from p tangent to span{X,Y }. The may absorb this term into T where it accounts for Gaussian curvature is defined as the product of two a contribution TΛ = −gΛ/κ. This has to be kept principal curvatures, each being measured in units
4 of an inverse length (the inverse of a principal ra- matrix, which we represent as follows by splitting dius). Hence the Gaussian curvature is measured off terms involving a time component: in units of an inverse length-squared. E −cM~ At each point p in spacetime the Einstein and T = . (11) µν 1 ~ Ricci tensors are symmetric bilinear forms on TpM. − c S Tmn Hence Ein and Ric are determined by the values p p Here all matrix elements refer to the matter’s en- Ein (W, W ) and Ric (W, W ) for all W ∈ T M. p p p ergy momentum distribution relative to the rest By continuity in W this remains true if we restrict frame of the observer who momentarily moves W to the open and dense set of vectors which are along e (i.e. with four-velocity u = ce ) and uses not null, i.e. for which g(W, W ) 6= 0. As we will 0 0 the basis {e , e , e } in his/her rest frame. Then see later on, we then have 1 2 3 E = T00 is the energy density, S~ = (s1, s2, s3) the N X1 (components of the) energy current-density, i.e. en- Ein(W, W ) = −g(W, W ) Sec , (9) ergy per unit surface area and unit time interval, ⊥W M~ the momentum density, and finally Tmn the N X2 (component of the) momentum current-density, i.e. Ric(W, W ) = +g(W, W ) Sec . (10) momentum per unit of area and unit time interval. kW Note that symmetry Tµν = Tνµ implies a simple re- lation between the energy current-density and the For the Einstein tensor the sum on the right-hand 1 momentum density side is over any complete set of N1 = 2 (n−1)(n−2) sectional curvatures of pairwise orthogonal planes S~ = c2 M~ . (12) in the orthogonal complement of W in TpM. For the Ricci tensor it is over any complete set of The remaining relations Tmn = Tnm express equal- N2 = n−1 sectional curvatures of pairwise orthog- ity of the m-th component of the current density onal planes containing W . If W is a timelike unit for n-momentum with the n-th component of the vector representing an observer, Ein(W, W ) is sim- current density for m-momentum. Note that the ply (−ε) times an equally weighted sum of space- two minus signs in front of the mixed components like sectional curvatures, whereas Ric(W, W ) is ε of (11) would have disappeared had we written times an equally weighted sum of timelike sectional down the contravariant components T µν . In flat µ curvatures. In that sense we may say that, e.g., spacetime, the four equations ∂ Tµν express the Ein(W, W ) at p ∈ M measures the mean Gaussian local conservation of energy and momentum. In curvature of the (local) hypersurface in M that is curved spacetime (with vanishing torsion) we have spanned by geodesics emanating from W orthogo- the identity (to be proven later; compare (90b)) nal to W . It, too, is measured in units of the square µ of an inverse length. ∇ Gµν ≡ 0 (13) implies via (7) µ 3.2 What aspects of matter? ∇ Tµν = 0 , (14) Now we turn to the right-hand side of Einstein’s which may be interpreted as expressing a local con- equations. We restrict to four spacetime dimen- servation of energy and momentum for the matter sions, though much of what we say will apply ver- plus the gravitational field, though there is no such batim to other dimensions. The tensor T on the thing as a separate energy-momentum tensor on right-hand side of (3) is the energy-momentum ten- spacetime for the gravitational field. sor of matter. With respect to an orthonormal Several positivity conditions can be imposed basis {e0, e1, ··· , en−1} with timelike e0 the com- on the energy momentum tensor T. The sim- ponents Tµν := T(eµ, eν ) form a symmetric 4 × 4 plest is known as weak energy-condition and reads
5 T(W, W ) ≥ 0 for all timelike W . It is equivalent to 2013) known with a relative standard uncertainty the requirement that the energy density measured of 1.2 × 10−4 and is thus by far the least well by any local observer is non negative. For a per- known of the fundamental physical constants. c = fect fluid of rest-mass density ρ and pressure p the 299 792 458 m · s−1 is the vacuum speed of light weak energy-condition is equivalent to both con- whose value is exact, due to the SI-definition of ditions ρ ≥ 0 and p ≥ −c2ρ. The strong energy- meter (“the meter is the length of the path trav- 1 condition says that T − 2 gTrg(T) (W, W ) ≥ 0 eled by light in vacuum during a time interval of again for all timelike W . This neither follows nor 1/299 792 458 of a second”). implies the weak energy-condition. For a perfect 2 The physical dimension of κ is fluid it is equivalent to both conditions p ≥ −c ρ 2 2 time /(mass · length), that is in SI-units and p ≥ −c ρ/3, i.e. to the latter alone if ρ is s2 · kg−1 · m−1 or m−2/(J · m−3), where J = positive and to the former alone if ρ is negative Joule = kg · m2 · s−2. It converts the common (which is not excluded here). Its significance lies physical dimension of all components Tµν , which in the fact that it ensures attractivity of gravity is that of an energy density (Joule per cubic as described by Einstein’s equations. It must, for meter in SI-units) into that of the components example, be violated if matter is to drive infla- of Ein, which is that of curvature (in dimension tion. Note that upon imposing Einstein’s equa- ≥ 2), i.e., the square of an inverse length (inverse tions the weak and the strong energy-conditions square-meter in SI-units). read Ein(W, W ) ≥ 0 and Ric(W, W ) ≥ 0 respec- tively. From (9) and (10) we can see that for fixed If we express energy density as mass density 2 2 2 W these imply conditions on complementary sets of times c , the conversion factor is κc = 8πG/c . sectional curvatures. For completeness we mention It can be expressed in various units that give a the condition of energy dominance, which states feel for the local “curving power” of mass-densities. 3 −3 that T(W, W ) ≥ |T(X,X)| for any pair of or- For that of water, ρW ≈ 10 kg · m , and nu- thonormal vectors W, X where W is timelike (and clear matter in the core of a neutron star (which hence X is spacelike). It is equivalent to the weak is more than twice that of atomic nuclei), ρN ≈ 17 −3 energy-condition supplemented by the requirement 5 × 10 kg · m , we get, respectively: ] that (iW T) be non spacelike for all timelike W . The second requirement ensures locally measured 1 2 1 2 densities of energy currents and momenta of matter κc2 ≈ · ρ−1 ≈ · ρ−1 , (16) 1.5 AU W 10 km N to be non spacelike.
3.3 How do geometry and matter where AU = 1.5 × 1011 m is the astronomical unit relate quantitatively? (mean Earth-Sun distance). Hence, roughly speak- ing, matter densities of water cause curvature radii We return to Einstein’s equations and finally dis- of the order of the astronomical unit, whereas the cuss the constant of proportionality κ on the right- highest known densities of nuclear matter cause hand side of (3). Its physical dimension is that of curvature radii of tens of kilometers. The curva- curvature (m−2 in SI units) divided by that of en- ture caused by mere mass density is that expressed ergy density (J ·m−3 in SI units, where J = Joule). in Ein(W, W ) when W is taken to be the unit time- It is given by like vector characterizing the local rest frame of the matter: It is a mean of spatial sectional curvatures 8πG m−2 −43 in the matter’s local rest frame. Analogous inter- κ := 4 ≈ 2.1 × 10 −3 , (15) c J · m pretations can be given for the curvatures caused where G ≈ 6.67384(80) × 10−11m3 · kg−1 · s−2 is by momentum densities (energy current-densities) Newton’s constant. It is currently (March and momentum current-densities (stresses).
6 3.4 Conserved energy-momentum isometries. We will discuss general Lie-group ac- tensors and globally conserved tions on manifolds in the Appendix at the end quantities of this contribution, containing detailed proofs of some relevent formulae. But in order not to inter- In this subsection we briefly wish to point out rupt the argument too much, let us recall at this that energy-momentum tensors T whose diver- point that an action of G on M is a map gence vanishes (14) give rise to conserved quanti- ties in case the spacetime (M, g) admits non-trivial Φ: G × M → M, (22a) isometries. We will stress the global nature of these (g, m) 7→ Φ(g, m) = Φg(m) , quantities and clarify their mathematical habitat. Conservation laws for the matter alone result which satisfies in the presence of symmetries, more precisely, if Killing fields for (M, g) exist. Recall that a vector Φe = IdM , (22b) field V is called a Killing field iff LV g = 0, where Φg ◦ Φh = Φgh . (22c) LV is the Lie derivative with respect to V . Recall that the Lie derivative can be expressed in terms of Here e ∈ G denotes the neutral element, IdM the the Levi-Civita covariant derivative with respect to identity map on M, and equation (22c) is valid for g, in which case we get the component expression: any two elements g, h of G. In fact, equation (22c) characterizes a left action. In contrast, for a right action we would have Φ instead of Φ on the (LV g)µν = ∇µVν + ∇ν Vµ = 0 . (17) hg gh right-hand side of (22c). Moreover, as the group We consider the one-form JV that results from acts by isometries for the metric g, we also have ∗ contracting T with V : Φhg = g for all h ∈ G. Now, this action defines a map, V , from Lie(G), µ ν JV := iV T = V Tµν dx . (18) the Lie algebra of G, into the vector fields on M. The vector field corresponding to X ∈ Lie(G) is As a result of Killing’s equation (17) it is divergence denoted V X . Its value at a point m ∈ M is defined free, by ∇ J µ = 0 . (19) d µ V X V (m) := Φ exp(tX), m . (23) This may be equivalently expressed by saying that dt t=0 1 the 3-form ?JV , which is the Hodge dual of the From this it is obvious that V : Lie(G) → ΓT0 M 1-form JV , is closed: is linear. Moreover, one may also show (compare (382b) in Appendix) that this map is a Lie anti- d ? JV = 0 . (20) homomorphism, i.e. that
[X,Y ] X Y Integrating ?JV over some 3-dimensional subman- V = −[V ,V ] . (24) ifold Σ results in a quantity (As shown in the Appendix, a right action would Z have resulted in a proper Lie homomorphism – see Q[V, Σ] := ?JV (21) Σ (382a) –, i.e. without the minus sign on the right- hand side, which however is not harmful.) The left which, because of (20), is largely independent of action of G on M extends to a left action on all Σ. More precisely, if Ω ⊂ M is an oriented domain tensor fields by push forward. In particular, the with boundary ∂Ω = Σ −Σ , then Stokes’ theorem X 1 2 push forward of V by Φg has a simple expression gives Q[V, Σ1] = Q[V, Σ2]. (see (383a) in Appendix) : Suppose now that V arises from a finite- X Adg (X) dimensional Lie group G that acts on (M, g) by Φg∗V = V , (25)
