Table Of Content

Wheeler-DeWitt quantization and singularities F. T. Falciano,1,∗ N. Pinto-Neto,1,† and W. Struyve2,‡ 1CBPF - Centro Brasileiro de Pesquisas F´ısicas, Xavier Sigaud st. 150, zip 22290-180, Rio de Janeiro, Brazil. 2Department of Physics, University of Liege, 4000 Liege, Belgium. (Dated: January 20, 2015) WeconsideraBohmianapproachtotheWheeler-DeWittquantizationoftheFriedmann-Lemaˆıtre- Robertson-Walker model and investigate the question whether or not there are singularities, in the sense that the universe reaches zero volume. We find that for generic wave functions (i.e., non- classical wave functions), there is a non-zero probability for a trajectory to be non-singular. This should be contrasted to the consistent histories approach for which it was recently shown by Craig andSinghthatthereisalwaysasingularity. Thisresultillustratesthatthequestionofsingularities depends much on which version of quantum theory one adopts. This was already pointed out by Pinto-Neto et al., albeit with a different Bohmian approach. Our current Bohmian approach 5 agreeswiththeconsistenthistoriesapproachbyCraigandSinghforsingle-timehistories,unlikethe 1 one studied earlier by Pinto-Neto et al. Although the trajectories are usually different in the two 0 Bohmian approach, their qualitative behavior is the same for generic wave functions. 2 n PACSnumbers: 98.80.Cq,04.60.Ds,98.80.Qc a J 7 I. INTRODUCTION However, while the dynamics of the Bohmian model is 1 very natural, there is no natural probability distribution onthesetofhistories. Assuch,thereisnoimmediateway ] Recently, some of us investigated the issue of singular- of making probabilistic statements, like about the prob- c ities in the Wheeler-DeWitt approach to quantum cos- q ability for a history to have a singularity for a particular mologyandshowedthattheanswerdependsstronglyon - wavefunction. Ontheotherhand,CraigandSingh[2,3] r theversionofquantummechanicsthatoneconsiders[1]. g showed that probabilistic statements can be made in the Two versions were compared: the consistent (or deco- [ consistent histories approach. This might seem puzzling herent) histories and the Bohmian approach (also called since, at least in the context of non-relativistic systems, 1 the deBroglie-Bohm or pilot-wave approach). Both ap- it is often claimed that consistent histories and Bohmian v proaches have a way of introducing possible histories of mechanics yield the same predictions for outcomes of 1 the universe. In the Bohmian approach, there is an ac- 8 measurements [4, 5]. The reason for this mismatch is tualconfigurationwhosetimeevolutionisdeterminedby 1 that in the consistent histories approach the scalar field the wave function. The possible paths of the configura- 4 is treated (at least formally) as a time variable, whereas 0 tion are the histories. On the other hand, in the consis- it is not in the Bohmian approach. The goal of this pa- . tent histories approach the (coarse-grained) histories are 1 peristoexploreanotherBohmiandynamicswhichisalso sequences of propositions at different moments in time. 0 based on treating the scalar field as a time variable. In 5 (For example, a possible proposition could be that the this case, there is a natural probability distribution for 1 system occupies a certain region in space). the scale factor which agrees with the one of the con- : v A simple quantum cosmological model was studied, sistent histories approach when single-time histories are i namely the quantized Friedmann-Lemaˆıtre-Robertson- considered. Hence, this allows for a better comparison X Walker model, with a homogeneous scalar field. For betweentheBohmianapproachandtheconsistenthisto- r the consistent histories approach, Craig and Singh [2, 3] riesapproachconcerningthequestionofsingularities. In a showed that the possible histories are always singular ir- particular, given a wave function, we can now calculate respective of the given wave function. Either they start the probability for a trajectory to run into a singularity. from a singularity or they end in a singularity. On the We will find that although the trajectories are different otherhand,accordingtotheBohmianapproachofPinto- from the ones in the model proposed in [1], they are of- Neto et al. [1], there are wave functions for which there tenqualitativelythesame. Inparticular,wewillfindthat are also histories that display a bounce (which corre- some wave functions allow for trajectories that are never sponds to a universe that contracts until it reaches a singular and instead display a bounce. More precisely, minimum volume and then starts expanding) and hence we will see that for a non-classical wave function there is do not have a singularity. always a non-zero probability for a bounce (while trajectories corresponding to a classical wave function always have a singularity). This should be contrasted with the fact that in the consistent histories approach the proba- ∗Electronicaddress: [email protected] bility for a singularity is always one, for any wave func- †Electronicaddress: [email protected] tion. ‡Electronicaddress: [email protected] 2 The outline of the paper is as follows. We start integration