ADM Formalism (Introduction)

Core Idea

The ADM (Arnowitt-Deser-Misner) formalism recasts general relativity as an initial-value (Cauchy) problem by decomposing four-dimensional spacetime into a foliation of three-dimensional spatial hypersurfaces evolving in time. The spacetime metric is decomposed into the spatial 3-metric γ_ij (geometry of each slice), the lapse function N (rate of proper time flow between slices), and the shift vector N^i (how spatial coordinates slide between slices). The Einstein equations split into constraint equations (Hamiltonian and momentum constraints, which the initial data must satisfy) and evolution equations (which propagate the data forward). The canonical variables are (γ_ij, π^{ij}), where π^{ij} is the conjugate momentum related to the extrinsic curvature K_ij. The ADM formalism is the foundation of numerical relativity (computational solution of Einstein's equations) and the starting point for canonical quantization of gravity.

Explainer

The Einstein field equations G_μν = (8πG/c⁴)T_μν are 10 coupled, nonlinear partial differential equations that treat space and time on equal footing — they are covariant, with no preferred time direction. But physical problems often require initial-value formulations: given the state of the gravitational field at one moment, predict its future evolution. The ADM formalism, developed by Arnowitt, Deser, and Misner in 1959-1962, provides exactly this by decomposing 4D spacetime into a sequence of 3D spatial slices (a foliation), each labeled by a time coordinate t.

The 4D metric is decomposed in terms of quantities on each slice. The spatial 3-metric γ_ij describes the intrinsic geometry of each slice (distances, angles, curvature within the slice). The extrinsic curvature K_ij describes how each slice is embedded in the surrounding 4D spacetime — roughly, it measures how the slice is "bent." The lapse function N specifies the proper time between adjacent slices (how fast time flows at each point), and the shift vector N^i specifies how spatial coordinates slide sideways between slices. The 4D line element becomes ds² = -N²c²dt² + γ_ij(dx^i + N^i c dt)(dx^j + N^j c dt). The lapse and shift are gauge variables — freely choosable — corresponding to the four coordinate degrees of freedom in GR.

The Einstein equations decompose into two types. The constraint equations — the Hamiltonian constraint and three momentum constraints — involve only the spatial metric γ_ij and the extrinsic curvature K_ij (no time derivatives). They must be satisfied on every spatial slice and correspond to the G_{0μ} components of the Einstein equations. The evolution equations — corresponding to the G_{ij} components — contain time derivatives and propagate (γ_ij, K_ij) from one slice to the next. The Bianchi identity guarantees that if the constraints are satisfied on the initial slice, the evolution equations preserve them automatically. This separation into constraints and evolution is the key structural insight that makes the initial-value problem well-defined.

The ADM formalism casts GR as a Hamiltonian system with canonical variables (γ_ij, π^{ij}), where π^{ij} is the momentum conjugate to γ_ij (related to K_ij by π^{ij} = √γ(K^{ij} - γ^{ij}K)). The Hamiltonian is a sum of constraints: H = ∫(NH + N^i H_i) d³x, where H = 0 and H_i = 0 are the constraint equations. This "vanishing Hamiltonian" structure is a consequence of the diffeomorphism invariance of GR and lies at the heart of the "problem of time" in quantum gravity — the Hamiltonian generates gauge transformations (coordinate changes) rather than physical time evolution. The ADM formalism is the starting point for both numerical relativity (where the 3+1 decomposition is implemented computationally to simulate black hole mergers, neutron star collisions, and cosmological dynamics) and canonical quantum gravity (where γ_ij and π^{ij} are promoted to operators, leading to the Wheeler-DeWitt equation).