7 where Ad denotes the adjoint representation of nature. In any case, we assume the isometric ac- G on Lie(G). In fact, relation (25) can be di- tion (22) to extend to an action of G on the set of rectly deduced from definition (23). Indeed, writ- matter variables µ, which we denote by µ 7→ Φg∗µ, ing Φ(g, p) = g ·p for notational simplicity, we have like the push-forward on tensor fields. This is also (see Appendix for more explanation) meant to indicate that we assume this to be a left action, i.e. Φg∗ ◦ Φh∗ = Φgh∗. X d We regard the energy-momentum tensor T as a (Φg∗V )(g · p) = g exp(tX) · p dt t=0 map from the space of matter variables to the space d = g exp(tX) g−1g · p of symmetric second-rank covariant tensor fields on dt t=0 M. We require this map to satisfy the following d covariance property: = exp tAdg(X) · (g · p) dt t=0 ∗ = V Adg (X)(g · p) . T[Φg∗µ] = Φg∗T[µ] := Φg−1 T , (28) (26) where Φg∗ is the ordinary push-forward of the ten- sor T. Since we take T to be covariant, its push This leads to (25) which we shall use shortly. forward is the pull back by the inverse diffeomor- Returning to the expression (21) we see that, for phism, as indicated by the second equality in (28). fixed Σ, it becomes a linear map from Lie(G) to : R For each specification µ of matter variables we X X can compute the quantitty Q[V , Σ, µ] as in (21). M : Lie(G) → R , M(X) := Q[V , Σ] . (27) Note that we now indicate the dependence on µ Hence each hypersurface Σ defines an element M ∈ explicitly. We are interested in computing how Q Lie∗(G) in the vector space that is dual to the Lie changes as µ is acted on by g ∈ G. This is done as algebra, given that the integral over Σ converges. follws: This is the case for spacelike Σ and energy momen- Z X tum tensors with spatially compact support (or at Q V , Σ, Φg∗µ = ? iV X T[Φg∗µ] Σ least sufficiently rapid fall off). The same argument Z ∗ as above using Stokes’ theorem and (20) then shows = ? iV X Φg−1 T[µ] that M is independent of the choice of spacelike Σ Z Σ. In other words, we obtain a conserved quantity ∗ = ? Φ −1 i X T[µ] ∗ g Φg−1∗V M ∈ Lie (G) for G-symmetric spacetimes (M, g) Σ Z and covariant divergence free tensors T. ∗ = ? Φ −1 i Ad (X) T[µ] g V g−1 So far we considered a fixed spacetime (M, g) and Σ a fixed energy-momentum tensor T, both linked by Z ∗ = Φ −1 ?i Ad (X) T[µ] Einstein’s equations. In this case the vanishing di- g V g−1 Σ vergence (14) is an integrability condition for Ein- Z = ? i Ad (X) T[µ] stein’s equation and hence automatic. However, V g−1 Φ (Σ) it is also of interest to consider the more general g−1 Ad −1 (X) case where (M, g) is merely a background for some = Q V g , Φg−1 (Σ), µ . matter represented by energy-momentum tensors (29) T[µ], all of which are divergence free (14) with re- spect to the background metric g. Note that we Here we used (28) in the second equality, the gen- ∗ ∗ do not assume (M, g) to satisfy Einstein’s equa- eral formula iV f T = f (if∗V T ) (valid for any dif- tions with any of the T[µ] on the right-hand side. feomorphism f, vector field V , and covariant ten- The µ stands for some matter variables which may sor field T ) in the third equality, (25) in the fourth be fundamental fields and/or of phenomenological equality, the formula ? f ∗F = f ∗ ?F in the fifth
8 equality (valid for any orientation preserving isom- nature does not matter in what follows]. The lin- etry f and any form-field F ; here we assume M ear isometries of (V, η) form the Lorentz group to be oriented), and finally the general formula for Lor ⊂ GL(V ) and the isometries G of (M, g) the integral of the pull back of a form in the sixth can be (non-naturally) identified with the semi- equality. direct product V o Lor, called the Poincar´egroup, Our final assumption is that Q does not de- Poin. Using g we can identify Lie∗(Poin) with pend on which hypersurface Φg(Σ) it is evaluated V ⊕(V ∧V ). The co-adjoint action of (a, A) ∈ Poin on. Since we assume (14) this is guaranteed if all on (f, F ) ∈ Lie∗(Poin) is then given by Φg(Σ) are in the same homology class or, more ∗ generally, if any two hypersurfaces Σ and Φg(Σ) Ad(a,A)(f, F ) = Af , (A ⊗ A)F − a ∧ Af . (32) are homologous to hypersurfaces in the comple- ment of the support of T. A typical situation Note that, e.g., the last term on the right hand side arising in physical applications is that of a source includes the law of change of angular momentum T[µ] with spatially compact support; then any two under spatial translations. In contrast, the adjoint sufficiently extended spacelike slices through the representation on Lie(Poin), the latter also identi- timelike support-tube of T[µ] is homologous to fied with V ⊕ (V ∧ V ), is given by the timelike cylindrical hypersurface outside this support-tube. In this case we infer from (29) that Ad(a,A)(f, F ) = Af − (A ⊗ A)F a , (A ⊗ A)F ,
X Ad −1 (X) (33) Q V , Σ, Φg∗µ e = Q V g , Σ, µ . (30) where the application of an element in V ∧ V Recall from (27) that for fixed Σ and T we have to an element in V is given by (u ∧ v)(w) := M ∈ Lie∗(G). Given the independence on Σ and u g(v, w) − v g(u, w), and linear extension. Note the depencence of T on µ, we now regard M as a the characteristic difference between (32) and (33), map from the matter variables µ to Lie∗(G). This which lies in the different actions of the subgroup map may be called the momentum map. (Compare of translations, whereas the subgroup of Lorentz the notion of a momentum map in Hamiltonian transformations acts in the same fashion. Physical mechanics; cf. Section 7.) Equation (30) then just momenta transform as in (32), as already exempli- states the Ad∗-equivariance of the momentum map: fied by the non-trivial transformation behavior of angular momentum under spatial translations. For ∗ M ◦ Φg∗ = Adg ◦ M . (31) a detailed discussion of the proper group-theoretic setting and the adjoint and co-adjoint actions, see Here Ad∗ denote the co-adjoint representation of ∗ the recent account [75]. G on Lie(G), which is defined by Adg(α) = α ◦ Adg−1 . From all this we see that the conserved “momentum” that we obtain by evaluating M on 4 Spacetime decomposition the matter configuration µ is a conserved quantity that is globally associated to all of spacetime, not In this section we explain how to decompose a given a particular region or point of it. It is an element spacetime (M, g) into “space” and ‘time”. For this ∗ of the vector space Lie (G) which carries the co- to be possible we need to make the assumption that adjoint representation of the symmetry group G. M is diffeomorphic to the product of the real line R and some 3-manifold Σ: In particular this applies to Special Relativ- ity, where M is the four-dimensional real affine M =∼ R × Σ . (34) space with associated (four-dimensional real) vec- tor space V and g a bilinear, symmetric, non- This will necessarily be the case for globally hy- degenerate form of signature (−, +, +, +) [the sig- perbolic spacetimes, i.e. spacetimes admitting a
9 Cauchy surface [65]. We assume Σ to be orientable, M Σs0 for, if it were not, we could take the orientable Es0 double cover of it instead. Orientable 3-manifolds Σ Es Σs are always parallelizable [111] , i.e. admit three E 00 globally defined and pointwise linearly independent s Σs00 vector fields. This is equivalent to the triviality of the tangent bundle. In the closed case this is known as Stiefel’s theorem (compare [100], problem 12- Figure 1: Spacetime M is foliated by a one-parameter B) and in the open case it follows, e.g., from the family of spacelike embeddings of the 3-manifold Σ. Here well known fact that every open 3-manifold can the image Σs0 of Σ under Es0 lies to the future (above) and 00 0 3 Σ 00 to the past (below) of Σ if s < s < s . ‘Future’ and be immersed in R [117]. Note that orientabil- s s 2 1 ‘past’ refer to the time function t which has so far not been ity is truly necessary; e.g., RP × S is not par- given any metric significance. allelizable. Since Cartesian products of paralleliz- able manifolds are again parallelizable, it follows We distinguish between the abstract 3-manifold that a 4-dimensional product spacetime (34) is also Σ and its image Σs in M. The latter is called the parallelizable. This does, of course, not generalize leaf corresponding to the value s ∈ R. Each point to higher dimensions. Now, for non-compact four- in M is contained in precisely one leaf. Hence there dimensional spacetimes it is known from [64] that is a real valued function t : M → R that assigns to parallelizability is equivalent to the existence of a each point in M the parameter value of the leaf it spin structure, without which spinor fields could lies on: not be defined on spacetime. So we see that the t(p) = s ⇔ p ∈ Σ . (36) existence of spin structure is already implied by s (34) and hence does not pose any further topolog- So far this is only a foliation of spacetime by 3- ical restriction. Note that the only other potential dimensional leaves. For them to be addressed as topological restriction at this stage is that imposed “space” the metric induced on them must be posi- from the requirement that a smooth Lorentz metric tive definite, that is, the leaves should be spacelike is to exist everywhere on spacetime. This is equiv- submanifolds. This means that the one-form dt is alent to a vanishing Euler characteristic (see, e.g., timelike: § 40 in [111]) which in turn is equivalently to the g−1(dt, dt) < 0 . (37) global existence of a continuous, nowhere vanish- ing vector field (possibly up to sign) on spacetime. The normalized field of one-forms is then But such a vector field clearly exists on any Carte- sian product with one factor being . We conclude [ dt R n := p . (38) that existence of a Lorentz metric and a spin struc- −g−1(dt, dt) ture on an orientable spacetime M = R × Σ pose no restrictions on the topology of an orientable Σ. As explained in section 2, we write n[ since we think As we will see later on, even Einstein’s equation of this one form as the image under g of the nor- poses no topological restriction on Σ, in the sense malized vector field perpendicular to the leaves: that some (physically reasonable) solutions to Ein- stein’s equations exist for any given Σ. Topological n[ = g(n, · ) . (39) restrictions may occur, however, if we ask for solu- tion with special properties (see below). The linear subspace of vectors in TpM which are k Now, given Σ, we consider a one-parameter fam- tangent to the leaf through p is denoted by Tp M; ily of embeddings hence
k Es :Σ → M, Σs := Es(Σ) ⊂ M. (35) Tp M := {X ∈ TpM : dt(X) = 0} . (40)
