Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Semiclassical limit: WKB phase obeys the Einstein–Hamilton–Jacobi equation

Proved
WheelerDeWittSuperspace.wdw_wkb

by Mazecto · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

general-relativitymathematical-physicsquantum-gravity

Let ℏ≠0\hbar\neq0ℏ=0, let SSS be a real C2C^2C2 function on configuration space and RRR any supplied potential. At every site,

H^ eiS/ℏ=eiS/ℏ(2κ Gabcd ∂abS ∂cdS−det⁡h2κ(R−2Λ)  −  iℏ 2κ Gabcd ∂ab∂cdS),\widehat{\mathcal H}\,e^{iS/\hbar}=e^{iS/\hbar}\Big(2\kappa\,G_{abcd}\,\partial_{ab}S\,\partial_{cd}S-\frac{\sqrt{\det h}}{2\kappa}(R-2\Lambda)\;-\;i\hbar\,2\kappa\,G_{abcd}\,\partial_{ab}\partial_{cd}S\Big),HeiS/ℏ=eiS/ℏ(2κGabcd​∂ab​S∂cd​S−2κdeth​​(R−2Λ)−iℏ2κGabcd​∂ab​∂cd​S),

where H^=−2κℏ2Gabcd∂ab∂cd−det⁡h2κ(R−2Λ)\widehat{\mathcal H}=-2\kappa\hbar^2G_{abcd}\partial_{ab}\partial_{cd}-\tfrac{\sqrt{\det h}}{2\kappa}(R-2\Lambda)H=−2κℏ2Gabcd​∂ab​∂cd​−2κdeth​​(R−2Λ) with the supermetric to the left. The ℏ\hbarℏ-independent part is the Einstein–Hamilton–Jacobi functional.

Preamble
import Definitions.Def_WheelerDeWittSuperspace

set_option autoImplicit false

open Matrix
Formal statement
namespace WheelerDeWittSuperspace

/-- Milestone 5 (semiclassical limit): for `Ψ = exp(iS/hbar)`, the Wheeler–DeWitt operator is
`exp(iS/hbar)` times the Einstein–Hamilton–Jacobi functional of `S`, plus an explicit
first-order correction in `hbar`. -/
theorem wdw_wkb {X : Type*} [Fintype X] [DecidableEq X] (kappa hbar Lam : ℝ) (hkappa : kappa ≠ 0)
    (hhbar : hbar ≠ 0)
    (R : Config X → X → ℝ) (S : Config X → ℝ) (hS : ContDiff ℝ 2 S)
    (h : Config X) (x : X) :
    wdw kappa hbar Lam R (fun h' => Complex.exp (Complex.I * (S h' : ℂ) / (hbar : ℂ))) h x =
      Complex.exp (Complex.I * (S h : ℂ) / (hbar : ℂ)) *
        ((hamiltonJacobi kappa Lam R S h x : ℂ) -
          Complex.I * (hbar : ℂ) * (2 * (kappa : ℂ)) *
            ∑ a, ∑ b, ∑ c, ∑ d, (deWitt (metricAt h x) a b c d : ℂ) *
              (partialR (partialR S x c d) x a b h : ℂ)) := by sorry

end WheelerDeWittSuperspace
Source
C. Kiefer, Quantum Geometrodynamics: whence, whither?, Gen. Relativ. Gravit. 41 (2009) 877-901, https://arxiv.org/abs/0812.0295, Section 3.1 eq. (10) and Section 3.2 eq. (11); A. Peres, Nuovo Cimento 26 (1962) 53-62, https://doi.org/10.1007/BF02754342
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Theorem wdw_wkb (read-back).

Setting. Let XXX be any finite type with decidable equality (a finite set of "sites"; it may a priori be empty, but the statement also takes a site x∈Xx \in Xx∈X, so it only has content when X≠∅X \neq \varnothingX=∅). A configuration is a function h:X→R3×3h : X \to \mathbb{R}^{3\times 3}h:X→R3×3, i.e. at each site yyy an arbitrary real 3×33\times 33×3 array hab(y)h_{ab}(y)hab​(y), a,b∈{0,1,2}a,b \in \{0,1,2\}a,b∈{0,1,2}. All nine entries at each site are independent coordinates: configurations are not required to be symmetric, positive definite, or nondegenerate. The configuration space C=(R3×3)X≅R9∣X∣\mathcal{C} = (\mathbb{R}^{3\times 3})^X \cong \mathbb{R}^{9|X|}C=(R3×3)X≅R9∣X∣ is a finite-dimensional real vector space. For a site yyy write m=h(y)m = h(y)m=h(y) for the 3×33\times 33×3 matrix with entries mab=hab(y)m_{ab} = h_{ab}(y)mab​=hab​(y).

