Exact Correspondences Between the Hartle–Hawking and Vilenkin Boundary Conditions Under Randall-Sundrum Brane Dynamics

1. Introduction

Singularity. That word concentrates everything recalcitrant about classical general relativity when pressed to its early-universe limits. The Penrose–Hawking singularity theorems [1,2] establish—under the null energy condition $R_{\mu \nu }{u}^{\mu }{u}^{\nu }\ge 0$ and the requirement that at least one trapped surface forms—that any spacetime satisfying Einstein’s field equations must be geodesically incomplete. Incompleteness is not a coordinate artefact; it is the statement that timelike geodesics terminate, in finite proper time, at a locus where curvature invariants diverge and every classical description fails. The Big Bang is the canonical instance: the standard Friedmann cosmology extrapolated backward reaches, at $t=0$, a state of infinite energy density with no classical antecedent, no initial condition, and no principled dynamical explanation for the geometry our universe inherited.

Quantum cosmology was constructed to dissolve this impasse. In its canonical form [4,5], the four-dimensional spacetime metric is promoted to a quantum-mechanical degree of freedom; the classical Hamiltonian constraint $\mathcal{H}=0$ of the ADM decomposition [3] becomes, after canonical quantization, the Wheeler–DeWitt equation $\widehat{\mathcal{H}}\Psi =0$—a constraint equation on the space of all three-metrics and matter configurations, known as superspace. Two principled proposals for selecting the physical solution have generated sustained debate. Hartle and Hawking [6] advocated the no-boundary wave function, computed via a Euclidean path integral summed over compact four-geometries without initial boundary; time—understood as the WKB phase of $\Psi$—simply ceases to exist in the deep quantum régime, as naturally as spatial directions cease at the south pole of a sphere, and no “before” can be asked. Vilenkin [7,8], by contrast, required only outgoing modes at classically forbidden boundaries: the universe nucleates by a quantum tunnel from a state of literally nothing, an event with no classical analog and no preceding epoch. These proposals differ not only in physical picture but in a sign: Hartle–Hawking selects the growing Euclidean mode $e^{+|S_E|/\hslash }$, strongly preferring large cosmological constants, while Vilenkin selects the decaying mode $e^{-|S_E|/\hslash }$, preferring small-$\Lambda$ inflationary seeds [22]. Neither prediction has been definitively tested; the tension has never been resolved within purely four-dimensional quantum cosmology.

Randall–Sundrum brane-world models [9,10] enrich this landscape without immediately resolving it, but they do so in a geometrically transparent way that admits precise mathematical correspondences with both boundary conditions. In RS1, our universe is a positive-tension 3-brane at the boundary of an AdS₅ bulk whose curvature radius is set by a negative bulk cosmological constant; the exponential warp factor of the RS metric generates the gauge hierarchy $M_{\mathrm{weak}}/M_{\mathrm{P}}\sim {10}^{-15}$ without parameter fine-tuning beyond the modest requirement $k{r}_c\approx 11$ [9]. In RS2, a single brane in an infinite AdS₅ bulk localizes four-dimensional gravity via a normalizable zero mode, producing Newtonian gravity at macroscopic scales with Kaluza–Klein corrections at sub-millimetre distances [10,29]. Cosmologically, the embedding of a Friedmann brane in the RS bulk modifies the Friedmann equation at high densities [12,13,14], introducing corrections quadratic in $\rho$ that alter the structure of the WdW effective potential and therefore the tunneling amplitudes of both quantum cosmological proposals.

The present paper develops three interlocking results. The first is a systematic derivation of the RS-modified WdW equation—a second-order functional differential equation on brane minisuperspace—and an identification of two structurally new terms that deform both the Hartle–Hawking and Vilenkin exponents in computable directions. The second is a table of exact correspondences, operating at the levels of Euclidean path-integral saddle points, WKB exponents, and AdS/CFT holographic duals, between four-dimensional quantum cosmological boundary conditions and five-dimensional RS dynamics. The third—and physically most consequential—is a thermodynamic argument demonstrating that classical information channels are categorically erased at a cosmological bounce, while the phase of the universal wave function propagates unitarily across the bounce, constituting the unique pre-geometric memory channel; we define and prove the invariance of a phase entropy functional $S_{\theta }$ and a phase fidelity${\mathcal{F}}_{bounce}$ under any unitary bounce operator, and identify indirect observational consequences in the CMB bispectrum and initial state of inflation.

2. Wheeler–DeWitt Equation and Quantum Cosmological Boundary Conditions

The starting point is the ADM form [3] of the four-dimensional Einstein–Hilbert action with a minimally coupled inflaton field $\varphi$:

$$S={\displaystyle \frac{M_{\mathrm{P}}^2}{2}}\int d^{4}x\sqrt{-g}R-\int d^{4}x\sqrt{-g}\left[{\displaystyle \frac{1}{2}}{(\partial \varphi )}^2+V\left(\varphi \right)\right].$$

(1)

Restricting to closed ($k=+1$) FLRW metrics $d{s}^2=-N^2\left(t\right)d{t}^2+a^2\left(t\right)d{\Omega }_3^2$, the minisuperspace (MSS) action is

$$S_{\mathrm{ms}}=\int dtN\left[-{\displaystyle \frac{3M_{\mathrm{P}}^2}{N^2}}\left(a{\dot{a}}^2-Na\right)+{\displaystyle \frac{a^3{\dot{\varphi }}^2}{2N^2}}-a^{3}V\left(\varphi \right)\right].$$

(2)