10 The orthogonal complement is just the span of n at For example, letting the horizontal projection of ⊥ p, which we denote by Tp M. This gives, at each the form ω act on the vector X, we get point p of M, the g-orthogonal direct sum k k ] [ P∗ ω(X) = (P ω ) (X) T M = T ⊥M ⊕ T kM. (41) p p p = g P kω],X (47) and associated projections (we drop reference to = g ω],P kX the point p) = ω P kX , P ⊥ :TM → T ⊥M, where we merely used the definitions (1) of [ and ] X 7→ ε g(X, n) n , (42a) in the second and fourth equality, respectively, and k P k : TM → T kM, the self-adjointness (44b) of P in the third equal- ity. The analogous relation holds for P ⊥ω(X). It X 7→ X − εg(X, n) n . (42b) ∗ k ⊥ is also straightforward to check that P∗ and P∗ As already announced in Section 2, we introduced are self-adjoint with respect to g−1 (cf. (2)). the symbol Having the projections defined for vectors and ε = g(n, n) (43) co-vectors, we can also define it for the whole ten- sor algebra of the underlying vector space, just in order to keep track of where the signature mat- by taking the appropriate tensor products of these ters. Note that the projection operators (42) are k k maps. All tensor products between P and P∗ will self-adjoint with respect to g, so that for all X,Y ∈ k TM we have then, for simplicity, just be denoted by P , the action on the tensor being obvious. Similarly for g P ⊥X,Y = g X,P ⊥Y , (44a) P ⊥. (For what follows we need not consider mixed projections.) The projections being pointwise op- g P kX,Y = g X,P kY . (44b) erations, we can now define vertical and horizontal A vector is called horizontal iff it is in the kernel projections of arbitrary tensor fields. Hence a ten- u of P ⊥, which is equivalent to being invariant under sor field T ∈ ΓTd M is called horizontal if and only k P k. It is called vertical iff it is in the kernel of P k, if P T = T . The space of horizontal tensor fields ku ⊥ which is equivalent to being invariant under P . of rank (u,d) is denoted by ΓT d M. All this can be extended to forms. We define As an example, the horizontal projection of the vertical and horizontal forms as those annihilating metric g is horizontal and vertical vectors, respectively: h := P kg := g P k · ,P k · = g − εn[ ⊗ n[ . (48) ∗⊥ ∗ k Tp M := {ω ∈ Tp M : ω(X) = 0 , ∀X ∈ Tp M} , k0 (45a) Hence h ∈ ΓT 2 M. Another example of a horizon- ∗k ∗ ⊥ tal vector field is the “acceleration” of the normal Tp M := {ω ∈ Tp M : ω(X) = 0 , ∀X ∈ Tp M} . field n: (45b) a := ∇nn . (49) Using the ‘musical’ isomorphisms (1), the self- Here ∇ denotes the Levi-Civita covariant derivative adjoint projection maps (42) on vectors define self- with respect to g. An observer who moves perpen- adjoint projection maps on co-vectors (again drop- dicular to the horizontal leaves has four-velocity ping the reference to the base-point p) u = cn and four-acceleration c2a. If L denotes the Lie derivative, it is easy to show that the accelera- ⊥ ⊥ ∗ ∗⊥ P∗ :=[ ◦ P ◦ ] : T M → T M, (46a) tion 1-form satisfies k k ∗ ∗k P∗ := [ ◦ P ◦ ] : T M → T M. (46b) [ [ a = Lnn . (50)
11 Moreover, as n is hypersurface orthogonal it is ir- representative of space. Instead of using the folia- rotational, hence its 1-form equivalent satisfies tion by 3-dimensional spatial leaves (35) we could have started with a foliation by timelike lines, plus [ [ dn ∧ n = 0 , (51a) the condition that these lines are vorticity free. which is equivalent to the condition of vanishing These two concepts are equivalent. Depending on horizontal curl: the context, one might prefer to emphasize one or the other. k [ P dn = 0 . (51b) The vector parallel to the worldline at p = Es(q) is, as usual in differential geometry, defined by its Equation (51a) can also be immediately inferred action on f ∈ C∞(M) (smooth, real valued func- directly from (38). Taking the operation in ◦ d (ex- tions): terior derivative followed by contraction with n) as ∂ df(E 0 (q)) f = s . (54) well as the Lie derivative with respect to n of (50) 0 0 ∂t Es(q) ds s =s shows da[ ∧ n[ = 0 , (52a) At each point this vector field can be decomposed into its horizontal component that is tangential to an equivalent expression being again the vanishing the leaves of the given foliation and its normal com- of the horizontal curl of a: ponent. We write k [ P da = 0 . (52b) 1 ∂ = α n + β , (55) This will be useful later on. c ∂t [ Note that a is a horizontal co-vector field, i.e. an where β is the tangential part; see Figure 2. The ku=0 element of ΓT d=1 M. More generally, for a purely covariant horizontal tensor field we have the follow- p0 ing results, which will also be useful later on: Let Σ k0 s+ds T ∈ ΓT d M, then 1 ∂ k c ∂t αn P LnT = LnT, (53a) L T = fL T, (53b) fn n p Σs for all f ∈ C∞(M). Note that (53a) states that the ‘ β Lie derivative in normal direction of a horizontal covariant tensor field is again horizontal. That this Figure 2: For fixed q ∈ Σ its image points p = Es(q) 0 is not entirely evident follows, e.g., from the fact and p = Es+ds(q) for infinitesimal ds are connected by the that a corresponding result does not hold for T ∈ vector ∂/∂t|p, whose components normal to Σs are α (one ku function, called lapse) and β (three functions, called shift) ΓT d M where u > 0. The proofs of (53) just use respectively. standard manipulations. A fixed space-point q ∈ Σ defines the worldline real-valued function α is called the lapse (function) (history of that point) R 3 s 7→ Es(q). The col- and the horizontal vector field β is called the shift lection of all worldlines of all space-points define a (vector-field) . foliation of M into one-dimensional timelike leafs. Each leaf is now labeled uniquely by a space point. 4.1 Decomposition of the metric We can think of “space”, i.e., the abstract mani- 0 fold Σ, as the quotient M/∼, where p ∼ p iff both Let {e0, e1, e2, e3} be a locally defined orthonor- 0 1 2 3 points lie on the same worldline. As any Σs in- mal frame with dual frame {θ , θ , θ , θ }. We call 0 [ tersects each worldline exactly once, each Σs is a them adapted to the foliation if e0 = n and θ = n .
12 0 1 2 3 A local coordinate system {x , x , x , x } is called Orthogonality of the ea implies for the chart adapted if ∂/∂xa are horizontal for a = 1, 2, 3. Note components of the spatial metric (48) that in the latter case ∂/∂x0 is not required to be 3 orthogonal to the leaves (i.e. it need not be par- X h := h ∂/∂xm, ∂/∂xn = Aa Aa , (62) allel to n). For example, we may take x0 to be mn m n a=1 proportional to t; say x0 = ct. In the orthonormal co-frame the spacetime met- and its inverse ric, i.e. the field of signature (ε, +, +, +) metrics 3 mn −1 m n X −1 m −1 n in the tangent spaces, has the simple form h := h dx , dx = [A ]a [A ]a . (63) a=1 3 X g = εθ0 ⊗ θ0 + θa ⊗ θa . (56) Inserting (60) into (56) and using (62) leads to a=1 the (3+1)-form of the metric in adapted coordi- The inverse spacetime metric, i.e. the field of signa- nates ture (ε, +, +, +) metrics in the co-tangent spaces, g = εα2 + h(β, β)c2 dt ⊗ dt has the form m m + cβm dt ⊗ dx + dx ⊗ dt (64) 3 m n −1 X + hmn dx ⊗ dx , g = εe0 ⊗ e0 + ea ⊗ ea . (57) a=1 n [ where βm := hmnβ are the components of β := The relation that expresses the coordinate basis g(β, · ) = h(β, · ) with respect to the coordinate ba- in terms of the orthonormal basis is of the form (in sis {∂/∂xm}. Likewise, inserting (61) into (57) and a self-explanatory matrix notation) using (63) leads to the (3+1)-form of the inverse metric in adapted coordinates (we write ∂t := ∂/∂t 0 a m ∂/∂x α β e0 and ∂m := ∂/∂x for convenience) m = a , (58) ∂/∂x 0 A ea m −1 −2 −2 g = εc α ∂t ⊗ ∂t a where β are the components of β with respect to −1 −2 m − εc α β ∂t ⊗ ∂m + ∂m ⊗ ∂t (65) the horizontal frame basis {ea}. The inverse of (58) mn m n is + h + εβ β ∂m ⊗ ∂n . −1 −1 m 0 e0 α −α β ∂/∂x Finally we note that the volume form on space- = −1 m m , (59) ea 0 [A ]a ∂/∂x time also easily follows from (60) m 0 1 2 3 where β are the components of β with respect dµg = θ ∧ θ ∧ θ ∧ θ to the horizontal coordinate-induced frame basis (66) = αpdet{h } cdt ∧ d3x , {∂/∂xm}. mn The relation for the co-bases dual to those in (58) where we use the standard shorthand d3x = dx1 ∧ is given by the transposed of (58), which we write dx2 ∧ dx3. as: a 0 a 0 m α β 4.2 Decomposition of the θ θ = dx dx a . (60) 0 Am covariant derivative The inverse of that is the transposed of (59): Given horizontal vector fields X and Y , the covari- ant derivative of Y with respect to X need not be −1 −1 m 0 m 0 a α −α β horizontal. Its decomposition is written as dx dx = θ θ −1 m . 0 [A ]a (61) ∇X Y = DX Y + nK(X,Y ) , (67)
13 where field. Symmetry follows from the vanishing torsion of ∇, since then k DX Y := P ∇X Y, (68) K(X,Y ) = ε g(n, ∇X Y ) K(X,Y ) := ε g(n, ∇X Y ) . (69) = ε g(n, ∇Y X + [X,Y ]) The map D defines a covariant derivative (in the (74) = ε g(n, ∇ X) sense of Kozul; compare [110], Vol 2) for horizon- Y tal vector fields, as a trivial check of the axioms = K(Y,X) reveals. Moreover, since the commutator [X,Y ] for horizontal X,Y . From (69) one sees that of two horizontal vector fields is always horizontal K(fX, Y ) = fK(X,Y ) for any smooth function (since the horizontal distribution is integrable by f. Hence K defines a unique symmetric tensor construction), we have field on M by stipulating that it be horizontal, i.e. D K(n, ·) = 0. It is called the extrinsic curvature of T (X,Y ) = DX Y − DY X − [X,Y ] the foliation or second fundamental form, the first k = P ∇X Y − ∇Y X − [X,Y ] (70) fundamental form being the metric. From (69) = 0 and the symmetry just shown one immediately in- fers the alternative expressions due to ∇ being torsion free. We recall that tor- 1 sion is a tensor field T ∈ ΓT2 M associated to each K(X,Y ) = −ε g(∇X n, Y ) = −ε g(∇Y n, X) . covariant derivative ∇ via (75) This shows the relation between the extrinsic cur- ∇ T (X,Y ) = ∇X Y − ∇Y X − [X,Y ] . (71) vature and the Weingarten map, Wein, also called the shape operator, which sends horizontal vectors We have T (X,Y ) = −T (Y,X). As usual, even to horizontal vectors according to though the operations on the right hand side of (71) involve tensor fields (we need to differentiate), X 7→ Wein(X) := ∇X n . (76) the result of the operation just depends on X and Y pointwise. This one proves by simply checking the Horizontality of ∇X n immediately follows from n 1 validity of T (fX, Y ) = fT (X,Y ) for all smooth being normalized: g(n, ∇X n) = 2 X g(n, n) = 0. functions f. Hence (70) shows that D is torsion Hence (75) simply becomes free because ∇ is torsion free. K(X,Y ) = −ε h Wein(X),Y Finally, we can uniquely extend D to all hori- (77) zontal tensor fields by requiring the Leibniz rule. = −ε h X, Wein(Y ) , Then, for X,Y,Z horizontal where we replaced g with h—defined in (48)— (DX h)(Y,Z) since both entries are horizontal. It says that K is (−ε) times the covariant tensor corresponding = X h(Y,Z) − h(DX Y,Z) − h(Y,DX Z) (72) to the Weingarten map, and that the symmetry = X g(Y,Z) − g(∇X Y,Z) − g(Y, ∇X Z) of K is equivalent to the self-adjointness of the = (∇X g)(Y,Z) = 0 Weingarten map with respect to h. The Wein- garten map characterizes the bending of the em- due to the metricity, ∇g = 0, of ∇. Hence D is bedded hypersurface in the ambient space by an- metric in the sense swering the following question: In what direction Dh = 0 . (73) and by what amount does the normal to the hy- persurface tilt if, starting at point p, you progress The map K from pairs of horizontal vector fields within the hypersurface by the vector X. The an- (X,Y ) into functions define a symmetric tensor swer is just Weinp(X). Self adjointness of Wein