of (3) yields a = eα = [3(cτ +c˜)]1/3, where by recalling the Wheeler-DeWitt quantization of the c˜is an integration constant, so that a = 0 for τ = −c˜/c Friedmann-Lemaˆıtre-Robertson-Walker model and the (andthereisabigbangifc>oandabigcrunchifc<0). Bohmian approach of [1]. In section III, we will discuss This means that the universe reaches the singularity in possible choices of Hilbert space that were considered in finite proper time. the consistent histories approach. In section IV, we will Canonical quantization of this theory leads to the develop and discuss the alternative Bohmian dynamics. Wheeler-DeWitt equation InsectionV,wecalculatetheprobabilityforatrajectory to be singular. Finally, in section VI, we consider the ∂2ψ−∂2ψ =0. (5) φ α probabilities for singularities in the consistent histories approach, extending the work of Craig and Singh from In the corresponding Bohmian theory [1], there is an ac- two-time histories to many-time histories, and compare tual scalar field φ and scale factor a=eα, which satisfy these to those predicted by the Bohmian approach. N N φ˙ = ∂ S, α˙ =− ∂ S, (6) e3α φ e3α α II. WHEELER-DEWITT QUANTIZATION AND BOHMIAN MECHANICS where ψ = |ψ|eiS. The function N is again the lapse function, which is arbitrary, and, as in the classical case, implies that the dynamics is time reparameterization in- AclassicalflatFriedmann-Lemaˆıtre-Robertson-Walker variant.2 space-time is described by a metric The Wheeler-De Witt equation implies ds2 =N(t)2dt2−e3α(t)γ dxidxj, (1) ij (∂ S)2−(∂ S)2+Q=0, (7) φ α where N is the lapse function, a = eα is the scale factor, and γ is the flat spatial 3-metric. Considering the matterconitjentdescribedbyafreemasslesshomogeneous ∂φ(cid:0)|ψ|2∂φS(cid:1)−∂α(cid:0)|ψ|2∂αS(cid:1)=0, (8) scalar field φ(t), the Lagrangian1 of the system can be written as [6] where (cid:32) (cid:33) 1 1 φ˙2 α˙2 Q=− ∂2|ψ|+ ∂2|ψ| (9) L=Ne3α − , (2) |ψ| φ |ψ| α 2N2 2N2 isthequantumpotential. IfQ=0, then(7)impliesthat where a dot means derivative with respect to coordinate (∂ S)2 = (∂ S)2. In addition, ∂ S is then conserved φ α α time t. The corresponding equations of motion lead to along a Bohmian trajectory. Hence, in this case, the Bohmian motion is reduced to the classical motion given N N φ˙ =± c, α˙ = c, (3) by (3). e3α e3α Equation (8) is a continuity equation and implies that the Bohmian dynamics preserves |ψ|2. However, since where c is an integration constant. N remains an arbi- |ψ|2 is not normalizable it can not be straightforwardly trary function of time. This implies that the dynamics used to make statistical predictions. In this paper, we is time reparameterization invariant. Different choices of will consider an alternative Bohmian dynamics, which N merely correspond to different choices of the parame- allows for immediate statistical predictions and which terization of the paths in (φ,α) space. In the case c=0, allows for a direct comparison with consistent histories the universe is static. For c(cid:54)=0, we have approaches discussed in [2, 3, 14]. α=±φ+c¯, (4) with c¯another integration constant. Theuniversereachesthesingularity(i.e.,zerovolume) 2 Notethatthetimereparameterizationinvarianceisaspecialfea- when a=eα =0. For a non-static universe the singular- ture of Bohmian approaches to mini-superspace models [7, 8]. For the usual formulation of Bohmian dynamics for the full ity is reached for either φ→−∞ or φ→∞. So the uni- Wheeler-DeWitt theory of quantum gravity, a particular space- verseeitherstartswithabigbangorendsinabigcrunch. like foliation of space-time or, equivalently, a particular choice Intermsofpropertimeτ,whichisdefinedbydτ =Ndt, of “initial” space-like hypersurface and lapse function, needs to beintroduced. Differentfoliations(orlapsefunctions)yielddif- ferent Bohmian theories [9–12]. So, in this case, the dynamics is not invariant under space-time diffeomorphisms. (Yet, while theusualBohmianformulationisnotdiffeomorphisminvariant, 1 For simplicity, we have dropped the constant factor 4πG/3 in this may perhaps be achieved with alternative approaches. For frontofα˙2. Thisfactorcouldberemovedbyasuitablerescaling a discussion of analogous issues concerning special relativity in ofthescalarfield. Bohmianmechanics,see[13].) 3 φ which satisfy3 (cid:112) i∂φψ± =H(cid:98)±ψ± =∓ −∂α2ψ±. (15) Theleft-andright-movingcomponentsofψ willbede- ± α noted by ψ and ψ . If one uses the scalarfield φ as L,± R,± a time variable, equation (15) represents a Schrödinger- like equation and given the initial wave function ψ0 at ψR ± ψL time φ0, the (formal) solution is given by √ ψ±(α,φ)=e±i −∂α2(φ−φ0) ψ±0(α). (16) FIG. 1: The wave functions ψ and ψ respectively move to R L The first choice of Hilbert space is H = L2(R,C2) the right and to the left without change in shape. 