The momenta conjugate to a and $\varphi$ are ${\pi }_a=-6M_{\mathrm{P}}^{2}a\dot{a}/N$ and ${\pi }_{\varphi }=a^3\dot{\varphi }/N$. Canonical quantization ${\pi }_a\to -i\hslash \partial /\partial a$, ${\pi }_{\varphi }\to -i\hslash \partial /\partial \varphi$ converts the Hamiltonian constraint $\mathcal{H}=0$ into the Wheeler–DeWitt equation:

$$\begin{array}{|c|}\hline \widehat{\mathcal{H}}\Psi [a,\varphi ]=\left[-{\displaystyle \frac{{\hslash }^2}{2M_{\mathrm{P}}^2}}{\displaystyle \frac{{\partial }^2}{\partial a^2}}+U_{\mathrm{eff}}(a,\varphi )\right]\Psi =0,\\ \hline\end{array}$$

(3)

where (in units $M_{\mathrm{P}}=c=1$, $k=+1$, operator-ordering $\xi =0$):

$$U_{\mathrm{eff}}(a,\varphi )=-a^2+a^{4}V\left(\varphi \right).$$

(4)

Three features of (3) are worth underscoring. No explicit time derivative appears—$\widehat{\mathcal{H}}\Psi =0$ is a timeless constraint, and time must be recovered as an emergent, derivative concept [25,26,27]. The equation is formally of Klein–Gordon type on MSS, with the DeWitt metric serving as the target-space metric and $U_{\mathrm{eff}}$ playing the role of a potential that changes sign at the classical turning point $a_0$, defined by $U_{\mathrm{eff}}(a_0,\varphi )=0$, i.e. $a_0=V{\left(\varphi \right)}^{-1/2}$. The region $a<a_0$ ($U_{\mathrm{eff}}>0$) is classically forbidden; the region $a>a_0$ ($U_{\mathrm{eff}}<0$) is classically allowed. Tunneling through the forbidden region is the quantum analogue of the classical Big Bang.

The WKB ansatz $\Psi =A(a,\varphi )exp(i{S}_{\mathrm{cl}}(a,\varphi )/\hslash )$, inserted into (3) and expanded in ℏ, yields at leading order the Hamilton–Jacobi equation ${\left({\partial }_a{S}_{\mathrm{cl}}\right)}^2-U_{\mathrm{eff}}=0$ in the classically allowed region. The WKB time parameter is then defined operationally by [26,27,28]:

$${\displaystyle \frac{\partial }{\partial t}}\equiv \left({\partial }_a{S}_{\mathrm{cl}}\right){\displaystyle \frac{\partial }{\partial a}}+\left({\partial }_{\varphi }{S}_{\mathrm{cl}}\right){\displaystyle \frac{\partial }{\partial \varphi }},$$

(5)

so that the time-dependent Schrödinger equation for matter perturbations is recovered in the semiclassical limit. Time, in this sense, is not put in by hand—it crystallizes from the timeless wave function whenever the WKB phase oscillates rapidly enough for the approximation to hold. The turning point $a_0$ is, precisely, the birthplace of time.

The two boundary conditions differ in how the wave function behaves in the forbidden region $a<a_0$. For a de Sitter minisuperspace model with $V\left(\varphi \right)=\Lambda /3$, so $a_0={(3/\Lambda )}^{1/2}$ and $S_E\left(a\right)={\int }_0^{a}d{a}^{\prime }\sqrt{U_{\mathrm{eff}}\left(a^{\prime }\right)}$, the Hartle–Hawking no-boundary wave function [6,22], constructed via the Euclidean path integral

$${\Psi }_{HH}[h_{ij},{\varphi }_b]={\int }_{\mathrm{compact}}\mathcal{D}\left[g_{\mu \nu }\right]\mathcal{D}\left[\varphi \right]e^{-S_E[g,\varphi ]/\hslash },$$

(6)

summed over compact Euclidean four-geometries without initial boundary, selects the growing Euclidean mode:

$${\Psi }_{HH}\left(a\right)\approx \left\{\begin{array}{cc}\hfill {\mathcal{N}}_{HH}exp\left(+{\displaystyle \frac{\pi }{2\Lambda G\hslash }}\right)cosh\left({\displaystyle \frac{S_E\left(a\right)}{\hslash }}\right)\hfill & \hfill a<a_0,\hfill \\ \hfill {\mathcal{N}}_{HH}exp\left(+{\displaystyle \frac{\pi }{2\Lambda G\hslash }}\right)cos\left({\displaystyle \frac{S_{\mathrm{cl}}\left(a\right)}{\hslash }}-{\displaystyle \frac{\pi }{4}}\right)\hfill & \hfill a>a_0.\hfill \end{array}\right.$$

(7)

Vilenkin’s tunneling wave function [7,8], imposing purely outgoing modes at singular boundaries of superspace, selects the decaying Euclidean mode:

$${\Psi }_V\left(a\right)\approx \left\{\begin{array}{cc}\hfill {\mathcal{N}}_{V}exp\left(-{\displaystyle \frac{\pi }{2\Lambda G\hslash }}\right)exp\left(-{\displaystyle \frac{S_E\left(a\right)}{\hslash }}\right)\hfill & \hfill a<a_0,\hfill \\ \hfill {\mathcal{N}}_{V}exp\left(-{\displaystyle \frac{\pi }{2\Lambda G\hslash }}\right)exp\left(+{\displaystyle \frac{i{S}_{\mathrm{cl}}\left(a\right)}{\hslash }}\right)\hfill & \hfill a>a_0.\hfill \end{array}\right.$$

(8)

The sign reversal in the exponent of the overall prefactor reflects the fundamental preference of each proposal: Hartle–Hawking assigns larger weight to geometries with larger $\Lambda$ (the prefactor diverges as $\Lambda \to 0$); Vilenkin assigns larger weight to smaller $\Lambda$.