14 then means that there always exist three (n − 1 assume a minimal and a maximal value, denoted in general) perpendicular directions in the hyper- by kmin(p) = k(p, vmin) and kmax(p) = k(p, vmax) surface along which the normal tilts in the same respectively. These are called the principal curva- direction. These are the principal curvature direc- tures of S at p and their reciprocals are called the tions mentioned above. The principal curvatures principal radii. It is clear that the principal direc- are the corresponding eigenvalues of Wein. tions vmin and vmax just span the eigenspaces of Finally we note that the covariant derivative of the Weingarten map discussed above. In particu- the normal field n can be written in terms of the lar, vmin and vmax are orthogonal. The Gaussian acceleration and the Weingarten map as follows curvature K(p) of S at p is then defined to be the product of the principal curvatures: ∇n = εn[ ⊗ a + Wein . (78) K(p) = kmin(p) · kmax(p) . (81) Recalling (77), the purely covariant version of this is This definition is extrinsic in the sense that essen- [ [ [ ∇n = −ε K − n ⊗ a . (79) tial use is made of the ambient R3 in which S is em- From (48) and (79) we derive by standard manip- bedded. However, Gauss’ theorema egregium states ulation, using vanishing torsion, that this notion of curvature can also be defined in- trinsically, in the sense that the value K(p) can be
Lnh = −2εK . (80) obtained from geometric operations entirely car- ried out within the surface S. More precisely, it is In presence of torsion there would be an addi- a function of the first fundamental form (the met- [ tional term +2(inT )s, where the subscript s de- ric) only, which encodes the intrinsic geometry of [ notes symmetrization; in coordinates [(inT )s]µν = S, and does not involve the second fundamental λ α n Tλ(µgν)α. form (the extrinsic curvature), which encodes how S is embedded into R3. Let us briefly state Gauss’ theorem in mathemat- 5 Curvature tensors ical terms. Let
We wish to calculate the (intrinsic) curvature ten- a b g = gab dx ⊗ dx (82) sor of ∇ and express it in terms of the curvature tensor of D, the extrinsic curvature K, and the be the metric of the surface in some coordinates, spatial and normal derivatives of n and K. Be- and fore we do this, we wish to say a few words on the c 1 cd definition of the curvature measures in general. Γab = 2 g −∂dgab + ∂agbd + ∂bgda , (83) All notions of curvature eventually reduce to that 3 certain combinations of first derivatives of the met- of curves. For a surface S embedded in R we have the notion of Gaussian curvature which comes ric coefficients, known under the name of Christof- c about as follows: Consider a point p ∈ S and a unit fel symbols . Note that Γab has as many indepen- vector v at p tangent to S. Consider all smooth dent components as ∂agbc and that we can calculate curves passing through p with unit tangent v. It the latter from the former via is easy to see that the curvatures at p of all such ∂ g = g Γn + g Γn . (84) curves is not bounded from above (due to the pos- c ab an bc bn ac sibility to bend within the surface), but there will Next we form even more complicated combinations be a lower bound, k(p, v), which just depends on of first and second derivatives of the metric coeffi- the chosen point p and the tangent direction repre- cients, namely sented by v. Now consider k(p, v) as function of v. a a a a n a n As v varies over all tangent directions k(p, v) will R b cd = ∂cΓdb − ∂dΓcb + ΓcnΓdb − ΓdnΓcb , (85)
15 which are now known as components of the Rie- From (88) and using (71) one may show that the mann curvature tensor. From them we form the Riemann tensor always obeys the first and second totally covariant (all indices down) components: Bianchi identities:
n X Rab cd = ganR b cd . (86) R(X,Y )Z (XYZ) They are antisymmetric in the first and second X n o = (∇ T )(Y,Z) − T X,T (Y,Z) , index pair: Rab cd = −Rba cd = −Rab dc, so that X (XYZ) R12 12 is the only independent component. Gauss’ theorem now states that at each point on S we have (89a) X R12 12 (∇X R)(Y,Z) K = 2 . (87) (XYZ) g11g22 − g 12 X = R X,T (Y,Z) , (89b) An important part of the theorem is to show that the right-hand side of (87) actually makes good (XYZ) geometric sense, i.e. that it is independent of the where the sums are over the three cyclic permu- coordinate system that we use to express the coef- tations of X, Y , and Z. For zero torsion these ficients. This is easy to check once one knows that identities read in component form: Rabcd are the coefficients of a tensor with the sym- X α metries just stated. In this way the curvature of R λ µν = 0 , (90a) a surface, which was primarily defined in terms of (λµν) curvatures of certain curves on the surface, can be X α ∇λR β µν = 0 . (90b) understood intrinsically. In what follows we will see (λµν) that the various measures of intrinsic curvatures of n-dimensional manifolds can be reduced to that of The second traced on (α, µ) and contracted with 2-dimensional submanifolds, which will be called gβν yields (−2) times (13). sectional curvatures. The covariant Riemann tensor is defined by Back to the general setting, we start from the Riem(W, Z, X, Y ) := g W, R(X,Y )Z . (91) notion of a covariant derivative ∇. Its associated curvature tensor is defined by For general covariant derivatives its only symme- try is the antisymmetry in the last pair. But for R(X,Y )Z = ∇X ∇Y − ∇Y ∇X − ∇[X,Y ] Z. (88) special choices it acquires more. In standard GR For each point p ∈ M it should be thought of we assume the covariant derivative to be metric compatible and torsion free: as a map that assigns to each pair X,Y ∈ TpM of tangent vectors at p a linear map R(X,Y ): ∇g = 0 , (92) T M → T M. This assignment is antisymmetric, p p T = 0 . (93) i.e. R(X,Y ) = −R(Y,X). If R(X,Y ) is applied to Z the result is given by the right-hand side of In that case the Riemann tensor has the symme- (88). Despite first appearance, the right-hand side tries of (88) at a point p ∈ M only depends on the values of X,Y , and Z at that point and hence de- Riem(W, Z, X, Y ) = −Riem(W, Z, Y, X) , (94a) fines a tensor field. This one again proves by show- Riem(W, Z, X, Y ) = −Riem(Z, W, X, Y ) , (94b) ing the validity of R(fX, Y )Z = R(X, fY )Z = Riem(W, X, Y, Z) + Riem(W, Y, Z, X) + R(X,Y )fZ = fR(X,Y )Z for all smooth real- Riem(W, Z, Y, X) = 0 , (94c) valued functions f on M. In other words: All terms involving derivatives of f cancel. Riem(W, Z, X, Y ) = Riem(X, Y, W, Z) . (94d)
16 Equation (94a) is true by definition (88), (94b) is Here X,Y is a pair of linearly independent tangent equivalent to metricity of ∇, and (94c) is the first vectors that span a 2-dimensional tangent subspace Bianchi identity in case of zero torsion. The last restricted to which g is non-degenerate. We will symmetry (94d) is a consequence of the preceding say that span{X,Y } is non-degenerate. This is the three. Together (94a), (94b), and (94d) say that, necessary and sufficient condition for the denomi- at each point p ∈ M, Riem can be thought of nator on the right-hand side to be non zero. The as symmetric bilinear form on the antisymmetric quantity Sec(X,Y ) is called the sectional curva- tensor product TpM ∧ TpM. The latter has di- ture of the manifold (M, g) at point p tangent to 1 mension N = 2 n(n − 1) if M has dimension n, span{X,Y }. From the symmetries of Riem it is and the space of symmetric bilinear forms has di- easy to see that the right-hand side of (99) does 1 mension 2 N(N + 1). From that number we have indeed only depend on the span of X,Y . That is, to subtract the number of independent conditions for any other pair X0,Y 0 such that span{X0,Y 0} = n 0 0 (94c), which is 4 in dimensions n ≥ 4 and zero span{X,Y }, we have Sec(X ,Y ) = Sec(X,Y ). otherwise. Indeed, it is easy to see that (94c) is The geometric interpretation of Sec(X,Y ) is as identically satisfied as a consequence of (94a) and follows: Consider all geodesics of (M, g) that pass (94b) if any two vectors W, Z, X, Y coincide (pro- through the considered point p ∈ M in a direc- portionality is sufficient). Hence the number # of tion tangential to span{X,Y }. In a neighborhood independent components of the curvature tensor is of p they form an embedded 2-surface in M whose Gaussian curvature is just Sec(X,Y ). #Riem = Now, Riem is determined by components of the 1 N(N + 1) − n = 1 n2(n2 − 1) for n ≥ 4 form Riem(X,Y,X,Y ), as follows from the fact 2 4 12 that Riem is a symmetric bilinear form on TM ∧ 6 for n = 3 TM. This remains true if we restrict to those X,Y 1 for n = 2 whose span is non-degenerate, since they lie dense 1 2 2 in TM ∧ TM and (X,Y ) 7→ Riem(X,Y,X,Y ) is = 12 n (n − 1) for all n ≥ 2 . (95) continuous. This shows that the full information of the Riemann tensor can be reduced to certain The Ricci and scalar curvatures are obtained Gaussian curvatures. by taking traces with respect to g: Let {e1, ··· , en} This also provides a simple geometric interpre- be an orthonormal basis, g(ea, eb) = δabεa (no sum- tation of the scalar and Einstein curvatures in mation) with εa = ±1, then terms of sectional curvatures. Let {X1, ··· ,Xn} be any set of pairwise orthogonal non-null vec- n 1 X tors. The 2 n(n − 1) 2-planes span{Xa,Xb} are Ric(X,Y ) = εa Riem(ea, X, ea,Y ) (96) non-degenerate and also pairwise orthogonal. It a=1 then follows from (97) and (99) that the scalar cur- n X vature is twice the sum of all sectional curvatures: Scal = εa Ric(ea, ea) . (97) n a=1 X Scal = 2 Sec(Xa,Xb) . (100) The Einstein tensor is a,b=1 a
1 Ein = Ric − 2 Scal g . (98) The sum on the right-hand side of (100) is the same 1 for any set of 2 n(n − 1) non-degenerate and pair- The sectional curvature is defined by wise orthogonal 2-planes. Hence the scalar curva- ture can be said to be twice the sum of mutually or- Riem(X,Y,X,Y ) thogonal sectional curvatures, or n(n−1) times the Sec(X,Y ) = 2 , (99) g(X,X)g(Y,Y ) − g(X,Y ) mean sectional curvature. Similarly for the Ricci
17 and Einstein curvatures. The symmetry of the their Kulkarni-Nomizu product is defined by Ricci and Einstein tensors imply that they are fully determined by their components Ric(W, W ) and k `(X1,X2,X3,X4) := k(X1,X3) `(X2,X4) ? Ein(W, W ). Again this remains true if we restrict + k(X2,X4) `(X1,X3) to the dense set of non-null W , i.e. g(W, W ) 6= 0. − k(X1,X4) `(X2,X3) Let now {X1, ··· ,Xn−1} be any set of mutually orthogonal vectors (again they need not be nor- − k(X2,X3) `(X1,X4) , malized) in the orthogonal complement of W . As (103) before the 1 (n − 1)(n − 2) planes span{X ,X } 2 a b or in components are non degenerate and pairwise orthogonal. From (96), (98), and (99) it follows that (k `)abcd = kac`bd+kbd`ac−kad`bc−kbc`ad . (104) ? n−1 X The Weyl tensor, Weyl, is of the same type as Ric(W, W ) = g(W, W ) Sec(W, Xa) (101) Riem but in addition totally trace-free. It is ob- a=1 tained from Riem by a projection map, PW , given by and Weyl := PW (Riem) n−1 1 1 X := Riem − n−2 Ric − 2(n−1) Scal g g . Ein(W, W ) = −g(W, W ) Sec(Xa,Xb) . ? a,b=1 (105) ascalar curvature denoted by an encircled wedge, , which is a bi- ? 1 linear symmetric product on the space of covari- Riem = Ric − 4 Scal g g (for n = 3) . (107) ant symmetric rank-two tensors with values in the ? covariant rank-four tensors that have the symme- A metric manifold (M, g) is said to be of constant tries (94) of the Riemann tensor. Let k and ` be curvature if two symmetric covariant second-rank tensors, then Riem = k g g (108) ?
