Wheeler–DeWitt equation

The Wheeler–DeWitt equation^[1] is an attempt to combine mathematically the ideas of quantum mechanics and general relativity, a step toward a theory of quantum gravity. In this approach, time plays no role in the equation, leading to the problem of time.^[2] More specifically, the equation describes the quantum version of the Hamiltonian constraint using metric variables. Its commutation relations with the diffeomorphism constraints generate the Bergmann-Komar "group" (which is the diffeomorphism group on-shell, but differs off-shell).

Because of its connections with the low-energy effective field theory, it inherits all the problems of the naively quantized GR, and thus it cannot be used at multi-loop level, etc., at least not according to the current knowledge.

The equation has not played a role in string theory thus far, since all properly defined and understood descriptions of string/M-theory deal with some fixed asymptotic conditions on the background. Thus, at infinity, the "right" choice of the time coordinate "t" is determined in every description, so there is a preferred definition of the Hamiltonian (with nonzero eigenvalues) to evolve states of the system forward in time. This avoids all the issues of the Wheeler-de Witt equation to dynamically generate a time dimension.

But at the end, there could exist a Wheeler-de Witt style manner to describe the bulk dynamics of quantum theory of gravity. Some experts believe that this equation still holds the potential for understanding quantum gravity; however, decades after the equation was first written down, it has not brought physicists as clear results about quantum gravity as some of the results building on completely different approaches, such as string theory.

Motivation and background

In canonical gravity, spacetime is foliated into spacelike submanifolds. The three-metric (i.e., metric on the hypersurface) is $\gamma_{ij}$ and given by

$g_{\mu\nu}\,\mathrm{d}x^{\mu}\,\mathrm{d}x^{\nu}=(-\,N^2+\beta_k\beta^k)\,\mathrm{d}t^2+2\beta_k\,\mathrm{d}x^k\,\mathrm{d}t+\gamma_{ij}\,\mathrm{d}x^i\,\mathrm{d}x^j.$

In that equation the Roman indices run over the values 1, 2, 3 and the Greek indices run over the values 1, 2, 3, 4. The three-metric $\gamma_{ij}$ is the field, and we denote its conjugate momenta as $\pi^{kl}$. The Hamiltonian is a constraint (characteristic of most relativistic systems)

$\mathcal{H}=\frac{1}{2\sqrt{\gamma}}G_{ijkl}\pi^{ij}\pi^{kl}-\sqrt{\gamma}\,{}^{(3)}\!R=0$

where $\gamma=\det(\gamma_{ij})$ and $G_{ijkl}=(\gamma_{ik}\gamma_{jl}+\gamma_{il}\gamma_{jk}-\gamma_{ij}\gamma_{kl})$ is the Wheeler-DeWitt metric.

Quantization "puts hats" on the momenta and field variables; that is, the functions of numbers in the classical case become operators that modify the state function in the quantum case. Thus we obtain the operator

$\widehat{\mathcal{H}}=\frac{1}{2\sqrt{\gamma}}\widehat{G}_{ijkl}\widehat{\pi}^{ij}\widehat{\pi}^{kl}-\sqrt{\gamma}\,{}^{(3)}\!\widehat{R}.$

Working in "position space", these operators are

$\hat{\gamma}_{ij}(t,x^k) \to \gamma_{ij}(t,x^k)$

$\hat{\pi}^{ij}(t,x^k) \to -i \frac{\delta}{\delta \gamma_{ij}(t,x^k)}.$

One can apply the operator to a general wave functional of the metric $\widehat{\mathcal{H}} \Psi[\gamma] =0$ where:

$\Psi[\gamma] = a + \int \psi(x) \gamma(x) dx^3+ \int\int \psi(x,y)\gamma(x)\gamma(y) dx^3 dy^3 +...$

Which would give a set of constraints amongst the coefficients $\psi(x,y,...)$. Which means the amplitudes for N gravitons at certain positions is related to the amplitudes for a different number of gravitons at different positions. Or one could use the two field formalism treating $\omega(g)$ as an independent field so the wave function is $\Psi[\gamma,\omega]$

Derivation from path integral

The Wheeler–DeWitt equation can be derived from a path integral using the gravitational action in the Euclidean quantum gravity paradigm:^[3]

$Z = \int_{C}\mathrm{e}^{-I[g_{\mu\nu},\phi]}\mathcal{D}\bold{g}\, \mathcal{D}\phi$

where one integrates over a class of Riemannian four-metrics and matter fields matching certain boundary conditions. Because the concept of a universal time coordinate seems unphysical, and at odds with the principles of general relativity, the action is evaluated around a 3-metric which we take as the boundary of the classes of four-metrics and on which a certain configuration of matter fields exists. This latter might for example be the current configuration of matter in our universe as we observe it today. Evaluating the action so that it only depends on the 3-metric and the matter fields is sufficient to remove the need for a time coordinate as it effectively fixes a point in the evolution of the universe.

We obtain the Hamiltonian constraint from

$\frac{\delta I_{EH}}{\delta N}=0$

where $I_{EH}$ is the Einstein-Hilbert action, and $N$ is the lapse function (i.e., the Lagrange multiplier for the Hamiltonian constraint). This is purely classical so far. We can recover the Wheeler–DeWitt equation from

$\frac{\delta Z}{\delta N}=0=\int \left.\frac{\delta I[g_{\mu\nu},\phi]}{\delta N}\right|_{\Sigma} \exp\left(-I[g_{\mu\nu},\phi]\right)\,\mathcal{D}\bold{g}\, \mathcal{D}\phi$

where $\Sigma$ is the three-dimensional boundary. Observe that this expression vanishes, implying that the functional derivative also vanishes, giving us the Wheeler–DeWitt equation. A similar statement may be made for the diffeomorphism constraint (take functional derivative with respect to the shift functions instead).

Mathematical formalism

The Wheeler–DeWitt equation^[1] is a functional differential equation. It is ill defined in the general case, but very important in theoretical physics, especially in quantum gravity. It is a functional differential equation on the space of three dimensional spatial metrics. The Wheeler–DeWitt equation has the form of an operator acting on a wave functional, the functional reduces to a function in cosmology. Contrary to the general case, the Wheeler–DeWitt equation is well defined in mini-superspaces like the configuration space of cosmological theories. An example of such a wave function is the Hartle–Hawking state. Bryce DeWitt first published this equation in 1967 under the name "Einstein–Schrödinger equation"; it was later renamed the "Wheeler–DeWitt equation".^[4]

Simply speaking, the Wheeler–DeWitt equation says

$\hat{H}(x) |\psi\rangle = 0$

where $\hat{H}(x)$ is the Hamiltonian constraint in quantized general relativity and $|\psi\rangle$ stands for the wave function of the universe. Unlike ordinary quantum field theory or quantum mechanics, the Hamiltonian is a first class constraint on physical states. We also have an independent constraint for each point in space.

Although the symbols $\hat{H}$ and $|\psi\rangle$ may appear familiar, their interpretation in the Wheeler–DeWitt equation is substantially different from non-relativistic quantum mechanics. $|\psi\rangle$ is no longer a spatial wave function in the traditional sense of a complex-valued function that is defined on a 3-dimensional space-like surface and normalized to unity. Instead it is a functional of field configurations on all of spacetime. This wave function contains all of the information about the geometry and matter content of the universe. $\hat{H}$ is still an operator that acts on the Hilbert space of wave functions, but it is not the same Hilbert space as in the nonrelativistic case, and the Hamiltonian no longer determines evolution of the system, so the Schrödinger equation $\hat{H} |\psi\rangle = i \hbar \partial / \partial t |\psi\rangle$ no longer applies. This property is known as timelessness. The reemergence of time requires the tools of decoherence and clock operators (or the use of a scalar field).

We also need to augment the Hamiltonian constraint with momentum constraints

$\vec{\mathcal{P}}(x) \left| \psi \right\rangle = 0$

associated with spatial diffeomorphism invariance.

In minisuperspace approximations, we only have one Hamiltonian constraint (instead of infinitely many of them).

In fact, the principle of general covariance in general relativity implies that global evolution per se does not exist; the time $t$ is just a label we assign to one of the coordinate axes. Thus, what we think about as time evolution of any physical system is just a gauge transformation, similar to that of QED induced by U(1) local gauge transformation $\psi \rightarrow e^{i\theta(\vec{r} )} \psi$ where $\theta(\vec{r})$ plays the role of local time. The role of a Hamiltonian is simply to restrict the space of the "kinematic" states of the Universe to that of "physical" states - the ones that follow gauge orbits. For this reason we call it a "Hamiltonian constraint." Upon quantization, physical states become wave functions that lie in the kernel of the Hamiltonian operator.

In general, the Hamiltonian vanishes for a theory with general covariance or time-scaling invariance.

See also

• ADM formalism
• Diffeomorphism constraint
• Euclidean quantum gravity
• Regge Calculus
• Canonical quantum gravity
• Peres metric
• Loop quantum gravity

References

1. DeWitt, B. S. (1967). "Quantum Theory of Gravity. I. The Canonical Theory". Phys. Rev. 160 (5): 1113–1148. Bibcode:1967PhRv..160.1113D. doi:10.1103/PhysRev.160.1113.
2. https://medium.com/the-physics-arxiv-blog/d5d3dc850933
3. See J. B. Hartle and S. W. Hawking, "Wave function of the Universe." Phys. Rev. D 28 (1983) 2960–2975, eprint.
4. Go to Arxiv.org to read "Notes for a Brief History of Quantum Gravity" by Carlo Rovelli

• Herbert W. Hamber and Ruth M. Williams (2011). "Discrete Wheeler-DeWitt Equation". Physical Review D 84: 104033. arXiv:1109.2530. Bibcode:2011PhRvD..84j4033H. doi:10.1103/PhysRevD.84.104033. Available at .
• Herbert W. Hamber, Reiko Toriumi and Ruth M. Williams (2012). "Wheeler-DeWitt Equation in 2+1 Dimensions". Physical Review D 86: 084010. arXiv:1207.3759. Bibcode:2012PhRvD..86h4010H. doi:10.1103/PhysRevD.86.084010. Available at .