Concretely, for GUT-scale inflation with inflaton potential $V^{1/4}\sim {10}^{16}$ GeV—giving $V\approx 4.5\times {10}^{-13}$ in reduced Planck units where $M_{\mathrm{P}}={\left(8\pi G\right)}^{-1/2}=2.44\times {10}^{18}$ GeV—the Vilenkin tunneling probability is:

$$P_V\sim exp\left(-{\displaystyle \frac{24{\pi }^2}{V/M_{\mathrm{P}}^4}}\right)\approx exp\left(-{\displaystyle \frac{24{\pi }^2}{4.5\times {10}^{-13}}}\right)\approx exp(-5.3\times {10}^{14}),$$

(9)

an extraordinarily small but strictly positive number, confirming that tunneling is not kinematically forbidden—only exponentially suppressed. Table 1 provides a structured comparison of the two proposals.

3. Randall–Sundrum Brane Cosmology and the Modified WdW Equation

The RS bulk action is [9,10]:

$$S_{RS}=M_5^3\int d^{5}x\sqrt{G}\left({\mathcal{R}}_5-2{\Lambda }_5\right)+{\int }_{\partial \mathcal{M}}{d}^{4}x\sqrt{g}\left(-\lambda +{\mathcal{L}}_{\mathrm{matter}}\right),$$

(10)

where $G_{AB}$ is the five-dimensional metric with Ricci scalar ${\mathcal{R}}_5$, ${\Lambda }_5<0$ is the bulk cosmological constant, and $\lambda >0$ is the brane tension. The RS1 static solution is the warped metric:

$$d{s}_5^2=e^{-2k\left|y\right|}{\eta }_{\mu \nu }d{x}^{\mu }d{x}^{\nu }+d{y}^2,$$

(11)

with $k=\sqrt{-{\Lambda }_5/\left(6M_5^3\right)}$ and the RS fine-tuning condition ${\Lambda }_5=-{\lambda }^2/\left(6M_5^3\right)$, which enforces a flat brane (${\Lambda }_4=0$). The four-dimensional Planck mass is:

$$M_{\mathrm{P}}^2={\displaystyle \frac{M_5^3}{k}}\left(1-e^{-2k\pi r_c}\right)\approx {\displaystyle \frac{M_5^3}{k}}(k{r}_c\gg 1),$$

(12)

and the mass hierarchy generated by the warp factor is:

$${\displaystyle \frac{M_{\mathrm{weak}}}{M_{\mathrm{P}}}}=e^{-k\pi r_c}.$$

(13)

Setting $M_{\mathrm{weak}}/M_{\mathrm{P}}={10}^{-15}$ (the observed ratio between the electroweak and Planck scales), one finds:

$$k\pi r_c=15ln10\approx 34.54\Rightarrow k{r}_c\approx 10.99\simeq 11,$$

(14)

so that with $k\sim 0.1M_5\sim 2.4\times {10}^{17}$ GeV, the interbrane separation is $r_c\approx 11/(\pi \times 2.4\times {10}^{17}\mathrm{GeV})\approx 1.46\times {10}^{-17}{\mathrm{GeV}}^{-1}\approx 2.9\times {10}^{-32}$ m—two orders of magnitude above the Planck length, yet geometrically sufficient to compress the full ${10}^{15}$ hierarchy into a single exponential.

Figure 1 illustrates the five-dimensional geometry of both RS models.

Derivation of the RS-Modified Wheeler–DeWitt Equation

When the brane undergoes cosmological expansion, the Israel junction conditions applied to the RS bulk yield—after using the Gauss–Codazzi relations—the modified Friedmann equation [12,14]:

$$H^2\equiv {\left({\displaystyle \frac{\dot{a}}{a}}\right)}^2={\displaystyle \frac{{\kappa }_4^2}{3}}\rho \left(1+{\displaystyle \frac{\rho }{2\lambda }}\right)-{\displaystyle \frac{k}{a^2}}+{\displaystyle \frac{\mu }{a^4}}+{\displaystyle \frac{{\Lambda }_4}{3}},$$

(15)

where ${\kappa }_4^2=8\pi G_N={\kappa }_5^4\lambda /6$ with ${\kappa }_5^2\equiv M_5^{-3}$, $\mu$ is the dark-radiation constant (a free integration constant of the five-dimensional equations), and ${\Lambda }_4={\kappa }_5^4{\lambda }^2/12+{\kappa }_5^2{\Lambda }_5/2$ vanishes under the RS fine-tuning. The quadratic correction ${\rho }^2/\left(2\lambda \right)$ dominates when $\rho \gtrsim \lambda$, i.e. at energies $E\gtrsim {\lambda }^{1/4}$.