18 for some function k. Then Ric = 2k(n − 1)g and a different meaning from that given to it in (48). Ein = −k(n − 1)(n − 2)g. We recall that mani- We recall that the Levi-Civita covariant derivative folds (M, g) for which the Einstein tensor (equiv- is uniquely determined by the metric. For ∇ this alently, the Ricci tensor) is pointwise proportional reads to the metric are called Einstein spaces. The twice contracted second Bianchi identity (13) shows that 2 g(∇X Y,Z) k must be a constant unless n = 2. For n = 3 = X g(Y,Z) + Y g(Z,X) − Z g(X,Y ) equation (107) shows that Einstein spaces are of − g X, [Y,Z])] + g Y, [Z,X])] + g Z, [X,Y ])]. constant curvature. (112)
5.1 Comparing curvature tensors Subtracting (112) from the corresponding formula with ∇ and g replaced by ∇ˆ andg ˆ yields, using Sometimes one wants to compare two different cur- T = 0, vature tensors belonging to two different covariant ˆ derivatives ∇ and ∇. In what follows, all quan- 2g ˆ ∆(X,Y ),Z = tities referring to ∇ˆ carry a hat. Recall that a covariant derivative can be considered as a map − (∇Z h)(X,Y ) + (∇X h)(Y,Z) + (∇Y h)(Z,X). 1 1 1 (113) ∇ :ΓT0 M × ΓT0 M → ΓT0 M,(X,Y ) 7→ ∇X Y , which is C∞(M)-linear in the first and a deriva- This formula expresses ∆ as functional of g andg ˆ. tion in the second argument. That is, for f ∈ There are various equivalent forms of it. We have C∞(M) have ∇ Z = f∇ Z + ∇ Z and fX+Y X Y chosen a representation that somehow minimizes ∇ (fY + Z) = X(f)Y + f∇ Y + ∇ Z. This X X X the appearance ofg ˆ. Note that g enters in h as implies that the difference of two covariant deriva- well as ∇, whereasg ˆ enters in h and via the scalar tives is C∞(M)- linear also in the second argument product on the left-hand side. The latter obstructs and hence a tensor field: expressing ∆ as functional of g and h alone. In ˆ 1 ∇ − ∇ =: ∆ ∈ ΓT2 M. (109) components (113) reads
Replacing ∇ˆ with ∇ + ∆ in the definition of the a 1 an ∆ = gˆ −∇nhbc + ∇bhcn + ∇chnb . (114) curvature tensor for ∇ˆ according to (88) directly bc 2 leads to Note that one could replace the components of h Rˆ(X,Y )Z = R(X,Y )Z with those ofg ˆ = g + h in the bracket on the right- hand side, since the covariant derivatives of g van- + (∇ ∆)(Y,Z) − (∇ ∆)(X,Z) X Y ish. + ∆ X, ∆(Y,Z) − ∆ Y, ∆(X,Z) Now suppose we consider h and its first and sec- + ∆ T (X,Y ),Z) . ond derivatives to be small and we wanted to know (110) the difference in the covariant derivatives and cur- vature only to leading (linear) order in h. To that Note that so far no assumptions have been made order we may replaceg ˆ with g on the left-hand side ˆ concerning torsion or metricity of ∇ and ∇. This of (113) and the right-hand side of (114). Moreover formula is generally valid. In the special case where we may neglect the ∆-squared terms in (110) and ˆ ∇ and ∇ are the Levi-Civita covariant derivatives obtain, writing δR for the first order contribution with respect to two metricsg ˆ and g, we set to Rˆ − R,
h :=g ˆ − g , (111) a a a δR bcd = ∇c∆db − ∇d∆cb . (115) which is a symmetric covariant tensor field. Note that here, and for the rest of this subsection, h has From this the first-order variation of the Ricci ten-
19 sor follows, writing hab =: δgab, tensor, which is related to the covariant form,
n n Weyl, in the same way (91) as the curvature ten- δRab = ∇n∆ab − ∇b∆na sor R is related to Riem (i.e., by raising the first 1 = 2 −∆g δgab − ∇a∇b δg (116) index of the latter). n n From (120) we also deduce the transformation + ∇ ∇a δgnb + ∇ ∇b δgna , properties of the Ricci tensor: ab ab where ∆g := g ∇a∇b and δg = g δgab. Finally, the variation of the scalar curvature is (note δgab = Ricgˆ =Ricg ac bd ab −1 −g g δgcd = −h ) − ∆gφ + (n − 2)g (dφ, dφ) g (122)
ab a − (n − 2)(∇∇φ − dφ ⊗ dφ) . δR = Rab δg + ∇aU , (117a) where, as above, ∆ denotes again the Lapla- where g cian/d’Alembertian for g. Finally, for the scalar a nm a an m U = g ∆nm − g ∆mn curvature we get abcd (117b) = G ∇b δgcd . −2φ Scalgˆ = e Scalg Here we made use of the De Witt metric, which de- fines a symmetric non-degenerate bilinear form on − 2(n − 1)∆gφ the space of symmetric covariant rank-two tensors and which in components reads: − (n − 1)(n − 2)g−1(dφ, dφ) .
abcd 1 ac bd ad bc ab cd (123) G = 2 g g + g g − 2g g . (118) We will later have to say more about it. This law has a linear dependence on the second and We also wish to state a useful formula that com- a quadratic dependence on first derivatives of φ. If pares the curvature tensors for conformally related the conformal factor is written as an appropriate metrics, i.e. power of some positive function Ω : M → R+ we gˆ = e2φ g , (119) can eliminate all dependence on first and just retain the second derivatives. In n > 2 dimensions it is where φ : M → R is smooth. Then easy to check that the rule is this:
2φh i 2φ 4 Riemgˆ = e Riemg + g K , (120a) e = Ω n−2 , (124) ? with then (123) becomes K := −∇2φ + dφ ⊗ dφ − 1 g−1(dφ, dφ) g . (120b) 2 4(n − 1) − n+2 Scalgˆ = − Ω n−2 DgΩ , (125a) (This can be proven by straightforward calculations n − 2 using either (88) and (112), or Cartan’s structure where equations, or, most conveniently, normal coordi- n − 2 nates.) From (120a) and the fact that the kernel D = ∆ − Scal . (125b) g g 4(n − 1) g of the map PW in (105) is given by tensors of the form g K it follows immediately that ? Dg is a linear differential operator which is ellip- 2φ tic for Riemannian and hyperbolic for Lorentzian Weylgˆ = e Weylg . (121) metrics g. If we set Ω = Ω1Ω2 and apply (125) This is equivalently expressed by the conformal in- twice, one time to the pair (ˆg, g), the other time variance of the contravariant version of the Weyl to (ˆg, Ω2g), we obtain by direct comparison (and
20 renaming Ω2 to Ω thereafter) the conformal trans- Here we used h(W, ∇X n) = −K(W, X) from (75), formation property for the operator Dg: and
− n+2 D 4 = M Ω n−2 ◦ Dg ◦ M(Ω) , (126) Riem(n, Z, X, Y ) Ω n−2 g (130) = ε (DX K)(Y,Z) − (DY K)(X,Z) . where M(Ω) is the linear operator of multiplica- tion with Ω. This is the reason why Dg is called Here and in the sequel we return to the meaning of the conformally covariant Laplacian (for Rieman- h given by (48). In differential geometry (129) is nian g) or the conformally covariant wave operator referred to as Gauss equation and (130) as Codazzi- (for Lorentzian g). As we will see, it has useful Mainardi equation. applications to the initial-data problem in GR. The remaining curvature components are those involving two entries in n direction. Using (79) 5.2 Curvature decomposition we obtain via standard manipulations (now using metricity and vanishing torsion) Using (67) we can decompose the various curvature tensors. First we let X,Y,Z be horizontal vector Riem(X, n, Y, n) fields. We use (67) in (88) and get the general = i ∇ ∇ − ∇ ∇ − ∇ n[ formula (i.e. not yet making use of the fact that ∇ X Y n n Y [Y,n] [ [ [ and D are metric and torsion free) = iX iY εLnK + K ◦ K + Da − εa ⊗ a . (131) R(X,Y )Z = RD(X,Y )Z −1 + (∇X n) K(Y,Z) − (∇Y n) K(X,Z) Here K ◦ K (X,Y ) := h (iX K, iY K) = ] + n (DX K)(Y,Z) − (DY K)(X,Z) iX K (iY K) and we used the following relation between covariant and Lie derivative (which will + n K T D(X,Y ),Z , have additional terms in case of non-vanishing tor- (127) sion): where ∇nK = LnK + 2ε K ◦ K. (132) D R (X,Y, )Z := DX DY − DY DX − D[X,Y ] Z Note also that the left-hand side of (131) is sym- (128) metric as consequence of (94d). On the right-hand is the horizontal curvature tensor associated to the side only Da[ is not immediately seen to be sym- Levi-Civita covariant derivative D of h. This for- metric, but that follows from (52b). Unlike (129) mula is general in the sense that it is valid for any and (130), equation (131) does not seem to have a covariant derivative. No assumptions have been standard name in differential geometry. made so far concerning metricity or torsion, and Equations (127), (129), and (130) express all this is why the torsion T D of D (defined in (70)) components of the spacetime curvature in terms of makes an explicit appearance. From now on we horizontal quantities and their Lie derivatives Ln shall restrict to vanishing torsion. We observe that in normal direction. According to (55) the latter the first two lines on the right-hand side of (127) can be replaced by a combination of Lie derivatives are horizontal whereas the last two lines are pro- along the time vector-field ∂/∂t and the shift β. portional to n. Decomposition into horizontal and From (53b) we infer that Lαn = αLn on horizontal normal components, respectively, leads to (where covariant tensor fields, therefore we may replace T D = 0 and X,Y,Z, and W are horizontal), −1 −1 k k L → α L ∂ − L → α L − L (133) D n β ∂ β Riem(W, Z, X, Y ) = Riem (W, Z, X, Y ) c∂t c∂t − εK(W, X)K(Z,Y ) − K(W, Y )K(Z,X) . on horizontal covariant tensor fields. Here we set (129) Lk = P k ◦L, i.e. Lie derivative (as operation in the
21 ambient spacetime) followed by horizontal projec- where ∇· denotes the divergence with respect to ∇ tion. Moreover, using (50), one easily sees that the and V is a vector field on M whose normal com- acceleration 1-form a[ can be expressed in terms of ponent is the trace of the extrinsic curvature and the spatial derivative of the lapse function: whose horizontal component is ε times the acceler- ation on n: a[ = −εα−1Dα . (134) c V = nKc + εa . (139) Hence the combination of accelerations appearing For the horizontal-horizontal components of Ein- in (130) may be written as stein’s equation it turns out to be simpler to use their alternative form (6b) with the Ricci tensor on Da[ − εa[ ⊗ a[ = −εα−1D2α . (135) the left hand side. For that we need the horizontal components of the Ricci tensor, which we easily get Note that D2α := DDα is just the horizontal co- from (129) and (131): variant Hessian of α with respect to h. D Ric(ea, eb) = Ric (ea, eb) + L K + 2εK Kc − εK Kc 6 Decomposing Einstein’s n ab ac b ab c + εDaab − aaab . equations (140)
The curvature decomposition of the previous sec- For later applications we also note the expression tion can now be used to decompose Einstein’s equa- for the scalar curvature. It follows, e.g., from tions. For this we decompose the Einstein ten- adding the horizontal trace of (140) to ε times sor Ein into the normal-normal, normal-tangential, (138). This leads to and tangential-tangential parts. Let {e0, e1, e2, e3} Scal = ScalD − ε K Kab − (Ka)2 + 2∇ · V. be an orthonormal frame with e0 = n, i.e. adapted ab a to the foliation as in Section 4.1. Then (102) to- (141) gether with (129) immediately lead to Here we made use of the relation between the ∇ and D derivative for the acceleration 1-form: ab a 2 D 2 Ein(e0, e0) = − KabK − (Ka ) − εScal , [ [ [ [ [ (136) ∇a = Da + ε n ⊗ ∇na + iaK ⊗ n , (142) where ScalD is the scalar curvature of D, i.e. of whose trace gives the following relation between the spacelike leaves in the metric h. Similarly we the ∇ and D divergences of a: obtain from (130),
