---
title: Causal Dynamical Triangulations
url: https://www.emergentmind.com/topics/causal-dynamical-triangulations
type: topic
---

# Causal Dynamical Triangulations

Causal Dynamical Triangulations (CDT) is a nonperturbative, background-independent approach to quantum gravity in which the gravitational path integral is defined as a statistical sum over causally well-behaved, piecewise-linear Lorentzian spacetimes. The method implements a lattice regularization by restricting the sum to simplicial manifolds with a fixed causal structure, enabling the emergence of semiclassical geometry and the systematic study of quantum spacetime at the Planck scale. CDT has yielded quantitative results regarding phase structure, critical phenomena, and continuum physics, and is central to contemporary research in quantum gravity.

## 1. Formalism and Defining Principles

CDT regularizes the continuum gravitational path integral,
\[
Z = \int \mathcal{D}[g]\, e^{iS_{EH}[g]},
\]
by replacing it with a sum over triangulations built from $d$-simplices, with lengths determined by the lattice spacing $a$ and a parameter $\alpha > 0$ distinguishing spacelike ($\ell^2_s = a^2$) and timelike ($\ell^2_t = -\alpha a^2$) edges. The fundamental principles underlying CDT are:

- **Global Proper-Time Foliation:** Triangulations are required to admit a global foliation by discrete, spatial slices of fixed topology, distinguishing time direction and enforcing causality at the microscopic (lattice) level [2401.09399, 1004.0352, 1302.2440, 1410.0670].
- **Causal Structure:** Only causal (Lorentzian-signature) triangulations compatible with fixed spatial topology are included, forbidding branching and spatial topology change [1111.6938, 1302.2173, 1410.0670].
- **Wick Rotation:** The ensemble is constructed such that analytic continuation $\alpha \mapsto -\alpha$ is possible, yielding a statistical model with Euclidean Boltzmann weights suitable for Monte Carlo simulations [2401.09399, 1302.2173].
- **Regge Action:** The discretized action is a Regge–Einstein–Hilbert action, typically of the form
  \[
  S_{EH} = -(\kappa_0 + 6\Delta) N_0 + \kappa_4(N_{4,1} + N_{3,2}) + \Delta\,[2N_{4,1} + N_{3,2}],
  \]
  with $N_0$ the number of vertices, $N_{4,1}$ and $N_{3,2}$ counts of distinct 4-simplices, and $(\kappa_0, \kappa_4, \Delta)$ bare couplings [1904.05755, 1302.2173].

## 2. Structure of the Path Integral and Phase Diagram

The regularized path integral is written as
\[
Z_E(\kappa_0, \kappa_4; \alpha) = \sum_T \frac{1}{C(T)}\, e^{-S_E(T,\alpha)},
\]
where $T$ runs over all triangulations with the prescribed causal and topological constraints, $C(T)$ is a symmetry factor, and $S_E$ is the Wick-rotated action [2401.09399, 1302.2173, 1004.0352]. Monte Carlo methods allow the statistical study of geometric observables, including the spatial volume profile, spectral dimension, and curvature-related quantities.

CDT in four dimensions exhibits a rich phase structure in the $(\kappa_0,\Delta)$ parameter space, with phases characterized as follows [1704.00577, 1904.05755, 1111.6938, 1512.08997]:

| Phase        | Geometry                    | Physical Significance          | Order of Transition |
|--------------|----------------------------|-------------------------------|---------------------|
| A            | Branched-polymer, oscillatory | No extended geometry          | 1st (to $C_{dS}$)   |
| B            | Collapsed, time-uniform     | Crumpled, unphysical          | 2nd or higher       |
| $C_{\mathrm{dS}}$ | Semiclassical, de Sitter-like | Physical, GR-like            | 1st (to A), 2nd+ (to B) |
| $C_b$ (bifurcation) | Clustering/black balls, inhomogeneous | Unusual correlations         | 2nd or higher       |

Second (or higher) order transitions are present at the $B$–$C_{\mathrm{dS}}$ and $C_{\mathrm{dS}}$–$C_b$ boundaries, opening the possibility of a nontrivial continuum limit via Wilsonian renormalization [1704.00577, 1904.05755].

## 3. Semiclassical Geometry and Effective Dynamics

CDT demonstrates the emergence of a semiclassical, four-dimensional de Sitter geometry in the $C_{\mathrm{dS}}$ phase. The expectation value of the spatial three-volume as a function of proper time follows [1111.6938, 2401.09399, 1302.2173, 1302.1093]
\[
\langle N_3(i)\rangle \propto N_4^{3/4} \cos^3\left(\frac{i}{s_0 N_4^{1/4}}\right),
\]
matching the Euclidean de Sitter universe in minisuperspace reduction. Fluctuations about this background scale as $N_4^{-1/4}$ and can be described by an effective action of the form
\[
S[n_t] = \frac{1}{\Gamma}\sum_t \left[\frac{(n_{t+1} - n_t)^2}{n_{t+1} + n_t} + \mu\, n_t^{1/3} - \lambda n_t\right],
\]
which is identical to the discretized minisuperspace action for the scale factor [1302.1093, 1302.2810, 1102.4643]. This structure is robust under inclusion of additional simplex types and persists across different topologies [1705.07653, 1904.05755].

The kinetic term in the effective action maintains a universal structure, while the potential term reflects topology: for spherical topology ($S^3$), a classical curvature term $\propto n_t^{1/3}$ appears, while for $T^3$ topology no such term is present and a quantum correction dominates [1705.07653].