To derive the modified WdW equation, one identifies the canonical momentum ${\pi }_a=-6M_{\mathrm{P}}^{2}a\dot{a}/N$ from the brane minisuperspace action, performs the Legendre transform, and imposes the Hamiltonian constraint. The result is:

$$\begin{array}{|c|}\hline \left[-{\displaystyle \frac{{\partial }^2}{\partial a^2}}+U_{RS}(a,\varphi )\right]{\Psi }_{RS}[a,\varphi ]=0,\\ \hline\end{array}$$

(16)

with the RS-modified effective potential (units ${\kappa }_4=1$, $k=+1$):

$$\begin{array}{cc}\hfill U_{RS}(a,\varphi )& \hfill =\underset{U_{GR}\left(a\right)}{\underbrace{a^2-{\displaystyle \frac{{\Lambda }_{\mathrm{eff}}}{3}}{a}^4}}-\underset{\mathrm{RS}\mathrm{brane}-\mathrm{energy}\mathrm{correction}}{\underbrace{{\displaystyle \frac{{\kappa }_4^4\rho {\left(\varphi \right)}^2}{6\lambda }}{a}^6}}+\underset{\mathrm{dark}\mathrm{radiation}}{\underbrace{{\displaystyle \frac{\mu {\kappa }_4^2}{a^2}}}},\hfill \end{array}$$

(17)

where ${\Lambda }_{\mathrm{eff}}={\kappa }_4^2\rho /3$ for constant energy density. Four structural differences between $U_{RS}$ and $U_{GR}$ deserve articulation. The $a^6$ correction, proportional to ${\rho }^2/\lambda$, is always negative and pulls the potential downward at intermediate a, narrowing the classical turning point compared to the four-dimensional case. The dark-radiation term $\mu /a^2$ diverges at $a\to 0$ with a sign that is a free parameter, raising the potential barrier (for $\mu >0$) or reducing it (for $\mu <0$). The coefficient ${\kappa }_4^2={\kappa }_5^4\lambda /6$ connects the four-dimensional gravitational coupling to the brane tension—all RS corrections scale with brane-tension combinations. The fine-tuning condition (10) is broken dynamically during brane expansion, generating a density-dependent effective cosmological constant absent from the static RS solution.

The classical turning point $a_0^{RS}$ defined by $U_{RS}(a_0^{RS},\varphi )=0$ satisfies, for $\mu =0$ and constant $\rho$:

$$a_0^{RS}\approx a_0^{GR}\left(1-{\displaystyle \frac{1}{4\lambda {\kappa }_4^2{\left(a_0^{GR}\right)}^2}}\right)+\mathcal{O}\left({\lambda }^{-2}\right),$$

(18)

where $a_0^{GR}={(3/{\Lambda }_{\mathrm{eff}})}^{1/2}$. The correction is always negative (the RS turning point is smaller), confirming that the RS geometry narrows the quantum barrier. Figure 2 displays $U_{GR}$, $U_{RS}$ (with $\mu >0$ and $\mu <0$), and the LQC-corrected potential for direct comparison.

RS Correction to the Tunneling Exponent

The tunneling amplitude through $U_{RS}\left(a\right)$ is:

$$|T_{RS}{|}^2\sim exp\left(-2{\int }_0^{a_0^{RS}}da\sqrt{U_{RS}\left(a\right)}\right).$$

(19)

Expanding around the GR baseline, the RS correction to the log-amplitude is:

$$\begin{array}{cc}\hfill \delta ln|T_{RS}{|}^2& \hfill =-2{\int }_0^{a_0^{RS}}da\left(\sqrt{U_{RS}}-\sqrt{U_{GR}}\right)\hfill \\ \hfill & \hfill \approx +{\int }_0^{a_0^{GR}}da{\displaystyle \frac{{\kappa }_4^4{\rho }^2{a}^6/\left(6\lambda \right)-\mu {\kappa }_4^2/a^2}{2\sqrt{U_{GR}\left(a\right)}}}+\mathcal{O}\left({\lambda }^{-2}\right).\hfill \end{array}$$

(20)

For $\mu =0$ this is positive—the RS correction increases the tunneling probability relative to four-dimensional GR, an enhancement that grows with the energy density at the bounce epoch and shrinks as $\lambda \to \infty$ (recovering GR). For $\mu <0$ a competing suppression may dominate at small a.

4. Exact Correspondences Between Boundary Conditions and RS Dynamics

The mapping between four-dimensional quantum cosmological boundary conditions and five-dimensional RS dynamics operates at three structurally distinct levels. We develop each in sequence.

The Euclidean path-integral level is the most fundamental. The Hartle–Hawking integral (6) generalizes to five dimensions as:

$${\Psi }_{HH}^{RS}[h_{ij},{\varphi }_b]={\int }_{\mathrm{compact}}\mathcal{D}\left[G_{AB}\right]\mathcal{D}\left[\varphi \right]e^{-S_E^{\left(5\right)}[G,\varphi ]/\hslash },$$

(21)

summed over compact Euclidean five-geometries with a closed brane and no initial boundary in the bulk. The Euclidean continuation of the Lorentzian RS metric ($t\to -i\tau$) yields a Euclidean AdS₅ bulk in which the brane traces a closed $S^4$ hypersurface; the no-boundary condition demands that this Euclidean bulk be smooth and compact, with the brane shrinking to a point at the Euclidean “south pole.” The correspondence is:

$$\underset{\mathrm{HH}\mathrm{saddle},4\mathrm{D}}{\underbrace{\mathrm{compact}{S}^4}}⟷\underset{\mathrm{HH}\mathrm{saddle},\mathrm{RS}/5\mathrm{D}}{\underbrace{\mathrm{compact}\mathrm{Eucl}.{\mathrm{AdS}}_5+\mathrm{closed}\mathrm{brane}}}.$$

(22)

The Euclidean action of the five-dimensional RS saddle is:

$$S_E^{RS}=-{\displaystyle \frac{24{\pi }^2{M}_5^3}{H^4}}·f\left({\displaystyle \frac{\lambda }{M_5^{3}H}}\right),$$

(23)

where H is the de Sitter Hubble rate on the brane and $f\left(x\right)\to 1$ as $x\to 0$ (recovering the four-dimensional result $S_E=-24{\pi }^2{M}_{\mathrm{P}}^2/H^2$). Vilenkin’s tunneling condition, applied to the RS path integral, maps to brane nucleation in the AdS₅ bulk—formally identical to vacuum bubble nucleation [30]—with probability $P_V^{RS}\sim exp(-2|S_E^{RS}|/\hslash )$.