b b ∇ · a = D · a − εh(a, a) . (143) Ein(e0, ea) = Ric(e0, ea) = −ε D Kab − DaKb . (137) Another possibility would have been to use (136) The normal-normal component of the Ricci ten- and (138) in Scal = −2ε(Ein(e0, e0)−Ric(e0, e0)). sor cannot likewise be expressed simply in terms of Using (136) and (137), and also using the horizontal quantities, the geometric reason being De Witt metric (118) for notational ease, we can that, unlike the Einstein tensor, it involves non- immediately write down the normal-normal and horizontal sectional curvatures (compare (101) and normal-tangential components of Einstein’s equa- (102)). A useful expression follows from taking the tions (3): trace of (131), considered as symmetric bilinear form in X and Y . The result is : abcd D G KabKcd + εScal = −2κ T(n, n) , (144a) abcd ab ab c 2 G DbKcd = −εκh T(n, eb) . (144b) Ric(e0, e0) = −KabK + (Kc ) + ε∇ · V, (138)
22 From (77) and (118) we notice that the bilinear in form on the left-hand side of (144a) can be written ˙ k as Kab := L ∂ K c∂t ab k = LβK ab + DaDbα abcd G(K,K) : = G KabKcd c c D + α −2εKacKb + εKabKc − Ric (ea, eb) 2 = Tr(Wein ◦ Wein) − Tr(Wein) . κ − αε n−2 habT(n, n) (145) 1 + ακ T − n−2 Trh(T) h (ea, eb) . (148) Here the trace is natural (needs no metric for its definition) since Wein is an endomorphism. In a Note that in the last term the trace of T is taken local frame in which Wein is diagonal with entries with respect to h and not g. The relation is ~ Tr (T) = Tr (T) − εT(n, n). k := (k1, k2, k3) we have h g The only remaining equation that needs to be added here is that which relates the time derivative ab a b G(K,K) := (δ − 3n n )kakb , (146) of h with K. This we get from (80) and (133):
h˙ := Lk h = Lk h − 2αK . (149) a ab ∂ ab β ab ab where n are the√ components of the normalized c∂t vector (1, 1, 1)/ 3 in eigenvalue-space, which we Equations (149) and (148) are six first-order in identify with 3 endowed with the standard Eu- R time evolution equations for the pair (h, K). This clidean inner product. Hence, denoting by θ the pair cannot be freely specified but has to obey the angle between ~n and ~k, we have four equations (144a) and (144b) which do not contain any time derivatives of h or K. Equa- 0 if | cos θ| = p1/3 tions (144a) and (144b) are therefore referred to as constraints, more specifically (144a) as scalar con- G(K,K) = > 0 if | cos θ| < p1/3 (147) p straint (also Hamiltonian constraint) and (144b) as < 0 if | cos θ| > 1/3 . vector constraint (also diffeomorphism constraint).
p We derived these equations from the 3+1 split Note that | cos θ| = 1/3 describes a double cone of a spacetime that we considered to be given. De- around the symmetry axis generated by ~n and ver- spite having expressed all equations in terms of hor- tex at the origin, whose opening angle just is right izontal quantities, there is still a relic of the ambi- 3 so as to contain all three axes of R . For eigenvalue- ent space in our equations, namely the Lie deriva- vectors inside this cone the bilinear form is nega- tive with respect to ∂/∂ct. We now erase this last tive, outside this cone positive. Positive G(K,K) relic by interpreting this Lie derivative as ordinary require sufficiently anisotropic Weingarten maps, partial derivative of some t-dependent tensor field or, in other words, sufficiently large deviations from on a genuine 3-dimensional manifold Σ, which is being umbilical points. not thought of as being embedded into a space- k The horizontal-horizontal component of Ein- time. The horizontal projection Lβ of the space- stein’s equations in the form (5) immediately fol- time Lie derivative that appears on the right-hand lows from (140). In the ensuing formula we use sides of the evolution equations above then trans- (133) to explicitly solve for the horizontal Lie lates to the ordinary intrinsic Lie derivative on Σ derivative of K with respect to ∂/c∂t and also (135) with respect to β. This is how from now on we shall to simplify the last two terms in (140). This results read the above equations. Spacetime does not yet
23 exist. Rather, it has to be constructed from the which shows immediately that the time derivatives evolution of the fields according to the equations, of the constraint functions are zero if the con- usually complemented by the equations that gov- straints vanished initially. This suffices for ana- ern the evolution of the matter fields. In these lytic data, but in the general case one has to do evolution equations α and β are freely specifiable more work. Fortunately the equations for the evo- functions, the choice of which is subject to mathe- lution of the constraint functions can be put into matical/computational convenience. Once α and β an equivalent form which is manifestly symmetric are specified and h as a function of parameter-time hyperbolic [62]. That suffices to conclude the has been determined, we can form the expression preservation of the constraints in general. In fact, (64) for the spacetime metric and know that, by symmetric hyperbolicity implies more than that. It construction, it will satisfy Einstein’s equations. ensures the well-posedness of the initial-value prob- To sum up, the initial-value problem consists in lem for the constraints, which not only says that the following steps: they stay zero if they are zero initially, but also that they stay small if they are small initially. This 1. Choose a 3-manifold Σ. is of paramount importance in numerical evolution 2. Choose a time-parameter dependent lapse schemes, in which small initial violations of the con- function α and a time-parameter dependent straints must be allowed for and hence the conse- shift vector-field β. quences of these violations need to be controlled. For a recent and mathematically more thorough 0 3. Find a Riemannian metric h ∈ ΓT2 Σ and a discussion of the Cauchy problem we refer to James symmetric covariant rank-2 tensor field K ∈ Isenberg’s survey [86]. 0 ΓT2 Σ that satisfy equations (144a) and (144b) Finally we wish to substantiate our earlier claim either in vacuum (T = TΛ; cf. (4)), or after that any Σ can carry some initial data. Let us show specifying some matter model. this for closed Σ. To this end we choose a matter 4. Evolve these data via (149) and (148), possibly model such that the right-hand side of (144b) van- complemented by the evolution equations for ishes. Note that this still allows for arbitrary cos- the matter variables. mological constants since TΛ(n, ea) ∝ g(n, ea) = 0. Next we restrict to those pairs (h, K) were K = λh 5. Construct from the solution the spacetime for some constant λ. Geometrically this means metric g via (64). that, in the spacetime to be developed, the Cauchy For this to be consistent we need to check that the surface will be totally umbilical (isotropic Wein- evolution according to (149) and (148) will pre- garten map). Due to this proportionality and the serve the constraints (144a) and (144b). At this previous assumption the vector constraint (144b) stage this could be checked directly, at least in will be satisfied. In the scalar constraint we have 2 the vacuum case. The easiest way to do this is G(K,K) = G(λh, λh) = −6λ so that it will be to use the equivalence of these equations with Ein- satisfied provided that stein’s equations and then employ the twice con- D 2 tracted 2nd Bianchi identity (13). It follows that −εScal = 2κT(n, n) − 6λ . (151) µν µν µν µν ∇µE ≡ 0, where E = G + λg . The four constraints (144a) and (144b) are equivalent For the following argument the Lorentzian signa- to E00 = 0 and E0m = 0, and the six second-order ture, ε = −1, will matter. For physical rea- equations Emn = 0 to the twelve first-order evolu- sons we assume the weak energy condition so that tion equations (149) and (148). In coordinates the κT(n, n) ≥ 0, which makes a positive contribution µν to the right-hand side of (151). However, if we identity ∇µE ≡ 0 reads choose the modulus of λ sufficiently large we can 0ν mν µ λν ν µλ ∂0E = −∂mE − ΓµλE − ΓµλE , (150) make the right-hand side negative somewhere (or
24 everywhere, since Σ is compact). Now, in dimen- and β = 0, so that g = −c2 dt2 + h. This means sions 3 or higher the following theorem of Kazdan that n = ∂/∂ct is geodesic. Taking such a gauge & Warner holds ([91], Theorem 1.1): Any smooth from the t = 0 slice in the Schwarzschild/Kruskal function on a compact manifold which is negative spacetime would let the slices run into the singu- somewhere is the scalar curvature for some smooth larity after a proper-time of t = πGM/c3, where Riemannian metric. Hence a smooth h exists which M is the mass of the black hole. In that short pe- solves (151) for any given T(n, n) ≥ 0, provided we riod of time the slices had no chance to explore a choose λ2 > |λ| sufficiently large. If Σ is not closed significant portion of spacetime outside the black a corresponding theorem may also be shown [118]. hole. The above argument crucially depends on the A gauge condition that one may anticipate to signs. There is no corresponding statement for pos- have singularity-avoiding character is that where α itive scalar curvatures. In fact, there is a strong is chosen such that the divergence of the normal topological obstruction against Riemannian met- field n is zero. This condition just means that the rics of strictly positive scalar curvature. It follows locally co-moving infinitesimal volume elements do from the theorem of Gromov & Lawson ([78], The- not change volume, for Lndµ = (∇ · n) dµ, where 3 orem 8.1) that a 3-dimensional closed orientable Σ dµ = det{hab}d x is the volume element of Σ. allows for Riemannian metrics with positive scalar From (79) we see that n has zero divergence iff curvature iff its prime decomposition consists of K has zero trace, i.e. the slices are of zero mean- prime-manifolds with finite fundamental group or curvature. The condition on α for this to be pre- “handles” S1 × S2. All manifolds whose prime list served under evolution follows from contains at least one so-called K(π, 1)-factor (a 3- ab ab ab manifold whose only non-trivial homotopy group is 0 = Ln(h Kab) = −K Lnhab + h LnKab . the first) are excluded. See, e.g., [67] for more ex- (152) planation of these notions. We conclude that the Here we use (80) to eliminate Lnhab in the first given argument crucially depends on ε = −1. term and (131) to eliminate LnKab in the second term, also making use of (135). This leads to the following equivalent of (152): 6.1 A note on slicing conditions ab ∆hα + ε Ric(n, n) + K Kab α = 0 . (153) The freedom in choosing