Data and hypotheses. The theorem is stated for

  • real numbers κ\kappaκ, ℏ\hbarℏ, Λ\LambdaΛ with hypotheses κ≠0\kappa \neq 0κ=0 and ℏ≠0\hbar \neq 0ℏ=0 (Λ\LambdaΛ is unrestricted);
  • an arbitrary function R:C×X→RR : \mathcal{C} \times X \to \mathbb{R}R:C×X→R, (h,y)↦R(h,y)(h, y) \mapsto R(h, y)(h,y)↦R(h,y) (no regularity, symmetry or locality assumed);
  • a function S:C→RS : \mathcal{C} \to \mathbb{R}S:C→R with the hypothesis that SSS is C2C^2C2 (twice continuously Fréchet differentiable over R\mathbb{R}R) on all of C\mathcal{C}C;
  • an arbitrary configuration h∈Ch \in \mathcal{C}h∈C and an arbitrary site x∈Xx \in Xx∈X (no physicality condition on hhh).

Definitions used (unfolded).

  • Volume density: for a real 3×33\times 33×3 matrix mmm, m:=det⁡m\sqrt{m} := \sqrt{\det m}m​:=detm​, where the real square root is taken to be 000 for negative arguments. Hence m=0\sqrt{m} = 0m​=0 whenever det⁡m≤0\det m \le 0detm≤0.
  • DeWitt supermetric:
Gabcd(m)=macmbd+madmbc−mabmcd2det⁡m,G_{abcd}(m) = \frac{m_{ac}m_{bd} + m_{ad}m_{bc} - m_{ab}m_{cd}}{2\sqrt{\det m}},Gabcd​(m)=2detm​mac​mbd​+mad​mbc​−mab​mcd​​,

with the convention that division by 000 gives 000. So Gabcd(m)=0G_{abcd}(m) = 0Gabcd​(m)=0 for all indices whenever det⁡m≤0\det m \le 0detm≤0. Since mmm need not be symmetric, GGG is formed literally from the given entries.

  • Coordinate direction: ex,ab∈Ce_{x,ab} \in \mathcal{C}ex,ab​∈C is the configuration that is 111 in slot (x,a,b)(x,a,b)(x,a,b) and 000 everywhere else (only that single ordered entry; ex,ab≠ex,bae_{x,ab} \neq e_{x,ba}ex,ab​=ex,ba​ for a≠ba \neq ba=b).
  • Partial derivatives: for F:C→RF : \mathcal{C} \to \mathbb{R}F:C→R (resp. C\mathbb{C}C), ∂ab(x)F(h):=DF(h) [ex,ab]\partial_{ab(x)} F(h) := DF(h)\,[e_{x,ab}]∂ab(x)​F(h):=DF(h)[ex,ab​], the real Fréchet derivative of FFF at hhh applied to ex,abe_{x,ab}ex,ab​ (for complex-valued FFF, C\mathbb{C}C is treated as a real vector space). If FFF is not Fréchet differentiable at hhh, this is defined to be 000. Iterated: ∂ab(x)∂cd(x)F\partial_{ab(x)}\partial_{cd(x)}F∂ab(x)​∂cd(x)​F means first take ∂cd(x)F\partial_{cd(x)}F∂cd(x)​F as a function on C\mathcal{C}C, then apply ∂ab(x)\partial_{ab(x)}∂ab(x)​ to it. Both derivatives are at the same site xxx.
  • Kinetic term: for Ψ:C→C\Psi : \mathcal{C} \to \mathbb{C}Ψ:C→C,
K[Ψ](h,x)=∑a,b,c,d=02Gabcd(h(x))  ∂ab(x)∂cd(x)Ψ(h).K[\Psi](h,x) = \sum_{a,b,c,d=0}^{2} G_{abcd}(h(x))\; \partial_{ab(x)}\partial_{cd(x)}\Psi(h).K[Ψ](h,x)=a,b,c,d=0∑2​Gabcd​(h(x))∂ab(x)​∂cd(x)​Ψ(h).
  • Wheeler–DeWitt operator at site xxx:
H[Ψ](h,x)=−2κℏ2 K[Ψ](h,x)  −  (2κ)−1det⁡h(x) (R(h,x)−2Λ) Ψ(h).\mathcal{H}[\Psi](h,x) = -2\kappa\hbar^2\, K[\Psi](h,x) \;-\; (2\kappa)^{-1}\sqrt{\det h(x)}\,\bigl(R(h,x) - 2\Lambda\bigr)\,\Psi(h).H[Ψ](h,x)=−2κℏ2K[Ψ](h,x)−(2κ)−1deth(x)​(R(h,x)−2Λ)Ψ(h).
  • Einstein–Hamilton–Jacobi functional:
HJ[S](h,x)=2κ∑a,b,c,dGabcd(h(x))  ∂ab(x)S(h)  ∂cd(x)S(h)  −  (2κ)−1det⁡h(x) (R(h,x)−2Λ),\mathrm{HJ}[S](h,x) = 2\kappa \sum_{a,b,c,d} G_{abcd}(h(x))\; \partial_{ab(x)}S(h)\; \partial_{cd(x)}S(h) \;-\; (2\kappa)^{-1}\sqrt{\det h(x)}\,\bigl(R(h,x) - 2\Lambda\bigr),HJ[S](h,x)=2κa,b,c,d∑​Gabcd​(h(x))∂ab(x)​S(h)∂cd(x)​S(h)−(2κ)−1deth(x)​(R(h,x)−2Λ),

a real number.