1 [2, 3, 15]. We have that H = H ⊕ H , with 1 1,+ 1,− H = L2(R,C) the positive and negative frequency 1,± III. WHEELER-DEWITT EQUATION AND Hilbert spaces. We write Ψ(α) = (ψ (α),ψ (α)) with + − HILBERT SPACES Ψ∈H and ψ ∈H .4 The inner product is given by 1 ± 1,± (cid:90) So far we have not introduced a Hilbert space for the (cid:104)Ψ|Φ(cid:105) = dα(cid:0)ψ∗φ +ψ∗φ (cid:1) . (17) 1 + + − − Wheeler-DeWitt equation (5). We will discuss two possible choices of Hilbert space that have appeared in the The dynamics for the positive and negative frequency literature [2, 3, 14–16]. Both are motivated by considering φ (at least formally) as a time variable. However, states is given by (15). Since the Hamiltonians H(cid:98)± are Hermitian operators on H , this dynamics preserves as we shall see, for the purpose of assigning histories to 1,± the inner product. the scale factor, either in the consistent histories or in The natural observable corresponding to the scale fac- the Bohmian framework, both choices can be considered tor is given by the projection-valued measure (PVM) equivalent,assumingasuitablechoiceof“observable”for the scale factor in each case. First, consider a general solution to the Wheeler- P(cid:98)1(dα)=P(cid:98)1,+(dα)⊕P(cid:98)1,−(dα), (18) DeWitt equation (5), which is of the form where ψ(α,φ)=ψ (α+φ)+ψ (α−φ), (10) L R P(cid:98)1,±(dα)=|α(cid:105)(cid:104)α|dα, (19) where the indices L and R denote respectively “left- moving” and “right-moving”. This terminology stems with |α¯(cid:105) = δ(α−α¯). (Since (cid:82) P(cid:98)1,±(dα) = 1(cid:98)±, with 1(cid:98)± from the fact that if φ is regarded as time, then ψ , as theidentityoperatoronH ,itimmediatelyfollowsthat L 1,± a function of α, moves to the “left” over time, without (cid:82) P(cid:98)1(dα)=1(cid:98), the identity operator on H1.) change in shape, whereas ψR moves to the right, see fig- For a state Ψ = (ψ+,ψ−), the corresponding density ure 1. In terms of the Fourier transform, we have for α is 1 (cid:90) ∞ ψ(α,φ)=√ dk ψ (k)eik(α+φ) ρ(α,φ)=|ψ (α,φ)|2+|ψ (α,φ)|2, (20) L + − 2π −∞ + √1 (cid:90) ∞ dk ψ (k)eik(α−φ). (11) i.e., (cid:104)Ψ|P(cid:98)1(dα)|Ψ(cid:105)1 = ρ(α,φ)dα. (The Hilbert space is 2π R thesameasthatofthePaulitheoryforanon-relativistic −∞ The general solution can also be decomposed into a pos- spin-1/2 particle, with P(cid:98)1(dα) being the analogue of the position PVM in the Pauli theory.) itive and negative frequency part The second choice of Hilbert space H is again the 2 ψ =ψ +ψ , (12) direct sum of positive and negative frequency Hilbert + − where 1 (cid:90) ∞ ψ =√ dk ψ (k)eik(α+φ) + 2π L 3 For a wave function ψ(α) = 1 (cid:82)∞ dkψ(k)eikα, the 0 2π −∞ + √12π (cid:90)−0∞dk ψR(k)eik(α−φ), (13) 4 aA21cπctti(cid:82)uo−∞na∞llyo,dfkin|kt[|h2ψe,(3k,H)e1ai5km]αoi.lntolynitahnepiossitdievfienferedquaesncy(cid:112)s−ec∂toα2rψo(fαH) 1=is 1 (cid:90) 0 considered, byappealtosuperselection. However, theBohmian ψ− =√ dk ψL(k)eik(α+φ) trajectorieswilldependonwhetherthewavefunctionisasuper- 2π −∞ position of both frequencies or not. (Following the terminology 1 (cid:90) ∞ of[17],onecouldsaythatfrequencyisweaklysuperselected,but + √ dk ψR(k)eik(α−φ), (14) notstrongly.) Therefore,ageneralanalysisshouldincludeboth 2π 0 frequencysectors. 4 spaces, i.e., H = H ⊕ H , with H = H In the consistent histories approach, (coarse-grained) 2 2,+ 2,− 2,+ 2,− [14, 16]. States in H are maps from R to C2 and histories for α can be introduced, by choosing a Hilbert 2 states in H are maps from R to C. We write Ψ¯(α)= space, a PVM (or positive operator-valued measure 2,± (ψ¯ (α),ψ¯ (α)) with Ψ¯ ∈H and ψ¯ ∈H (the bar is (POVM)) and a Hamiltonian. The results will be iden- + − 2 ± 2,± usedinordertodistinguishthestatesfromthoseinH1). tical when choosing either (H1,P(cid:98)1,H(cid:98)) or (H2,P(cid:98)2,H(cid:98)), The inner product on H is given by 2,± whereH(cid:98) istheHamiltonianoperatordeterminedby(15) (cid:90) (cid:90) (the former is considered in [2, 3], albeit just for positive (cid:104)ψ¯|φ¯(cid:105)2,± =2 dαψ¯∗(∓H(cid:98)±)φ¯=2 dαψ¯∗(cid:112)−∂α2φ¯, frequencies, the latter in [14]).6 The reason is that there (21) is a unitary mapping H2 → H1, which maps P(cid:98)2 → P(cid:98)1 and leaves the Hamiltonian invariant. This mapping so that the inner product for H reads5 2 is defined by ψ¯ (α) → ψ (α) = (cid:104)α,NW|ψ¯ (cid:105) (or ± ± ± 2,± (cid:104)Ψ¯|Φ¯(cid:105)2 =2(cid:90) dα(cid:16)ψ¯+∗(cid:112)−∂α2φ¯++ψ¯−∗(cid:112)−∂α2φ¯−(cid:17) . (22) |ψα,(Nk)W=(cid:105) (cid:112)→2||kα|(cid:105)ψ¯in(kt)erimnsteorfmbsaosfisFsotuartieesr,coormψp¯±on(ken)ts→). ± ± Thismapisunitarysince(cid:104)Ψ|Φ(cid:105) =(cid:104)Ψ¯|Φ¯(cid:105) (whichfollows 1 2 The dynamics for the positive and negative frequency from the fact that (cid:82) P(cid:98)2(dα) = 1(cid:98), the identity operator states is given again by (15). Since the Hamiltonians H(cid:98)± are Hermitian operators on H2,±, the dynamics also on H2). Since it takes P(cid:98)2 →P(cid:98)1, we have preFsoerrvtehsetnhoisrminnoefrapwroadvuectfu.nction Ψ¯, the integrand on (cid:104)Ψ¯|P(cid:98)2(dα)|Ψ¯(cid:105)2 =(cid:104)Ψ|P(cid:98)1(dα)|Ψ(cid:105)1 =ρ(α,φ)dα. (27) the right hand side of (22) reads As such, we clearly get the same result for probabilities 2(cid:16)ψ¯∗(cid:112)−∂2ψ¯ +ψ¯∗(cid:112)−∂2ψ¯ (cid:17) . (23) ofWhiestocrainesawlsitohd(Hev1e,loP(cid:98)p1)aorB(oHh2m,Pi(cid:98)a2n).approach given a + α + − α − Hilbert space, a POVM and a Hamiltonian. For a One might be tempted to regard this quantity as the given wave function, the distribution corresponding to densityofαatagiventimeφforthestateΨ¯. However,it