the lapse and shift func- tions can be of much importance, theoretically and This is a linear elliptic equation for α. The case in numerical evolution schemes. This is particu- of interest to us in GR is ε = −1. In the closed larly true for the lapse function α, which deter- case we immediately deduce by standard argu- mines the amount of proper length by which the ments that α = 0 is the only solution, provided Cauchy slice advances in normal direction per unit the strong energy-condition holds (which implies parameter interval. If a singularity is to form in Ric(n, n) ≥ 0). In the open case, where we might spacetime due to the collapse of matter within a impose α → 1 as asymptotic condition, we de- bounded spatial region, it would clearly be advan- duce existence and uniqueness again under the as- tageous to not let the slices run into the singularity sumption of the strong energy condition. Hence we ab before the outer parts of it have had any chance may indeed impose the condition h Kab = 0, or to develop a sufficiently large portion of spacetime Tr(Wein) = 0, for non-closed Σ. It is called the that one might be interested in, e.g. for the study maximal slicing condition or York gauge [119]. of gravitational waves produced in the past. This Whereas this gauge condition has indeed the de- means that one would like to slow down α in regions sired singularity-avoiding character it is also not which are likely to develop a singularity and speed easy to implement due to the fact that at each new up α in those regions where it seems affordable. stage of the evolution one has to solve the elliptic Take as an example the “equal-speed” gauge α = 1 equation (153). For numerical studies it is easier
25 to implement evolution equations for α. Such an The inverse metric to (157) is given by equation is, e.g., obtained by asking the time func- tion (36) to be harmonic, in the sense that −1 −1 ∂ ∂ G(λ) = G(λ) abcd ⊗ , (158a) ∂hab ∂hcd 0 = t := gµν ∇ ∇ t g µ ν where 1 1 (154) − 2 2 µν = | det{gαβ}| ∂µ | det{gαβ}| g ∂ν t . 1 G−1 = h h + h h − 2µ h h . (λ) abcd 2 ac bd ad bc ab cd This is clearly just equivalent to (158b) The relation between λ and µ is 1 µ0 ∂µ | det{gµν }| 2 g = 0 , (155) λ + µ = nλµ , (159) which can be rewritten using (65) and (66) to give so that ∂α α˙ : = = L α − εKaα2 GabnmG−1 = 1 δaδb + δaδb . (160) c∂t β a (156) (λ) (λ) nmcd 2 c d d c 2 = Lβα + Tr(Wein) α . In ordinary GR n = 3, λ = 1, and µ = 1/2. Note that there are good reasons the keep the su- This is called the harmonic slicing condition. Note perscript −1 even in component notation, that is, that we can still choose β = 0 and try to determine to write G−1 rather than just G , since α as function of the trace of Wein. There also (λ) abcd (λ) abcd −1 2 klmn exist generalizations to this condition where α on G(λ) abcd does not equal hakhblhcmhdnG(λ) unless the right-hand side is replaced with other functions λ = 2/n, in which case λ = µ. f(α). If we change coordinates according to
1 n τ : = ln det{hab} , 6.2 A note on the De Witt metric (161) 1 n At each point p on Σ the De Witt metric (118) can rab : = hab/ det{hab} , be regarded as a symmetric bilinear form on the where τ parametrizes conformal changes and r space of positive-definite inner products h of T Σ. ab p the conformally invariant ones, the metric (157) The latter is an open convex cone in T ∗ Σ ⊗ T ∗ Σ. P P reads We wish to explore its properties a little further. A frame in T Σ induces a frame in T ∗Σ ⊗ T ∗Σ ac bd p p p G(λ) = n(1−λn) dτ ⊗dτ +r r drab⊗drcd , (162) (tensor product of the dual frame). If hab are the an a components of h then we have the following repre- where r rnb = δb . Since h is positive definite, so is sentation of the generalized De Witt metric r. Hence the second part is positive definite on the 1 n(n+1)−1– dimensional vector space of trace- abcd 2 G(λ) = G(λ) dhab ⊗ dhcd , (157a) free symmetric tensors. Hence the De Witt metric is positive definite for λ < 1/n, Lorentzian for λ > where 1/n, and simply degenerate (one-dimensional null space) for the critical value λ = 1/n. In the GR abcd 1 ac bd ad bc ab cd G = h h + h h − 2λ h h . (157b) case we have λ = 1 and n = 3, so that the De Witt (λ) 2 metric is Lorentzian of signature (−, +, +, +, +, +). Here we introduced a factor λ in order to Note that this Lorentzian signature is independent parametrize the impact of the negative trace term. of ε, i.e. it has nothing to do with the Lorentzian We also consider Σ to be of general dimension n. signature of the spacetime metric.
26 In the Hamiltonian formulation it is not G but We stress once more that the signature of the rather a conformally related metric that is impor- De Witt metric is not related to the signature of p tant, the conformal factor being det{hab}. If we spacetime, i.e. independent of ε. For example, for set the GR values λ = 1 and n = 3, it is Lorentzian ˆ 1/2 G(λ) := det{hab} G(λ) (163) even if spacetime were given a Riemannian met- ric. Moreover, by integrating over Σ, the pointwise and correspondingly metric (166) defines a bilinear form on the infinite dimensional space of Riemannian structures on Σ, ˆ−1 −1/2 −1 G(λ) := det{hab} G(λ) , (164) the geometry of which may be investigated to some limited extent [70][74]. ˆ we can again write G(λ) in terms of (τ, rab). In fact, the conformal rescaling clearly just corresponds to p nτ/2 multiplying (162) with det{hab} = e . Set- 7 Constrained Hamiltonian ting systems 1/2 T := 4(1 − nλ)/n enτ/4 (165) In this section we wish to display some charac- we get, excluding the degenerate case λ = 1/n, teristic features of Hamiltonian dynamical systems with constraints. We restrict attention to finite- ˆ G(λ) = sign(1 − nλ) dT ⊗ dT (166) dimensional systems in order to not overload the 2 ac bd + T C r r drab ⊗ drcd , discussion with analytical subtleties. Let Q be the n-dimensional configuration man- where C = n/(16|1 − nλ|) (= 3/32 in GR). This ifold of a dynamical system that we locally co- 1 n is a simple warped product metric of R+ with ordinatize by (q , ··· , q ). By TQ we denote the left-invariant metric on the homogeneous space its tangent bundle, which we coordinatize by 1 n 1 n GL(3, R)/SO(3) × R+ of symmetric positive defi- (q , ··· , q , v , ··· , v ), so that a tangent vector nite forms modulo overall scale, the warping func- X ∈ TQ is given by X = va∂/∂qa. The dynamics 2 tion being just T if T is the coordinate on R+. of the system is described by a Lagrangian Now, generally, quadratic warped-product metrics of the form ±dT ⊗ dT + T 2g, where g is indepen- L : TQ → R , (167) dent of T , are non-singular for T & 0 iff g is a metric of constant curvature ±1 (like for a unit which selects the dynamically possible trajectories n in TQ as follows: Let 3 t 7→ x(t) ∈ Q be a (at sphere in R , with T being the radius coordinate, R or the unit spacelike hyperboloid in n-dimensional least twice continuously differentiable) curve, then Minkowski space, respectively). This is not the it is dynamically possible iff the following Euler case for (166), which therefore has a curvature sin- Lagrange equations hold (we set dx/dt =:x ˙): gularity for small T , i.e. small det{hab}. Note that " # this is a singularity in the space of metrics (here ∂L d ∂L a − a = 0 . (168) at a fixed space point), which has nothing to do ∂q q=x(t) dt ∂v q=x(t) v=x ˙ (t) v=x ˙ (t) with spacetime singularities. In the early days of Canonical Quantum Gravity this has led to specu- Performing the t-differentiation on the second lations concerning “natural” boundary conditions term, this is equivalent to for the wave function, whose domain is the space of b metrics [52]. The intention was to pose conditions Hab x(t), x˙(t) x¨ = Va x(t), x˙(t) , (169) such that the wave function should stay away from where such singular regions in the space of metrics; see ∂2L(q, v) also [92] for a more recent discussion. Hab(q, v) := , (170) ∂va∂vb
27 and are s functions φα, α = 1, ··· , s such that 2 ∂L(q, v) ∂ L(q, v) b C∗ := (q, p) ∈ T ∗Q : φ (q, p) = 0 , α = 1, ··· , s . Va(q, v) := − v . (171) α ∂qa ∂va∂qb (175) This is called the constraint surface in phase space. Here we regard H and V as function on TQ with It is given as the intersection of the zero-level sets of values in the symmetric n × n matrices and n R s independent functions. Independence means that respectively. In order to be able to solve (169) for at each p ∈ C∗ the s one-forms dφ | ··· , dφ | are the second derivativex ¨ the matrix H has to be 1 p s p linearly independent elements of T ∗T ∗Q. invertible, that is, it must have rank n. That is p The dynamical trajectories of our system will the case usually encountered in mechanics. On the ∗ other hand, constrained systems are those where stay entirely on C . The trajectories themselves are the rank of H is not maximal. This is the case we integral lines of a Hamiltonian flow. But what is are interested in. the Hamiltonian function that generates this flow? We assume H to be of constant rank r < n. To explain this we first recall the definition of the Then, for each point on TQ, there exist s = (n − energy function for the Lagrangian L. It is a func- tion E : TQ → R defined through r) linearly independent kernel elements K(α)(q, v), α = 1, ··· , s, such that Ka (q, v)H (q, v) = 0. (α) ab ∂L(q, v) E(q, v) := va − L(q, v) . (176) Hence any solution x(t) to (169) must be such that ∂va the curve t 7→ x(t), x˙(t) in TQ stays on the subset At first sight this function cannot be defined on C := (q, v) ∈ TQ : ψα(q, v) = 0 , α = 1, ··· , s , phase space, for we cannot invert FL to express v (172a) as function of q and p which we could insert into where E(q, v) in order to get E(q, v(q, p)). However, one may prove the following: There exists a function a ψα(q, v) = K(α)(q, v)Va(q, v) . (172b) ∗ HC∗ : C → R , (177a) We assume C ⊂ TQ to be a smooth closed subman- ifold of co-dimension s, i.e. of dimension 2n − s = so that n + r. E = HC∗ ◦ FL . (177b) Now we consider the cotangent bundle T ∗Q over Q. On T ∗Q we will use so-called canonical coor- A local version of this is seen directly from tak- 1 n dinates, denoted by {q , ··· , q , p1, ··· , pn}, the ing the differential of (176), which yields dE = precise definition of which we will give below. The vad(∂L/∂va) − (∂L/∂qa)dqa, expressing the fact ∗ Lagrangian defines a map FL : TQ → T Q, which that dE(q,v)(X) = 0 if FL∗(q,v)(X) = 0 for X ∈ in these coordinates reads T(q,v)TQ, or in simple terms: E does not vary if q and p do not vary. ∂L(q, v) ∗ FL(q, v) = q, p := . (173) So far the function HC∗ is only defined on C . By ∂v our regularity assumptions there exists a smooth ∗ extension of it to T Q, that is a function H0 : From what has been said above it follows that the ∗ T Q → such that H | ∗ = H ∗ . This is clearly Jacobian of that map has constant rank n + r. R 0 C C not unique. But we can state the following: Let H Given sufficient regularity, we may further assume 0 and H both be smooth (at least continuously dif- that ∗ ferentiable) extensions of H ∗ to T Q, then there C∗ := FL C ⊂ T ∗Q (174) C exist s smooth functions λα : T ∗Q → R such that is a smoothly embedded closed submanifold in ∗ α phase space T Q of co-dimension s. Hence there H = H0 + λ φα . (178)