Conclusion. Let Ψ(h′)=exp⁡ ⁣(i S(h′)/ℏ)\Psi(h') = \exp\!\bigl(i\,S(h')/\hbar\bigr)Ψ(h′)=exp(iS(h′)/ℏ) for h′∈Ch' \in \mathcal{C}h′∈C (complex exponential, SSS and ℏ\hbarℏ regarded as complex numbers). Then, for the given hhh and xxx,

H[Ψ](h,x)  =  eiS(h)/ℏ(HJ[S](h,x)  −  i ℏ (2κ)∑a,b,c,d=02Gabcd(h(x))  ∂ab(x)∂cd(x)S(h)),\mathcal{H}[\Psi](h,x) \;=\; e^{iS(h)/\hbar}\Bigl(\mathrm{HJ}[S](h,x) \;-\; i\,\hbar\,(2\kappa)\sum_{a,b,c,d=0}^{2} G_{abcd}(h(x))\;\partial_{ab(x)}\partial_{cd(x)}S(h)\Bigr),H[Ψ](h,x)=eiS(h)/ℏ(HJ[S](h,x)−iℏ(2κ)a,b,c,d=0∑2​Gabcd​(h(x))∂ab(x)​∂cd(x)​S(h)),

an equality of complex numbers, where the real quantities HJ[S](h,x)\mathrm{HJ}[S](h,x)HJ[S](h,x), GabcdG_{abcd}Gabcd​ and ∂ab(x)∂cd(x)S(h)\partial_{ab(x)}\partial_{cd(x)}S(h)∂ab(x)​∂cd(x)​S(h) are embedded in C\mathbb{C}C. The second-derivative term uses the same index order as the kinetic term (∂cd\partial_{cd}∂cd​ applied first, then ∂ab\partial_{ab}∂ab​).

Edge cases and remarks.

  • No positivity or symmetry is assumed on h(x)h(x)h(x). If det⁡h(x)≤0\det h(x) \le 0deth(x)≤0, then det⁡h(x)=0\sqrt{\det h(x)} = 0deth(x)​=0 and every Gabcd(h(x))=0G_{abcd}(h(x)) = 0Gabcd​(h(x))=0, so both the potential and all supermetric terms vanish on both sides; the identity then reads 0=00 = 00=0. If det⁡h(x)>0\det h(x) > 0deth(x)>0 but h(x)h(x)h(x) is not symmetric, the formulas are applied verbatim to the non-symmetric entries.
  • Derivatives are with respect to each of the nine ordered entries separately; there is no symmetrization (no factor relating ∂ab\partial_{ab}∂ab​ and ∂ba\partial_{ba}∂ba​).
  • The potential term on the left contains Ψ(h)=eiS(h)/ℏ\Psi(h) = e^{iS(h)/\hbar}Ψ(h)=eiS(h)/ℏ and appears on the right inside HJ\mathrm{HJ}HJ; the function RRR enters only through its value R(h,x)R(h,x)R(h,x), so no regularity of RRR is needed or assumed.
  • The hypothesis ℏ≠0\hbar \neq 0ℏ=0 excludes the case where iS/ℏiS/\hbariS/ℏ would be evaluated with the convention 1/0=01/0 = 01/0=0 (making Ψ≡1\Psi \equiv 1Ψ≡1); κ≠0\kappa \neq 0κ=0 excludes (2κ)−1=0(2\kappa)^{-1} = 0(2κ)−1=0. Both hypotheses are satisfiable, as is C2C^2C2-smoothness of SSS (e.g. SSS constant or linear), so the statement is not vacuous. With XXX nonempty, all binders can be instantiated.
  • The C2C^2C2 hypothesis is global on C\mathcal{C}C; it guarantees the Fréchet derivatives involved exist, so the "derivative =0=0=0 when not differentiable" convention does not intervene for SSS or Ψ\PsiΨ.
  • The identity is asserted pointwise for every single hhh and xxx; it says nothing about H[Ψ]\mathcal{H}[\Psi]H[Ψ] vanishing (no constraint equation is imposed) and makes no asymptotic claim in ℏ\hbarℏ: it is an exact equality with an explicit term linear in ℏ\hbarℏ.
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Mazecto · Oct 8, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me