the PVM will be the Bohmian equilibrium distribution. isnotalwayspositivedefinite,evenifwerestrictourselves Thedynamicscanthenbedefinedinsuchawaythatthis to positive frequencies [18, 19]. distribution is preserved [22, 23]. In the next section, One possibility to obtain a positive density is by using we will consider a dynamics that preserves the density the Newton-Wigner states [14, 20] (20). So also in the Bohmian approach, the two choices 1 (cid:90) ∞ dk (H1,P(cid:98)1,H(cid:98)) and (H2,P(cid:98)2,H(cid:98)) yield an identical dynamics |α¯,NW(cid:105)= eik(α−α¯). (24) for the scale factor. (cid:112) 2π 2|k| −∞ These states are normalized according to IV. ALTERNATIVE BOHMIAN DYNAMICS (cid:104)α,NW|α¯,NW(cid:105) = δ(α − α¯) and form a basis of 2 H . The PVM then reads 2,± The density ρ(α,φ), given in (20), is not preserved by theBohmiandynamics(6),i.e.,ifthedensityofαisgiven P(cid:98)2(dα)=P(cid:98)2,+(dα)⊕P(cid:98)2,−(dα), (25) by ρ(α,φ ) at a certain time φ , then the Bohmian dy- 0 0 namicsingeneralleadstoadensitydifferentfromρ(α,φ) where now atothertimesφ[1]. However,onecanpostulateadiffer- ent dynamics that preserves ρ [23–25]. According to this P(cid:98)2,±(dα)=|α,NW(cid:105)(cid:104)α,NW|dα. (26) dynamics, possible trajectories α(φ) satisfy Note that, as before, (cid:82) P(cid:98)2(dα) = 1(cid:98), being understood (cid:90) α(φ) (cid:90) α(0) that the action of these operators on the states in the dα¯ρ(α¯,φ)= dα¯ρ(α¯,0), (28) Hilbert space H2 is through the inner product defined in −∞ −∞ Eq. (22). or, equivalently, dα 1 (cid:90) α =v (α,φ)=− dα¯∂ ρ(α¯,φ). (29) dφ 2 ρ(α,φ) φ −∞ 5 Perhaps a more familiar way of writing the inner products (21) and(22)isintermsofsolutionstothewaveequations(15)[14]. Since ρ satisfies the continuity equation The inner product (21) then corresponds to the Klein-Gordon innerproduct(cid:104)ψ¯|φ¯(cid:105)KG=−i(cid:82) dα(cid:0)ψ¯∗∂φφ¯−φ¯∂φψ¯∗(cid:1)fortwopos- ∂φρ+∂α(v2ρ)=0, (30) itive frequency solutions to (15). The inner product (22) then correspondsto(cid:104)Ψ¯|Φ¯(cid:105)2 =(cid:104)ψ¯+|φ¯+(cid:105)KG−(cid:104)ψ¯−|φ¯−(cid:105)KG. Notethat thisinnerproductispositivedefinite,unliketheusualdefinition fortheinnerproductoftwosolutionstotheKlein-Gordonequa- tion,whichdoesnotcontaintherelativeminussignbetweenthe 6 See [21] for similar considerations on equivalence of quantum positiveandnegativefrequencypart. theories. 5 it is preserved by the Bohmian dynamics. functionandhencewedonothaveastraightforwarddef- Let us compare this dynamics to the one formulated inition of proper time. So, it is not immediately clear in section II. First, consider a real wave function ψ, i.e. whetherthesingularitiesarereachedinfinitepropertime Imψ = 0.7 According to the dynamics of section II, ornot. Nonetheless,sincethesetrajectoriesareclassical, the possible solutions are α(t) = α(0), φ(t) = φ(0), so it seems natural to introduce a proper time that agrees that the universe is static, with a constant scale factor with that of the classical theory. As such, the conclusion and constant scalar field. Since these solutions can not isagainthatthesingularitiesarereachedinfiniteproper be expressed as α(φ) they must be different from the time. trajectories given by (28). This dynamics seems less natural than the one given NowassumethatthetrajectoriesforthefirstBohmian earlier. The reason is of course that φ plays the role dynamics can be expressed as α(φ) (at least locally). If of time in this dynamics. However, there seems to be no ∂ S (cid:54)=0, then the vector field tangent to these trajecto- reasontogiveitthatdistinguishedrole. Forexample,we φ ries is given by could equally well have taken α to play the role of time. This would have resulted in completely different paths. ∂ S v (α,φ)=− α . (31) Of course, this is only a simplified model. In a more 1 ∂φS serious model, we expect that some of the matter fields will represent time on an effective level. But this should Thevelocityfieldsv andv aregenericallynotthesame. 1 2 follow from analyzing a more fundamental Bohmian dy- Using (7) and (8), to evaluate the integral in (29), we namics, similar to (6) (see e.g. [9, 26, 27]), rather than have that be postulated a priori. 1(cid:90) α ρ (cid:104) v =v − dα¯ ∂ Q 2 1 ρ 2(∂ S)2 φ −∞ φ V. SINGULARITIES AND BOHMIAN 2∂ S (cid:105) + φ {∂ (ρ∂ S)−∂ (ρ∂ S)} . MECHANICS ρ φ φ α α (32) In this section, we consider the possibility of singularities (zero scale factor) for trajectories given by the So a wave function ψ will lead to the same trajectories Bohmian dynamics of section IV . In particular, using only if the equilibrium distribution (20) as probability distribu- 2∂ S tion,wecalculatetheprobabilityforatrajectorytohave ∂ Q+ φ {∂ (ρ∂ S)−∂ (ρ∂ S)}=0. (33) φ ρ φ φ α α a singularity, for a given wave function. Westartbyconsideringtheasymptoticbehaviorofthe Generically the above condition will not hold. For ex- Bohmiantrajectories. First,supposethatthewavefunc- ample, consider a positive or negative frequency wave tion has only right-moving components, i.e., Ψ(α,φ) = function. Then the term in curly brackets is zero (be- Ψ (α − φ) = (ψ (α − φ),ψ (α − φ)). Then the R R,+ R,− cause the density ρ then equals |ψ|2 and hence satisfies support of the wave function is localized on α < 0 for (8)) and the condition is reduced to ∂φQ = 0, which is φ→−∞ and α>0 for φ→+∞. This follows from the generically not satisfied. fact that On the other hand, if the wave function ψ is ei- (cid:90) 0 ther only left- or right-moving, i.e., ψ = ψL(α+φ) or lim dαρΨR(α,φ) ψ =ψR(α−φ), then the condition (33) is automatically φ→−∞ −∞ satisfied. In this case, the trajectories are the same for (cid:90) 0 bothBohmianapproaches. Inaddition,thequantumpo- = lim dα(cid:0)|ψR,+(α−φ)|2+|ψR,−(α−φ)|2(cid:1) tential (9) is zero, so that the trajectories are classical. φ→−∞ −∞ (Conversely, if Q=0, then the wave function ψ is either = lim (cid:90) −φdν(cid:0)|ψ (ν)|2+|ψ (ν)|2(cid:1) left-orright-movingandthetrajectoriesarethesamefor R,+ R,− φ→−∞ −∞ both Bohmian approaches.) As shown in section II, the =||Ψ ||2 =1 (34) classical trajectories are given by α = ±φ+c, with c a R 1 constant. Thepositiveandnegativesignrespectivelycor- and, similarly, for the other limit we have respond to a right- and left-moving wave function. The trajectories reach the singularity a=eα =0 respectively (cid:90) ∞ lim dαρΨR(α,φ) at φ→−∞ and φ→∞. Note, however, that in our sec- φ→+∞ 0 ond Bohmian approach, we have not introduced a lapse (cid:90) ∞ = lim dα(cid:0)|ψ (α−φ)|2+|ψ (α−φ)|2(cid:1) R,+ R,− φ→+∞ 0 (cid:90) ∞ = lim dν(cid:0)|ψ (ν)|2+|ψ (ν)|2(cid:1) R,+ R,− 7 NotethatinsectionII,wewrotethewavefunctionasψ=ψ++ φ→+∞ −φ ψ−, whereasinsectionIIIwewroteitasΨ=(ψ+,ψ−). These =||Ψ ||2 =1. (35) notationsareofcourseequivalent. R 1 6 In the same manner, for a wave function that has only φ left-moving components, i.e., Ψ(α,φ) = Ψ (α + φ) = L (ψ (α+φ),ψ (α+φ)), we have that L,+ L,− (cid:90) ∞ lim dαρΨL(α,φ) φ→−∞ 0 (cid:90) ∞ α = lim dα(cid:0)|ψ (α+φ)|2+|ψ (α+φ)|2(cid:1) L,+ L,− φ→−∞ 0 (cid:90) ∞ = lim dν(cid:0)|ψ (ν)|2+|ψ (ν)|2(cid:1) L,+ L,− φ→−∞ φ =||Ψ ||2 =1 (36) L 1 FIG. 2: Asymptotic classical behavior of the Bohmian tra- and jectories for an arbitrary wave function. (cid:90) 0 lim dαρΨL(α,φ) φ→+∞ −∞ 3. starts with a big bang and ends in a big crunch (cid:90) 0 = lim dα(cid:0)|ψ (α+φ)|2+|ψ (α+φ)|2(cid:1) L,+ L,− φ→+∞ −∞ 4. undergoes a bounce, i.e., contracts until it reaches (cid:90) φ a minimal size and then expands again. = lim dν(cid:0)|ψ (ν)|2+|ψ (ν)|2(cid:1) L,+ L,− φ→+∞ −∞ Only trajectories of type 4 are non-singular. As we shall =||ΨL||21 =1, (37) see, allfourtypesoftrajectoriesmayoccur. Though, for a given wave function, there can not be trajectories of sothatthesupportisnowlocalizedonα>0forφ→−∞ both types 1 and 2. This is an immediate consequence and α<0 for φ→+∞. of the no-crossing property of the Bohmian dynamics So in summary, we have that (whichstatesthattrajectoriescannotcrossatequaltimes  inconfigurationspace),sincesuchtrajectoriesmustcross α<0 for φ→−∞  each other. ΨR has support on , Intheprevioussection,weshowedthatwavefunctions α>0 for φ→∞ that are purely left- or right-moving give rise to classical  trajectories, which are either of type 1 or type 2, and α>0 for φ→−∞  which are singular. There are also wave functions for ΨL has support on . which there are trajectories of type 3 or 4. α<0 for φ→∞ As a simple example, consider a wave function which is even or odd in α. For such a wave function, the ve- locity field v is odd in α. As such, for every possible This analysis shows that for any wave function Ψ = 2 trajectory α(φ), also −α(φ) is a possible trajectory. Be- Ψ +Ψ , its left- and right-moving components will be- R L cause of the no-crossing property, this implies that no come asymptotically non-overlapping in α-space. As a trajectory will cross the φ-axis.9 In other words, trajec- consequence, asymptotically, Bohmian trajectories will tories that have positive value of α at any given time φ, be determined either by the Ψ or Ψ . Hence, asymp- R L must have α>0 at all times and hence avoid the singu- totically,thetrajectoriesareclassical,givenbyα=φ+c larity. These trajectories correspond to a bouncing uni- or α = −φ+c, see figure 2.8 (The same holds for the verse. Theyasymptoticallystartwithα=−φ+c ,reach BohmiandynamicsofsectionII,inthecasethatthetra- i aminimalvalueα andthenasymptoticallyevolveac- jectories case be expressed as α(φ).) min cording to α=φ+c for φ→∞. See figure 3 for an ex- This asymptotic classical behavior implies that there f ample. Conversely,trajectoriesthatasymptoticallystart may be four different types of trajectories. Namely, tra- with α=φ+c˜ originate from a singularity. Eventually jectories that represent a universe that: i they reach a maximal vale α and then for φ → ∞ max they move again to the singularity according to the tra- 1. startswithabigbangandkeepsexpandingforever jectory α = −φ+c˜ . So these trajectories are big bang f 2. keeps contracting until a big crunch - big crunch solutions. 8 There might also be a trajectory that does not display this 9 Their might also be the trajectory α(φ) = 0 which acts as a asymptotic behavior and acts as a bifurcation line between tra- bifurcation line between trajectories with different asymptotic jectorieswithdifferentpossibleasymptoticbehavior. behavior. 