At the WKB level, the tunneling exponent under RS corrections follows directly from (20): both proposals preserve their respective signs (positive for HH, negative for Vilenkin) but acquire RS corrections that shift the numerical value of the exponent in a brane-tension-dependent direction, with the amplitude of the shift controlled by the dimensionless ratio $\rho /\lambda$ at the turning point.

At the AdS/CFT level [11], the RS2 brane at position $y_0$ in AdS₅ is holographically dual to a four-dimensional conformal field theory with UV cutoff ${\Lambda }_{UV}\sim M_5{e}^{-k{y}_0}$ [31,32]. The brane wave function maps to the CFT partition function:

$${\Psi }_{brane}\left[h_{ij}\right]=Z_{CFT}[h_{ij};{\Lambda }_{UV}].$$

(24)

The no-boundary Euclidean path integral over bulk geometries is dual to the CFT path integral over field configurations on the boundary; the Vilenkin boundary condition corresponds to a specific RG-flow state of the CFT defined by purely outgoing modes in the flow. Table 2 records the complete three-level mapping.

5. The Problem of the Beginning of Time and Its Resolution

Time does not enter the WdW equation. This is no gauge ambiguity or calculational convenience—it is an irresistible consequence of diffeomorphism invariance, which promotes the Hamiltonian of general relativity to a constraint $\mathcal{H}=0$ rather than a generator of evolution in some pre-existing background time [4,28]. The WdW equation is timeless in the most literal sense: no derivative $\partial /\partial t$ appears, and the wave function $\Psi [h_{ij},\varphi ]$ does not carry an explicit time label. Time must be recovered as a derivative concept—a function of the dynamical variables themselves.

The WKB mechanism achieves this recovery [25,26,27]. In the classically allowed region ($a>a_0$), the wave function oscillates rapidly: $\Psi \approx A{e}^{i{S}_{\mathrm{cl}}/\hslash }$. The phase $S_{\mathrm{cl}}(a,\varphi )$ satisfies the classical Hamilton–Jacobi equation. Equation (5) defines a vector field on minisuperspace whose integral curves are the classical Friedmann solutions; the parameter along these curves is cosmic time t. Time, in this formulation, is not fundamental—it is the WKB label that tracks which classical trajectory the wave function is currently “near.” Before the crystallization ($a<a_0$, the Euclidean region), no such classical trajectory exists, the phase is real, and no time parameter is defined. The turning point $a_0$ is, structurally, the moment at which time itself first becomes meaningful.

Three mechanisms for time emergence, one per framework, follow from this analysis. In the Hartle–Hawking case, the Euclidean-to-Lorentzian transition at $a=a_0$ is geometrically smooth: the analytic continuation $\tau \to it$ converts the round $S^4$ (Euclidean) into a de Sitter hyperboloid (Lorentzian) without any singular intermediate state. Time “appears” by analytic continuation from imaginary time; the boundary between the two régimes is the equator of the Euclidean sphere. In the Vilenkin case, the transition is discontinuous in classical terms—the universe tunnels from $a=0$ to $a=a_0$ with no Lorentzian trajectory connecting them—and time appears sharply after the nucleation, on the classical expanding side. In the RS case, a third mechanism is available: the position y of the brane in the fifth dimension serves as a York time variable [42,43]. As the brane moves through the AdS₅ bulk, its induced metric evolves cosmologically, and the integral ${\tau }_{brane}=\int dy\sqrt{G_{55}\left(y\right)}$ constitutes the four-dimensional cosmic time. Time on the brane is motion in the extra dimension—a concrete geometric realization of temporal emergence from a higher-dimensional, essentially static, bulk geometry.

Figure 3 illustrates the WKB transition and the birth of time for both boundary conditions under the standard and RS-modified potentials.

6. Wave-Function “Collapse” and the Big Bounce

Collapse requires an observer. In single-particle quantum mechanics the Copenhagen measurement postulate assigns a definite post-measurement state to a system after an external agent interacts with it. No such agent exists for the universe: by definition, no observer occupies a position exterior to the total system being described by $\Psi [h_{ij},\varphi ]$. The universal wave function does not collapse—it evolves, unitarily, according to the WdW equation. What appears to be collapse, from the vantage of an internal observer, is decoherence [19,40,41]: entanglement between large-scale geometry (the “system”) and short-wavelength radiation and matter perturbations (the “environment”) suppresses off-diagonal elements of the reduced density matrix:

$${\rho }_{\mathrm{red}}[h,h^{\prime }]=\int \mathcal{D}\left[{\varphi }_e\right]\Psi [h,{\varphi }_e]{\Psi }^{\ast }[h^{\prime },{\varphi }_e]\stackrel{\mathrm{decoherence}}{\to }{\left|\mathcal{A}\left[h\right]\right|}^2\delta [h-h_{\mathrm{cl}}],$$

(25)

selecting one classical geometry $h_{\mathrm{cl}}$ without any genuine global collapse. The universe does not “choose” a branch; it entangles with its internal environment so thoroughly that the interference between branches becomes unmeasurable.

