Skip to content

Minisuperspace Quantization

Scrollable formula. Use Left and Right arrow keys to reveal hidden content.

Scrollable table. Use Left and Right arrow keys to reveal additional columns.

Scrollable figure. Use Left and Right arrow keys to inspect hidden content.

Minisuperspace quantization retains finitely many homogeneous variables before quantization. It turns the gravitational constraint into a tractable wave equation, but the truncation removes local gravitational waves, density perturbations, and mode couplings at the classical level. Its results are therefore exact for the reduced model, not automatically controlled approximations to the full quantum theory.

Helpful background. Flat-Spacetime Quantum Field Theory supplies canonical quantization; Einstein Equations supplies Friedmann dynamics; Quantum Information supplies state and inner-product concepts; Quantum Gravity supplies truncation criteria; Causality and Global Structure supplies singular and boundary interpretation; and Quantum Fields in Curved Spacetime supplies field modes on evolving backgrounds.

Homogeneous action and the lapse constraint

Section titled “Homogeneous action and the lapse constraint”

Take a spatially flat homogeneous metric and scalar,

ds2=−N(t)2dt2+a(t)2dx2,ϕ=ϕ(t),ds^2=-N(t)^2dt^2+a(t)^2d\mathbf x^2, \qquad \phi=\phi(t),

and integrate over a fiducial comoving cell of coordinate volume V0V_0. After including the gravitational boundary term, the reduced Lagrangian is

L=−3V0aa˙28πGN+V0a3ϕ˙22N−NV0a3V(ϕ).L=-\frac{3V_0a\dot a^2}{8\pi G N} +\frac{V_0a^3\dot\phi^2}{2N} -NV_0a^3V(\phi).

The momenta are

pa=−3V0aa˙4πGN,pϕ=V0a3ϕ˙N,pN=0.p_a=-\frac{3V_0a\dot a}{4\pi G N}, \qquad p_\phi=\frac{V_0a^3\dot\phi}{N}, \qquad p_N=0.

The vanishing lapse momentum is a primary constraint. The Legendre transform gives H=NHH=N\mathcal H with

H=−2πG3V0apa2+12V0a3pϕ2+V0a3V(ϕ).\mathcal H= -\frac{2\pi G}{3V_0a}p_a^2 +\frac{1}{2V_0a^3}p_\phi^2 +V_0a^3V(\phi).

Varying NN imposes H=0\mathcal H=0. Substituting the momenta recovers the flat Friedmann equation, so the sign and normalization have an immediate classical check.

Canonical quantization sends pa↦−i∂ap_a\mapsto-i\partial_a and pϕ↦−i∂ϕp_\phi\mapsto-i\partial_\phi. Because aa and pap_a do not commute, a−1pa2a^{-1}p_a^2 has inequivalent orderings. For the explicitly declared measure ap daa^p\,da, a symmetric one-parameter family is

a−1pa2⟼−a−p∂a ⁣(ap−1∂a).a^{-1}p_a^2\longmapsto -a^{-p}\partial_a\!\left(a^{p-1}\partial_a\right).

Integration by parts shows symmetry because the quadratic form contains ap−1(∂aψ)∗(∂aχ)a^{p-1}(\partial_a\psi)^*(\partial_a\chi) when boundary flux vanishes. Different measures and orderings alter the first-derivative and effective inverse-square terms. The Wheeler–DeWitt equation H^Ψ=0\widehat{\mathcal H}\Psi=0 therefore requires an ordering, domain, and measure, not only the classical constraint.

For the free massless scalar, the normalization can be made explicit. Set α=ln⁡(a/a0)\alpha=\ln(a/a_0) for a fixed reference scale a0a_0, use pα=apap_\alpha=ap_a, and multiply the classical constraint by 2V0a32V_0a^3. This gives

pϕ2−4πG3pα2=0.p_\phi^2-\frac{4\pi G}{3}p_\alpha^2=0.

Define the dimensionless clock φ=4πG/3 ϕ\varphi=\sqrt{4\pi G/3}\,\phi. Choosing the flat ordering for this rescaled constraint then yields

(∂φ2−∂α2)Ψ(α,φ)=0.\left(\partial_\varphi^2-\partial_\alpha^2\right)\Psi(\alpha,\varphi)=0.