7 φ the probability for trajectories coming from α > 0 to keep contracting towards the singularity is P = contracting 0. Similarly, for the case that P > P , one can R,i R,f show that P = P , P = P − P , bounce R,f contracting L,f L,i P =P and P =0. recollapsing L,i expanding In summary, for an arbitrary state Ψ, we have α P =P =min(P ,P )=min(P ,P ), bounce recollapsing L,i L,f R,i R,f P =max(P −P ,0)=max(P −P ,0), expanding L,i L,f R,f R,i P =max(P −P ,0)=max(P −P ,0). contracting L,f L,i R,i R,f (40) ThisimpliesthattheprobabilityP =1−P FIG. 3: Some trajectories for a wave function that is sym- singularity bounce to run into a singularity satisfies metric under α→−α. The trajectories on the left represent universeswhichstartwithabigbangandendinabigcrunch. 1 Thetrajectoriesontherightrepresentuniversesthatbounce. 2 (cid:54)Psingularity (cid:54)1, (41) i.e., the probability to run into a singularity is at least Let us now analyze the general case and establish the one half. Maximum probability is reached when there probability to have singular trajectories for a general are only left- or right-moving components. In that case, wavefunctionΨ=Ψ +Ψ . CallP andP theprob- every trajectory either runs into a singularity or starts R L L,i R,i ability for the universe to start respectively with α < 0 from a singularity. On the other hand, for a superpo- (from the left) and α > 0 (from the right) in the limit sition of left- or right-moving components, there is al- φ→−∞. Similarly, P and P denote the probabil- ways a non-zero probability for a bounce. For a state L,f R,f ities for the universe to end up respectively with α < 0 Ψ, the maximum probability that can ever be attained and α>0 in the limit φ→∞. Since Ψ and Ψ do not foratrajectorytobenon-singularis1/2, whichhappens R L overlap asymptotically, we have when P =P . This happens for the examples of the L,i L,f symmetric and anti-symmetric states discussed before.10 (cid:90) 0 P = lim dαρΨ(α,φ)=||Ψ ||2, The trajectories for the first Bohmian approach, that L,i φ→−∞ −∞ R 1 was presented in section II, are different from those con- (cid:90) ∞ sidered here. However, the trajectories in the first ap- PR,i = lim dαρΨ(α,φ)=||ΨL||21, proachthatareexpressibleasα(φ)havethesameasymp- φ→−∞ 0 totic behavior as those considered here and hence are (cid:90) 0 P = lim dαρΨ(α,φ)=||Ψ ||2, qualitatively the same. However, there is no natural L,f L 1 φ→∞ −∞ probability distribution over trajectories which allows to (cid:90) ∞ calculate the probability for trajectories to be singular. P = lim dαρΨ(α,φ)=||Ψ ||2, (38) R,f R 1 φ→∞ 0 and VI. SINGULARITIES AND CONSISTENT HISTORIES P =1−P =P =1−P . (39) L,i L,f R,f R,i According to the consistent histories approach the We can now find the probability for the universe to probability for a singularity is always one. Coarse- start or to end in a singularity for a given wave func- grainedhistoriesareeitheroftype1or2. Thiswasshown tion Ψ. First, assume that P < P . In this case, R,i R,f for two-time histories in [2, 3], with the times being the trajectories starting from α > 0 as φ → −∞ can not infinite past and future. end up moving to α < 0 as φ → ∞. If a trajec- In [1], some of us argued that introducing a third in- tory did do that, then all the trajectories that started termediate time would not lead to a family of consistent on the α < 0 would must also end up with α < 0 as φ → ∞ because of the no-crossing property. This implies that P (cid:62) P and hence, because of (39), that L,f L,i P (cid:62)P , but this contradicts our assumption. Thus, R,i R,f 10 Theexamplethatwasusedin[4,5]tocomparetheBohmianap- inthecasethatP <P ,theprobabilityforabounce R,i R,f proachtotheconsistenthistoriesapproachinthecontextofnon- is Pbounce = PR,i. In addition, it implies that the prob- relativisticquantummechanicsisactuallycompletelyanalogous ability for trajectories to start from a singularity and to the case of a symmetric state in our cosmological model. In to keep expanding is P = P −P and the [4,5],thewavefunctioniscomposedoutoftwopacketsthatcross expanding R,f R,i eachotherintime. Thewavefunctionisconsideredtobecom- probability for trajectories to start and end in a singu- pletely symmetric so that Bohmian trajectories will never cross larity (i.e., trajectories with a big bang and big crunch) thesymmetryaxis,whilethehistoriesintheconsistenthistories is Precollapsing = PL,i −(PR,f −PR,i) = PR,i. Finally, approachcorrespondtoeitherleft-orright-movingpaths. 