Three candidate mechanisms for the cosmological bounce are physically distinct and deserve separate treatment. The decoherence mechanism [19,39] produces an effective classical geometry from the WdW wave function without any singular bounce event: the WKB approximation breaks down at small a and is re-established at large a, with the quantum-gravity epoch playing the role of a “scrambler” that re-initializes the classical description without physically bouncing. The Penrose objective-reduction mechanism [20] posits that gravitational self-energy drives a genuine non-unitary state reduction when the mass-energy difference between superposed geometries satisfies:

$${\tau }_{OR}\sim {\displaystyle \frac{\hslash }{\Delta E_{\mathrm{grav}}}}\sim {\displaystyle \frac{\hslash M_{\mathrm{Pl}}^2}{G\Delta \rho {\ell }^3}}\sim \ell \left({\displaystyle \frac{M_{\mathrm{P}}^3}{\Delta \rho {\ell }^3}}\right);$$

(26)

at Planckian density $\Delta \rho \sim M_{\mathrm{P}}^4$ and Planckian scale $\ell \sim {\ell }_{\mathrm{P}}$, one finds ${\tau }_{OR}\sim \ell \sim {10}^{-43}$ s, so objective reduction would occur instantaneously by any macroscopic clock. In the RS framework, the brane-collision mechanism of the ekpyrotic/cyclic scenario [17,18] is the most physically transparent: two branes approach each other in AdS₅, collide—generating a hot plasma that reheats the universe—and bounce apart. The wave function of the two-brane system $\Psi [a,y_{12}]$ (where $y_{12}$ is the inter-brane separation) propagates through $y_{12}\to 0$ unitarily and re-emerges on the far side. The brane collision is literally a bounce in the five-dimensional geometry, with the bounce operator mapping:

$${\Psi }_{\mathrm{after}}[a,\varphi ]={\widehat{U}}_{\mathrm{bounce}}{\Psi }_{\mathrm{before}}[a,\varphi ],{\widehat{U}}_{\mathrm{bounce}}^{†}{\widehat{U}}_{\mathrm{bounce}}=\mathbb{1}.$$

(27)

Figure 4 summarizes the information channels through the bounce.

7. Entropy Argument, Phase Memory, and the Unitary Bridge

Thermal entropy is bounded, and that bound requires a substrate. The Bekenstein bound [24] $S\le 2\pi RE/\hslash c$ applies to any system of energy E enclosed in a sphere of radius R; the Bekenstein–Hawking black-hole entropy $S_{BH}=A/\left(4{\ell }_{\mathrm{P}}^2\right)$ applies to any region bounded by a horizon of area A. Both expressions are properties of fields defined on a spacetime manifold: they count independent field configurations compatible with specified geometric constraints. When the spacetime manifold itself ceases to exist—as at a classical Big Crunch—the manifold on which entropy is defined disappears. Thermal entropy $S_{\mathrm{thermal}}$, Weyl curvature entropy $S_{\mathrm{Weyl}}$ in Penrose’s sense [21], and every other functional of classical field configurations are therefore categorically erased at the bounce. They require a geometric substrate; the bounce removes the substrate itself.

This erasure mechanism differs in kind—not merely in degree—from information loss into a black hole. A gravitationally collapsing body creates a horizon within a pre-existing, spatially infinite spacetime; an external observer inhabits an intact region and can, in principle, recover information via Hawking radiation [23] or holographic encoding on the horizon surface. The horizon is a membrane within spacetime. At a cosmological bounce, no exterior region exists: the compression is of the entire spacetime, fields, and geometry simultaneously, so no external observer position and no undisturbed substrate remain. Stated in terms of the metric:

$$\underset{\mathrm{black}\mathrm{hole}\text{:}\mathrm{substrate}\mathrm{intact}}{\underbrace{g_{\mu \nu }^{\mathrm{exterior}}\ne 0}}\mathrm{vs}.\underset{\mathrm{bounce}\text{:}\mathrm{no}\mathrm{substrate}.}{\underbrace{g_{\mu \nu }\to 0\mathrm{everywhere}}}$$

(28)

The CMB, primordial gravitational waves, and every other classical observable of our universe are post-bounce phenomena. They are defined on the post-bounce spacetime and carry, by the above argument, zero imprint of pre-bounce classical configurations. The argument is categorical, not probabilistic.

Phase Entropy and Phase Fidelity

The wave function $\Psi [h_{ij},\varphi ]=\left|\Psi \right|e^{i\theta [h_{ij},\varphi ]}$ is a complex functional on superspace. Its modulus squared ${\left|\Psi \right|}^2$ defines a probability measure; its phase $\theta [h_{ij},\varphi ]=arg\left(\Psi \right)$ encodes the semiclassical flow—in the WKB limit, $\theta =S_{\mathrm{cl}}/\hslash$ and the gradient ${\nabla }_h\theta$ is proportional to the canonical momentum ${\pi }_{ij}$, recording the entire classical trajectory from which $\Psi$ was derived. All information about which classical history the universe has followed is stored in the phase, not in the modulus.