The solutions are F(α−φ)+G(α+φ)F(\alpha-\varphi)+G(\alpha+\varphi). This is a specified quantization choice: multiplying a classical constraint by a nonconstant function does not establish unitary equivalence of arbitrary quantum orderings. The scalar serves as an internal clock on sectors where it is monotonic. This is relational evolution; the fundamental equation still has no external time. Kiefer and Sandhöfer derive the reduced constraint and discuss ordering and the inner-product problem Kiefer and Sandhöfer 2022, § 3 and appendix, Eqs. (22)–(31).

For noncompact flat slices, V0V_0 is an infrared regulator. Predictions cannot depend on its arbitrary coordinate size. Under V0↦cV0V_0\mapsto cV_0, canonical momenta and the normalization of homogeneous variables rescale. Dimensionless relational quantities—such as density as a function of ϕ\phi—should remain invariant after states are transformed consistently.

Using physical volume v=V0a3v=V_0a^3 clarifies this check, but vv itself scales if the chosen cell is changed. One must distinguish changing the regulator cell from comparing physically different compact universes. A wave-function normalization tied to da dϕda\,d\phi can also differ from one tied to the minisuperspace DeWitt metric.

Setting all inhomogeneous modes to zero before quantization eliminates their vacuum energy, backreaction, decoherence, and interactions with the background. It also suppresses anisotropic instabilities unless anisotropy variables are retained. Quantizing the reduced theory is not generally the same as quantizing the full theory and projecting onto homogeneous states, because constraints and quantum fluctuations need not commute with truncation.

A systematic use of minisuperspace therefore states a regime: long wavelengths, near-homogeneous states, and a hierarchy keeping omitted modes perturbative. One may then add perturbations and test whether their stress energy remains small. Singular behavior of the reduced wave function, or its absence, does not alone prove singularity resolution; a physical inner product and curvature observables are required.

Insert Ψ=AeiS\Psi=Ae^{iS} into the Wheeler–DeWitt equation and order terms by derivatives of the slowly varying amplitude. The leading real term replaces momenta by phase gradients and must reproduce the Hamilton–Jacobi constraint H(a,ϕ;∂S)=0\mathcal H(a,\phi;\partial S)=0. This verifies the relative sign between gravitational and scalar kinetic terms. The next order is a transport equation for AA along the classical minisuperspace trajectories; failure of that transport equation signals an inconsistent ordering or measure.

For a Klein–Gordon form, two solutions admit the conserved current

JA=i2 Ψ∗∇A↔Ψ,∇AJA=0.J^A=\frac{i}{2}\,\Psi^* \overset{\leftrightarrow}{\nabla^{A}}\Psi, \qquad \nabla_AJ^A=0.

Current conservation checks the chosen superspace measure and boundary flux. It does not make JϕJ^\phi positive for arbitrary superpositions. Positivity requires selecting a frequency sector relative to the proposed internal clock, a selection that can fail when the potential mixes positive and negative frequency. Finally, the G→0G\to0 or slowly varying geometry limit should recover quantum matter on a prescribed background only after a WKB branch supplies time.

For V(ϕ)=0V(\phi)=0, use the classical constraint to show that ϕ\phi is monotonic on any solution with pϕ≠0p_\phi\ne0. Why does this make it a useful but non-global clock?

Solution

Hamilton’s equation gives

ϕ˙=N∂H∂pϕ=NpϕV0a3.\dot\phi=N\frac{\partial\mathcal H}{\partial p_\phi} =N\frac{p_\phi}{V_0a^3}.

The free scalar equation conserves pϕp_\phi. For positive lapse and a>0a\gt0, the sign of ϕ˙\dot\phi is therefore fixed when pϕ≠0p_\phi\ne0, so each classical trajectory can be parametrized by ϕ\phi. The clock fails on the pϕ=0p_\phi=0 sector and can fail after adding a potential that produces turning points. In the quantum theory, superpositions of clock momenta and the choice of positive-frequency sector introduce further restrictions.

  • Kiefer, Claus, and Barbara Sandhöfer (2022), “Quantum Cosmology,” Zeitschrift für Naturforschung A 77, 543–559, arXiv:0804.0672v3.