8 histories, unless the state is classical (i.e. the state is a Given a wave function Ψ (and given that φ is suf- 1 left- or right-moving wave function). As such, for non- ficiently far in the past or φ sufficiently far in the fu- n classical states, the consistent histories approach would ture), we canalwaysfind acollections ofsets {∆k } such ik be unable to deal with the properties of the universe at that the decoherence condition is satisfied. The reason intermediate times. This would of course undermine the is that the wave function is a superposition of a left- consistenthistoriesapproachforthiscosmologicalmodel. moving and a right-moving packet. For each time φ , k However,theargumentin[1]isincomplete,mainlybe- we can choose the regions {∆k } such that support of ik cause only one particular family of histories was consid- Ψ and Ψ are each approximately within one of the L R ered. That is, the class of propositions (which are repre- ∆k (but not necessarily with the same i ). We then ik k saesnttheadtbiynp[2r,o3je]cftoironalloptherreaetotrism)ews.asAtnadkefnortothbisectlhaesss,atmhee have that for each time φk that P(cid:98)∆kkikΨL(α,φk) ≈ 0 for historiesarenotconsistent(unlesstheintermediatetime all but one ∆kik, which we denote ∆kL. For ∆kL, we have is chosen sufficiently early or late). However, this does thatP(cid:98)∆kkΨL(α,φk)≈ΨL(α,φk). Wethenalsohavethat not exclude other possible families of histories for which L thehistoriesareconsistent. Wewillshowthatasuitable C(cid:98)hLΨL(α,φ0)≈ΨL(α,φn), with hL =(∆1L,...,∆nL). In other words, there is a coarse-grained history that ap- class of propositions can be chosen such that consistent proximatelyfollowsthetrackofthepacketΨ . Similarly, historiescanbeconsideredforanarbitrarynumberofin- L there is a coarse-grained history h that approximately termediate times.11 With this choice, the probability for R follows the track of the packet Ψ . a singularity is always one. This generalizes the results R If one of the φ is such that Ψ and Ψ are approxi- in [2, 3] to an arbitrary number of times. k L R matelynon-overlappingatthattime(forexamplebytak- To setup the consistent histories framework, we will ing φ sufficiently early or φ sufficiently late, since we 1 n use the triplet (H1,P(cid:98)1,H(cid:98)). As explained in section III, know that the overlap vanishes in the limit φ → ±∞.), (H2,P(cid:98)2,H(cid:98)) would lead to the same results. then A coarse-grained history is a sequence of regions ∆1,...,∆n in α-space at a sequence of times φ1,...,φn. C(cid:98)hLΨ(α,φ0)≈ΨL(α,φn), C(cid:98)hRΨ(α,φ0)≈ΨR(α,φn) (46) These are obtained by considering an exhaustive set of regions {∆kik}, ik = 1,2,..., of α-space for each time and φ , k = 1,...,n. For each of these regions, there is a k projection operator C(cid:98)hΨ(α,φ0)≈0 , for all h(cid:54)=hL,hR. (47) (cid:90) Since12 P(cid:98)∆kkik = ∆kik P(cid:98)1(dα) (42) (cid:104)C(cid:98)hLΨ|C(cid:98)hRΨ(cid:105)≈0, (48) we have that the decoherence condition (44) is satisfied. defined in terms of the PVM P(cid:98)1(dα). In terms of these Theonlytwocoarse-grainedhistorieswithnon-zeroprob- projectionoperators,theclassoperatorforahistoryh= ability are h and h , with probabilities (∆1 ,...,∆n ) is L R i1 in P ≈||Ψ ||2 , P ≈||Ψ ||2. (49) C(cid:98)h =P(cid:98)∆nninU(cid:98)(φn,φn−1)P(cid:98)∆n−nin−1−11U(cid:98)(φn−1,φn−2) ··· In orderhLto consLide1r the questiohnR whethRer1there is a ··· P(cid:98)∆22i2U(cid:98)(φ2,φ1)P(cid:98)∆11i1U(cid:98)(φ1,φ0), (43) sTinhgenula(4ri6t)y-,(4w9e) hshoolduladndta,kuesiφn1g a→si−m∞ilarannodtaφtnion→as∞in. the previous section, we have where φ0 is some initial time and U(cid:98)(φk,φk−1) = e−iH(cid:98)(φk−φk−1) is the unitary time evolution from time Pbounce =Precollapsing ≈0, φkI−n1otrodφerk.to associate probabilities to a family of histo- Pexpanding =PhL ≈||ΨL||21, P =P ≈||Ψ ||2. (50) ries, the decoherence condition contracting hR R 1 Theprobabilityforabounceisnegligiblysmall. Itcan (cid:104)C(cid:98)h(cid:48)Ψ|C(cid:98)hΨ(cid:105)1 ≈0, for h(cid:48) (cid:54)=h, (44) actually be made arbitrarily small by suitably choosing needs to be satisfied. The probability for a history h is the set {∆1 }. So according to the consistent histories i1 then given by approach, one can always find a consistent family of histories where the probability for a singularity is one. Ph =||C(cid:98)hΨ||21. (45) 12 Eq. (48) follows from the fact that a left-moving packet and 11 Theanalysisiscompletelyanalogoustothatof[4,5]. Thereason right-movingpacketareorthogonal. Sincetheyhavenocommon is that, as already noted in footnote 10, we are considering two support at the far past they are clearly orthogonal then. Since packetsthatmoveacrosseachother,justlikeintheexampleof theinnerproductispreservedovertime,theymustbeorthogonal [4,5]. atalltimes. 9 VII. CONCLUSION DeWitt quantization. It would be interesting to also study the Bohmian approach to loop quantization. For We have analyzed a Bohmian approach to the the consistent histories approach to loop quantization, it Wheeler-DeWitt quantization of the Friedmann- was recently shown that histories do not have singulari- Lemaˆıtre-Robertson-Walker model. This Bohmian tiesforgenericwavefunctions[28]. Itisunclearwhether approach agrees with the consistent histories approach this is also true for a Bohmian approach. concerning the probabilities for single-time histories. However, it makes different predictions for the probability of trajectories or histories to have a singularity. In Acknowledgments theconsistenthistoriesapproach,atleastforthefamilies considered in the present paper, the probability for a history to have a singularity is one. On the other hand, F.T.F.andN.P.-N.wouldliketothankCNPqofBrazil in the Bohmian approach, for generic wave functions for financial support. W.S. acknowledges support from (i.e., non-classical wave functions), there is a non-zero the Actions de Recherches Concertées (ARC) of the Bel- probability for a trajectory to be non-singular and have gium Wallonia-Brussels Federation under contract No. a bounce. 