We define the phase entropy as a functional on the distribution of phases over superspace:

$$\begin{array}{ccc}\hfill S_{\theta }\left[\Psi \right]& \hfill \equiv \hfill & -{\int \mathcal{D}\left[h\right]\mathcal{D}\left[\varphi \right]\left|\Psi [h,\varphi ]\right|}^{2}ln{\displaystyle \frac{{\left|\Psi [h,\varphi ]\right|}^2}{\mathcal{Z}}}\hfill \\ \hfill & \hfill \equiv \hfill & -{\displaystyle \frac{1}{2\pi }}{\int }_0^{2\pi }d\theta \int \mathcal{D}\left[h\right]|{\nabla }_h{\Psi }_{\theta }\left[h\right]{|}^2,\hfill \end{array}$$

(29)

where $\mathcal{Z}$ is the normalization of ${\left|\Psi \right|}^2$ over superspace and ${\Psi }_{\theta }=\Psi e^{i\theta }$ denotes a phase-shifted copy. The first term is the Shannon entropy of the modulus-squared distribution (amplitude entropy); the second term captures the coherence of the phase gradient across superspace, and is minimized when the phase is deterministic (pure WKB) and maximized when the phase is uniformly distributed. We define the phase fidelity:

$${\mathcal{F}}_{\mathrm{bounce}}\equiv {\left|\int \mathcal{D}\left[h\right]\mathcal{D}\left[\varphi \right]{\Psi }_{\mathrm{after}}^{\ast }[h,\varphi ]·{\widehat{U}}_{\mathrm{bounce}}{\Psi }_{\mathrm{before}}[h,\varphi ]\right|}^2.$$

(30)

For a unitary bounce, ${\mathcal{F}}_{\mathrm{bounce}}=1$; for a classically chaotic bounce (BKL Mixmaster oscillations [33,34]) the bounce operator acquires a random unitary component and ${\mathcal{F}}_{\mathrm{bounce}}<1$.

Proposition 1

(Phase entropy is bounce-invariant). Let $\widehat{U}$ be any unitary operator on the Hilbert space of the universal wave function, i.e. ${\widehat{U}}^{†}\widehat{U}=\mathbb{1}$. Then $S_{\theta }\left[\widehat{U}\Psi \right]=S_{\theta }\left[\Psi \right]$.

Proof. 

Unitarity implies $|\widehat{U}{\Psi \left[h\right]|}^2={\left|\Psi \left[{\widehat{U}}^{†}h\right]\right|}^2$ up to a Jacobian factor that equals unity for $\widehat{U}$ induced by a field redefinition on superspace. Consequently, the amplitude entropy $-{\int \left|\Psi \right|}^{2}ln{\left|\Psi \right|}^2$ is invariant under $\widehat{U}$, since the measure $\mathcal{D}\left[h\right]$ transforms covariantly. For the phase-coherence term: $|{\nabla }_h\left(\widehat{U}\Psi \right){|}^2=|\widehat{U}{\nabla }_h{\Psi |}^2={\left|{\nabla }_h\Psi \right|}^2$, since $\widehat{U}$ acts as an isometry on the Hilbert space. Both terms of (29) are therefore invariant, and $S_{\theta }\left[\widehat{U}\Psi \right]=S_{\theta }\left[\Psi \right]$. □

Corollary 1.

Under unitary quantum gravity, $S_{\theta }$ is a conserved quantity throughout any cosmological evolution, including the Big Bounce. It is the natural quantum-cosmological analogue of the classical Liouville invariant.

Table 3 contrasts all relevant entropy measures across the bounce.

8. Observational Consequences

Phase memory is, by definition, not directly observable—$\Psi$ is not a classical field whose value can be measured at a detector. Its imprints must manifest as statistical properties of quantum perturbations that deviate from what a universe without pre-bounce history would predict. Three channels are physically motivated.

The phase of the post-bounce wave function, via the WKB relation, encodes specific correlations between long-wavelength modes set before the bounce and modes generated during inflation. In standard slow-roll inflation, the primordial bispectrum $B(k_1,k_2,k_3)=\langle \delta {\varphi }_{k_1}\delta {\varphi }_{k_2}\delta {\varphi }_{k_3}\rangle$ is suppressed below cosmic-variance sensitivity for single-field models in the Bunch–Davies vacuum [35]. If pre-bounce phase imprints a coherent pattern of inter-mode entanglement, the bispectrum acquires an oscillatory correction:

$$\Delta B(k_1,k_2,k_3)\sim {\mathcal{F}}_{\mathrm{bounce}}·B_0(k_1,k_2,k_3)·sin\left(\sum _i{k}_i{\eta }_{\mathrm{bounce}}\right),$$

(31)

where ${\eta }_{\mathrm{bounce}}$ is the conformal time at the bounce and the sine factor oscillates with period $\Delta k\sim {\eta }_{\mathrm{bounce}}^{-1}$, set by the pre-bounce Hubble radius. The amplitude is suppressed by ${\mathcal{F}}_{\mathrm{bounce}}$, which may itself be exponentially small—but its phase is a direct imprint of $S_{\theta }$ carried across the bounce.

In the RS ekpyrotic setting, the brane collision generates tensor perturbations with a power spectrum modified relative to the inflationary prediction [36]. The RS correction to the scalar primordial spectrum is [13,14]:

$${\mathcal{P}}_S^{RS}\left(k\right)={\mathcal{P}}_S^{GR}\left(k\right){\left(1+{\displaystyle \frac{{\rho }_{\ast }}{2\lambda }}\right)}^{-1}\left(1+{\displaystyle \frac{2{\rho }_{\ast }}{\lambda }}\right),$$