28 Locally a proof is simple: Let f : T ∗Q → R be for all α, β ∈ {1, ··· , s}. Following Dirac [55], continuously differentiable and such that f|C∗ ≡ constraints which satisfy this condition are said to 0. Consider a point p ∈ C∗ and coordinates be of first class. By the result shown (locally) (x1, ··· , x2n−s, y1, ··· ys) in a neighborhood U ⊂ above in (179) this is equivalent to the existence ∗ 1 2 T Q of p, where the x’s are coordinates on the con- of 2 s (s − 1) (at least continuously differentiable) γ γ ∗ straint surface and the y’s are just the functions φ. real-valued functions Cαβ = −Cβα on T Q, such In U the constraint surface is clearly just given by that y1 = ··· = ys = 0. Then γ {φα, φβ} = Cαβ φγ . (182) Z 1 d Note that as far as the intrinsic geometric proper- f|U (x, y) = dt f(x, ty) 0 dt ties of the constraint surface are concerned (181) Z 1 ∂f and (182) are equivalent. = dt (x, ty) yα = λ (x, y) yα , ∂yα α The indeterminacy of the Hamiltonian due to the 0 α (179a) freedom to choose any set of λ seems to imply an s-dimension worth of indeterminacy in the dy- where namically allowed motions. But the difference in Z 1 ∂f these motions is that generated by the constraint λ (x, y) := dt (x, ty) . (179b) α α functions on the constraint surface. In order to ac- 0 ∂y tually tell apart two such motions requires observ- For a global discussion see [80]. ables (phase-space functions) whose Poisson brack- As Hamiltonian for our constraint system we ets with the constraints do not vanish on the con- address any smooth (at least continuously differen- straint surface. The general attitude is to assume tiable) extension H of HC∗ . So if H0 is a somehow that this is not possible, i.e. to assume that phys- given one, any other can be written as ical observables correspond exclusively to phase- α H = H0 + λ φα (180) space functions whose Poisson bracket with all con- straints vanish on the constraint surface. This is for some (at least continuously differentiable) real- expressed by saying that all motions generated by α ∗ valued functions λ on T Q. the constraints are gauge transformations. This en- Here we have been implicitly assuming that the tails that they are undetectable in principle and Hamiltonian dynamics does not leave the con- merely correspond to a mathematical redundancy straint surface (174). If this were not the case in the description rather than to any physical de- we would have to restrict further to proper sub- grees of freedom. It is therefore more correct to ∗ manifolds of C such that the Hamiltonian vector speak of gauge redundancies rather than of gauge fields evaluated on them lie tangentially. (If no symmetries, as it is sometimes done, for the word such submanifold can be found the theory is sim- “symmetry” is usually used for a physically mean- ply empty). This is sometimes expressed by saying ingful operation that does change the object to that the primary constraints (those encountered which it is applied in at least some aspects (other- first in the Lagrangian/Hamiltonian analysis) are wise the operation is the identity). Only some “rel- completed by secondary, tertiary, etc. constraints evant” aspects, in the context of which one speaks for consistency. of symmetry, are not changed. Here we assume that our system is already dy- namically consistent. This entails that the Hamil-
tonian vector-fields Xφα for the φα are tangential 7.1 Geometric theory to the constraint surface. This is equivalent to Being first class has an interpretation in terms of X (φ )| ∗ = 0, or expressed in Poisson brackets: φα β C symplectic geometry. To see this, we first recall a
{φα, φβ} C∗ = 0 , (181) few facts and notation from elementary symplec-
29 tic geometry of cotangent bundles. Here some sign The special fibre-preserving diffeomorphisms we conventions enter and the reader is advised to com- wish to mention are given by adding to each mo- pare carefully with other texts. mentum p ∈ T ∗Q the value σ π(p) of a section A symplectic structure on a manifold is a non- σ : Q → T ∗Q: degenerate closed two-form. Such structures al- ways exist in a natural way on cotangent bun- F (p) = p + σπ(p) . (186) dles, where they even derive from a symplectic po- This transforms the symplectic potential at p ∈ tential. The latter is a one-form field θ on T ∗Q T ∗Q into whose general geometric definition is as follows: Let π : T ∗Q → Q be the natural projection from (F ∗θ) = θ ◦ F the co-tangent bundle of Q (phase space) to Q it- p F (p) ∗p self. Then, for each p ∈ T ∗Q, we define (183) = F (p) ◦ π ◦ F ∗p
θp := p ◦ π∗p . (183) (185) = F (p) ◦ π∗p (187) ∗ So in order to apply θp to a vector X ∈ TpT Q, (186) = θp + σπ(p) ◦ π∗p we do the following: Take the differential π∗ of the projection map π, evaluate it at point p and ∗ = θp + π σ . ∗ p apply it to X ∈ TpT Q in order to push it forward to the tangent space Tπ(p)Q at point π(p) ∈ Q. Hence Then apply p to it, which makes sense since p is, ∗ ∗ by definition, an element of the co-tangent space F θ = θ + π σ , (188a) at π(p) ∈ Q. F ∗ω = ω − π∗dσ . (188b) The symplectic structure, ω, is now given by This is a canonical transformation if σ is a closed ω = − dθ . (184) covector field on Q. By Poincar´e’sLemma such a The minus sign on the right-hand side has no signif- σ is locally exact, but this need not be the case globally. Obstructions to global exactness are the icance other than to comply with standard conven- 1 tions. Let us stress that θ, and hence ω, is globally first De Rahm cohomology class HDR(Q), which is defined. This is obvious from the global definition just defined to be the vector space of closed mod- (183). Therefore ω is not only closed, dω = 0, but ulo exact covector fields on Q. The dimension of even globally exact for any Q. Non-degeneracy of ω this vector space equals the rank of the free part will be immediate from the expression in canonical of the ordinary first homology group H1(Q, Z) on coordinates to be discussed below (cf. (196b)). Q with integer coefficients. This latter group is A diffeomorphism F : T ∗Q → T ∗Q is called a always abelian and isomorphic to the abelianiza- canonical transformation or symplectic morphism tion of the (generally non-abelian) first homotopy if it preserves ω, that is, if F ∗ω = ω. We explicitly group π1(Q). Hence for non-simply connected Q mention two kinds of canonical transformations, the possibility of canonical transformations exist which in some sense are complementary to each which change the symplectic potential by a closed other. yet non-exact covector field. The second set of canonical transformations that The first set of canonical transformations are ∗ fibre-preserving ones. This means that, for each we wish to mention are natural extensions to T Q q ∈ Q, points in the fibre π−1(q) are moved to of diffeomorphisms of Q. These extensions not only points in the same fibre π−1(q). This is equivalent leave invariant the symplectic structure ω but also to the simple equation the symplectic potential θ. To see this we note that any diffeomorphism f : Q → Q has a natural lift π ◦ F = π . (185) to T ∗Q. We recall that a lift of a diffeomorphism
30 f of the base manifold Q is a diffeomorphism F : the first set of n component functions by za = qa ∗ ∗ n+a T Q → T Q such that (for a = 1, ··· , n) and the second set by z = pa (for a = 1, ··· , n). For the first set we define π ◦ F = f ◦ π . (189) qa(λ) := xa π(λ) , (193a) This is equivalent to saying that the following dia- and for the second gram of maps commutes (a tailed arrow indicates ! injectivity and a double-headed arrow surjectivity) ∂ pa(λ) := λ a , (193b) ∂x π(λ) F T ∗QQT/ / / T ∗Q for any λ ∈ V . Note that (193b) just says that λ = p (λ) dxa| . In this way we get a “canonical” π π (190) a π(λ) extension of any chart on Q with domain U to a chart on T ∗Q with domain V = π−1(U). From the QQQ / / / Q f definition it is clear that ∂ ∂ Here the map F is just the pull-back of the inverse π∗λ a = a (194a) f −1. Hence the image of p ∈ T ∗Q is given by ∂q λ ∂x π(λ) and F (p) = p ◦ f −1 . (191) ∗f(π(p)) ∂ π∗λ = 0 . (194b) ∂pa From that it follows that the symplectic potential λ is invariant under all lifts of diffeomorphisms on Q: It immediately follows from the definition (183) that ∂ ∗ (F θ)p = θF (p) ◦ F∗p θλ a = pa(λ) (195a) ∂q λ (183) = F (p) ◦ π ◦ F and ∗p ∂ (189) θλ = 0 . (195b) = F (p) ◦ f∗π(p) ◦ π∗p ∂pa λ (192) (191) Hence, in canonical coordinates, the symplectic po- = p ◦ f −1 ◦ f ◦ π ∗f(π(p)) ∗π(p) ∗p tential and structure take on the form −1 = p ◦ f ◦ f ◦ π∗p a ∗π(p) θ|V = pa dq , (196a) a ω|V = dq ∧ dpa . (196b) = θp . Note again that (196) is valid in any canoni- So far we deliberately avoided intoducing local cal completion of a chart on Q. As advertised coordinates in order to stress global existence of above, it is immediate from (196b) that ω| is the quantities in question. We now introduce con- V non-degenerate at any point p ∈ V . Since non- venient coordinates in which the symplectic poten- degeneracy is a pointwise property and valid in any tial and structure take on the familiar form. These canonical chart, it follows that ω is non-degenerate are called canonical coordinates, which we already everywhere. In the sequel we shall drop the explicit mentioned above and the definition of which we mention of the chart domain V . now give. Let (x, U) be a local chart on Q such The non-degeneracy of ω allows to uniquely asso- that x : Q ⊃ U → n is the chart map with com- R ciate a vector field X to any real-valued function ponent functions xa. This chart induces a chart f f on T ∗Q through (z, V ) on T ∗Q, where V = π−1(U) ⊂ T ∗Q and 2n z : V → R . We follow general tradition and label iXf ω = df . (197)
31 It is called the Hamiltonian vector field of f. An the first term. As ω is non-degenerate comparison immediate consequence of (197) and dω = 0 is that with (197) leads to (202). Next we recall that the ω has vanishing Lie derivative with respect to any exterior differential of a general k-form field α, ap- Hamiltonian vector field: plied to the k + 1 vectors X0,X1, ··· ,Xk, can be written as
LXf ω = (iXf ◦ d + d ◦ iXf )ω = 0 . (198) dα(X0, ··· ,Xk) In coordinates Xf looks like this: X i = (−1) Xi α(X0, ··· , Xˆi, ··· ,Xk) 0≤i≤k ∂f ∂ ∂f ∂ (204) Xf = a − a . (199) X i+j ∂pa ∂q ∂q ∂pa + (−1) α [Xi,Xj],X0, ··· , 0≤i