12-17/02. W.S. carried out part of this work at Rutgers So, as was already emphasized in [1], where a different University, with a grant from the John Templeton Foun- Bohmian model was considered to make the compari- dation. The opinions expressed in this publication are son with the consistent histories approach, the question those of the authors and do not necessarily reflect the whetherornottheWheeler-DeWittquantizationleadsto views of the John Templeton Foundation. W.S. is also singularities depends very much on the version of quan- grateful to CBPF where part of this work was carried tum theory one adopts. out and to Sheldon Goldstein and Roderich Tumulka for In this paper, we have only analyzed the Wheeler- helpful discussions. [1] N. Pinto-Neto, F.T. Falciano, R. Pereira and E. Sergio [11] E. Sergio Santini, PhD Thesis, CBPF-Rio de Janeiro Santini,“Wheeler-DeWittquantizationcansolvethesin- (2000) and arXiv:gr-qc/0005092. gularity problem”, Phys. Rev. D 86, 063504 (2012) and [12] N.Pinto-NetoandE.SergioSantini,“TheConsistencyof arXiv:1206.4021 [gr-qc]. CausalQuantumGeometrodynamicsandQuantumField [2] D.A.CraigandP.Singh,“ConsistentHistoriesinQuan- Theory”, Gen. Rel. and Grav. 34, 505-532 (2002) and tum Cosmology, Found. Phys. 41, 371-379 (2011) and arXiv:gr-qc/0009080. arXiv:1001.4311 [gr-qc]. [13] D. Du¨rr, S. Goldstein, T. Norsen, W. Struyve and [3] D.A. Craig and P. Singh, “Consistent probabilities in N. Zangh`ı, “Can Bohmian mechanics be made rela- Wheeler-DeWitt quantum cosmology, Phys. Rev. D 82, tivistic?”, Proc. R. Soc. A 470, 20130699 (2014) and 123526 (2010) and arXiv:1006.3837 [gr-qc]. arXiv:1307.1714 [quant-ph]. [4] R.B.Griffiths,“Bohmianmechancisandconsistenthisto- [14] J.J. Halliwell and J. Thorwart, “Decoherent histories ries”,Phys.Lett.A261,227-324(1999)andarXiv:quant- analysis of the relativistic particle”, Phys. Rev. D 64, ph/9902059. 124018 (2001) and arXiv:gr-qc/0106095. [5] J.B. Hartle, “Bohmian histories and decoherent histo- [15] A. Ashtekar, T. Pawlowski and P. Singh, “Quantum na- ries”, Phys. Rev. A 69, 042111 (2004) and arXiv:quant- tureofthebigbang: Improveddynamics”,Phys. Rev. D ph/0209104. 74, 084003 (2006) and arXiv:gr-qc/0607039. [6] J.J. Halliwell, “Introductory lectures on quantum cos- [16] A.Ashtekar,A.CorichiandP.Singh,“Robustnessofkey mology”, in Quantum Cosmology and Baby Universes, features of loop quantum cosmology”, Phys. Rev. D 77, eds. S. Coleman, J.B. Hartle, T. Piran and S. Wein- 024046 (2008) and arXiv:0710.3565 [gr-qc]. berg, World Scientific, Singapore, 159-243 (1991) and [17] S. Colin, T. Durt and R. Tumulka, “On superselection arXiv:0909.2566 [gr-qc]. rulesinBohmBelltheories”,J. Phys. A39,15403-15419 [7] J.AcaciodeBarrosandN.Pinto-Neto,“TheCausalIn- (2006) and arXiv:quant-ph/0509177. terpretation of Quantum Mechanics and the Singularity [18] A.Kyprianidis,“Particletrajectoriesinrelativisticquan- Problem and Time Issue in Quantum Cosmology”, Int. tum mechanics”, Phys. Lett. A 111, 111-116 (1985). J. Mod. Phys. D 07, 201-213 (1998). [19] P.R. Holland, The quantum Theory of Motion, Cam- [8] F.T. Falciano, N. Pinto-Neto and E. Sergio Santini, bridge University Press, Cambridge (1993). “Inflationarynonsingularquantumcosmologicalmodel”, [20] T.D.NewtonandE.P.Wigner,“LocalizedStatesforEl- Phys.Rev.D76,083521(2007)andarXiv:0707.1088[gr- ementarySystems”,Rev.Mod.Phys.21,400-406(1949). qc]. [21] S. Goldstein, J. Taylor, R. Tumulka and N. Zangh`ı, [9] Y.V. Shtanov, Phys. Rev. D 54, “Pilot wave quantum “Fermionic Wave Functions on Unordered Configura- cosmology”, 2564-2570 (1996) and arXiv:gr-qc/9503005. tions”, arXiv:1403.3705 [quant-ph]. [10] N. Pinto-Neto and E. Sergio Santini, “Must Quantum [22] D.Du¨rr,S.Goldstein,R.TumulkaandN.Zangh`ı,“Bell- Spacetimes Be Euclidean?”, Phys. Rev. D 59, 123517 type quantum field theories”, J. Phys. A 38, R1-R43 (1999) and arXiv:gr-qc/9811067. (2005) and arXiv:quant-ph/0407116v1. 10 [23] W. Struyve and A. Valentini, “de BroglieBohm guid- tions of quantum gravity”, in Physics Meets Philosophy ance equations for arbitrary Hamiltonians” J. Phys. A at the Planck Scale, eds. C. Callender and N. Huggett, 42, 035301 (2009) and arXiv:0808.0290v3 [quant-ph]. Cambridge University Press, Cambridge, 275-289 (2004) [24] K.Berndl,“GlobalexistenceanduniquenessofBohmian and arXiv:quant-ph/9902018. trajectories”, in Bohmian Mechanics and quantum [27] N. Pinto-Neto, “The Bohm Interpretation of Quan- Theory: An Appraisal, eds. J.T. Cushing, A. Fine tum Cosmology”, Found. Phys. 35, 577-603 (2005) and and S. Goldstein, Kluwer, Dordrecht, 77 (1996) and arXiv:gr-qc/0410117. arXiv:quant-ph/9509009. [28] D.A. Craig and P. Singh, “Consistent probabilities in [25] S. Goldstein, “Absence of chaos in Bohmian dynamics”, loop quantum cosmology”, Class. Quantum Grav. 30, Phys. Rev. E 60, 7578-7579 (1999) and arXiv:quant- 205008 (2013) and arXiv:1001.4311 arXiv:1306.6142 [gr- ph/9901005. qc]. [26] S.GoldsteinandS.Teufel,“Quantumspacetimewithout observers: ontologicalclarityandtheconceptualfounda-