(32)

where ${\rho }_{\ast }$ is the energy density at Hubble exit for mode k. For high-scale RS

$$\lambda \sim M_5^4,$$

the correction is sub-percent on CMB scales; for low-scale RS ($\lambda \sim {\left(10\mathrm{TeV}\right)}^4$) it can reach order unity on large angular scales ($\ell \lesssim 10$), potentially explaining the observed CMB quadrupole anomaly. where $\lambda$ is a coupling parameter of mass dimension four, characterized by the scaling relation

$$\lambda \sim M_5^4,$$

where $M_5$ denotes the fundamental five-dimensional mass scale. The initial quantum state of inflation—whether Bunch–Davies, a Bogoliubov-transformed vacuum, or something more structured—carries an imprint of the pre-inflationary (and, under our framework, pre-bounce) wave function. A non-Bunch–Davies initial state modifies the scalar power spectrum with oscillatory features [37,38]:

$${\mathcal{P}}_S\left(k\right)={\mathcal{P}}_S^{BD}\left(k\right)\left[1+{ϵ}_{\beta }cos\left(2{\displaystyle \frac{k}{k_{\ast }}}+{\varphi }_0\right)\right],$$

(33)

where ${ϵ}_{\beta }$ is the Bogoliubov coefficient amplitude (proportional to ${\mathcal{F}}_{\mathrm{bounce}}$), $k_{\ast }\sim {\eta }_{\mathrm{bounce}}^{-1}$ is the reference scale set by the pre-bounce Hubble radius, and ${\varphi }_0$ is the phase inherited from $\theta \left[{\Psi }_{\mathrm{before}}\right]$. Detection of such oscillatory features with a specific phase relationship across CMB multipoles—and consistency between the phase measured in the temperature and E-mode polarization spectra—would constitute indirect but physically motivated evidence for quantum phase memory across the bounce. Future experiments (CMB-S4, LiteBIRD) are sensitive to $|{ϵ}_{\beta }|\gtrsim {10}^{-2}$ at $2\sigma$ confidence, and primordial gravitational-wave detectors (LISA, DECIGO, BBO) could constrain the tensor-to-scalar ratio in the RS ekpyrotic scenario [36].

9. Discussion and Conclusions

This paper establishes, to our knowledge for the first time, a systematic three-level correspondence between the Hartle–Hawking no-boundary proposal and the Vilenkin tunneling proposal on the one side, and the dynamical structure of Randall–Sundrum brane cosmology on the other, mediated by the modified Wheeler–DeWitt equation derived from the RS bulk action. At the Euclidean path-integral level, the no-boundary saddle maps to a compact Euclidean AdS₅ geometry with a closed brane; Vilenkin’s outgoing-mode condition maps to brane nucleation from the bulk, formally identical to vacuum bubble nucleation in five dimensions. At the WKB level, the RS correction to the tunneling exponent is calculable from (20) and is positive for $\mu =0$, predicting an enhanced tunneling probability relative to four-dimensional GR by a factor that depends on the ratio $\rho /\lambda$ at the turning point. At the holographic level, the AdS/CFT correspondence recasts the cosmological wave function as a CFT partition function with a UV cutoff, providing a non-perturbative definition of both boundary conditions in terms of CFT data.

The entropy argument for classical memory erasure at the bounce is, we believe, more decisive than is commonly appreciated in the literature on cyclic cosmology. The distinction between compression within spacetime (black hole) and compression of spacetime (bounce) is categorical: in the former case an intact exterior spacetime provides a substrate for information retrieval via Hawking radiation [23]; in the latter case no such substrate exists, and every classical observable—CMB, gravitational waves, thermal relics, Weyl entropy—is categorically erased. This does not imply that a cyclic universe has no memory; it implies that the memory must be encoded in the phase structure of the quantum wave function, not in classical fields. The phase entropy $S_{\theta }$ defined here is invariant under any unitary bounce operator (Proposition 1) and therefore constitutes a genuine conserved quantity of quantum cosmology—the natural quantum analogue of the classical Liouville invariant—that threads unbroken through an unlimited number of bounce cycles.

Three limitations of the present framework deserve explicit acknowledgement. The mini-superspace approximation truncates the infinite-dimensional superspace to two degrees of freedom (a and $\varphi$), discarding all inhomogeneous perturbation modes; including these would generate an infinite tower of harmonic-oscillator corrections to the WdW equation and could modify both $S_{\theta }$ and ${\mathcal{F}}_{\mathrm{bounce}}$. The unitarity of ${\widehat{U}}_{\mathrm{bounce}}$ is assumed—it holds in loop quantum cosmology [15,16], which provides a specific lattice regularization of the WdW equation, but is not a derived result in full quantum gravity. The AdS/CFT dictionary, while precise in the string-theoretic RS2 context [11], has not been rigorously established for the cosmological brane at finite y-position, where the standard GKPW prescription [31] must be modified. These are avenues for future development, not flaws in the structural correspondences established here.

The deepest open question is the value of ${\mathcal{F}}_{\mathrm{bounce}}$ in realistic quantum gravity. If quantum gravity is fully unitary across the bounce, ${\mathcal{F}}_{\mathrm{bounce}}=1$ and successive cycles carry perfect quantum memory. If some information is genuinely lost—as in the pre-Page-time black-hole scenario—${\mathcal{F}}_{\mathrm{bounce}}<1$ and cycles become progressively more quantum-stochastic, with the wave function approaching a Haar-random state after sufficiently many bounces. The observational handle on this question lies in the amplitude ${ϵ}_{\beta }$ of oscillatory features in the primordial power spectrum (equation (33)) and in the phase consistency between scalar and tensor perturbations—both of which are within reach of forthcoming CMB-S4 and space-borne gravitational-wave observatories, even if the signal amplitude is suppressed by the exponentially small ratio ${\mathcal{F}}_{\mathrm{bounce}}/P_V$ that sets the overall scale of pre-bounce imprints.