## 4. Critical Phenomena, Order Parameters, and Topology

Critical behavior and the continuum limit in CDT are analyzed via order parameters derived from global geometric quantities, most notably
\[
OP_2 \equiv \frac{N_{3,2}}{N_{4,1}},
\]
with associated observables: susceptibility $\chi_{OP_2} = \langle OP_2^2 \rangle - \langle OP_2 \rangle^2$ and Binder cumulant $B_{OP_2}$ [1904.05755]. Finite size scaling of the pseudocritical point is characterized by
\[
\kappa_0^{\mathrm{crit}}(N_{4,1}) = \kappa_0^{\mathrm{crit}}(\infty) - C N_{4,1}^{-1/\gamma},
\]
where the shift exponent $\gamma \simeq 1$ in both spherical and toroidal topologies, indicating a first order $A$–$C$ transition independent of topology [1904.05755]. Universal scaling of susceptibilities and Binder cumulants is found.

Spatial topology is fixed in all simulations. The order, critical exponents, and phase structure are remarkably robust against the choice of topology (spherical $S^3$ vs. toroidal $T^3$) and other simulation details (time slicing, volume fixing), with only the geometry and finite size effects differing markedly [1904.05755, 1705.07653].

## 5. Spectral Methods and Multiscale Geometry

Spectral analysis, particularly of the discrete Laplace–Beltrami operator on spatial slices, yields significant geometric and critical information [1912.11311]. For a dual graph encoding the triangulation, the spectral gap (smallest nonzero Laplacian eigenvalue) serves as an order parameter for distinguishing phase behavior:

- Large, non-decaying gap: collapsed (B-like) geometries.
- Vanishing gap with increasing volume: extended (C-like) or fractal geometries.

Effective spectral dimension $d_{\mathrm{EFF}}(\lambda)$, defined via
\[
d_{\mathrm{EFF}}(\lambda) \equiv 2 \frac{d\log(n/V)}{d\log \lambda},
\]
reveals running dimensionality and fractal properties at different scales [1912.11311, 1512.08997]. The scaling of spectral gaps across transitions (critical scaling of eigenvalues)
\[
\langle \lambda_n \rangle_{\infty} = A_n (\Delta_c - \Delta)^{2\nu}
\]
provides evidence for second order transitions and universality. The spectral data enable the identification of scale-dependent observables, criticality, and universality classes for continuum quantum gravity [1912.11311].

## 6. Connections to Statistical Field Theory and Effective Field Theory

Applying Landau theory to CDT, with the spatial slice volume profile as the order parameter, clarifies the phase structure and effective field theory description [2212.11043]. In particular:

- In 2D, the volume profile is homogeneous; effective action maps to 2D Hořava–Lifshitz gravity in proper-time gauge.
- In 3D (and analogously in 4D), condensation phenomena ("blob" or "droplet" phases) arise, corresponding to a nontrivial spatial volume distribution.
- Effective theories derived from CDT possess only foliation-preserving diffeomorphism invariance (FPD), not the full spacetime diffeomorphism invariance of general relativity, unless additional parameter fine-tuning is performed.

CDT's critical structure aligns naturally with Hořava–Lifshitz gravity: the continuum limit lies generically in the HL theory space with FPD symmetry, while recovery of general relativity would require enhanced symmetry via parameter fine-tuning [2212.11043].

| Dimension | Order Parameter | Effective Action                    | Continuum Symmetry   |
|-----------|----------------|-------------------------------------|----------------------|
| 2         | Slice length   | $S_{2d-HL} = \int dt\, \dot{\ell}^2/4\ell$ | FPD                 |
| 3         | Slice area     | $S_{3d-HL-mini} = \int dt(\dot{\phi}^2 - \xi/\phi^2)$ | FPD (GR for tuned parameters) |

## 7. Phenomenological and Observational Implications

CDT predicts several phenomena of empirical relevance:

- **Dimensional Reduction:** The spectral dimension runs from $d_S \simeq 4$ at large scales to $d_S \simeq 2$ at short, Planckian scales—a feature observed across various quantum gravity scenarios [1512.08997].
- **Modified Dispersion:** The scale dependence of the Laplace spectrum implies modified photon and graviton dispersion relations, subject to constraints from astrophysical observation (e.g., gamma-ray bursts: $E_\ast > 6.7\times 10^{10}\,\mathrm{GeV}$ at 95% confidence) [1512.08997].
- **Cosmological Fluctuations:** Modified spectral running could imprint on primordial perturbations, constraining CDT scenarios through CMB data. Some low-energy dimensional reduction scenarios are empirically excluded.
- **Gravitational Defects:** If CDT phase transitions occurred in cosmological history and were second order, formation of gravitational topological defects could occur. Their non-observation constrains allowed scenarios without or with insufficient inflation [1512.08997].

## 8. Computational Techniques and Universality

CDT simulations use ergodic local moves (Pachner moves and their Lorentzian causal extensions) to generate ensembles of triangulations with prescribed global topology, causality, and spacetime volume. The discretization implements a statistical field theory model amenable to finite-size scaling and renormalization group analysis. Extensive studies confirm the universality of continuum CDT results in 2D and higher, independent of regularization details, supporting minimal dependence on the discretization scheme [1302.2440]. Recent computational advances incorporate spectral estimators, transfer matrix extraction, and sophisticated sampling (e.g., parallel tempering) [1302.1093, 1912.11311, 1904.05755].

## 9. Outlook and Theoretical Significance

CDT achieves a nonperturbative definition of quantum gravity that is mathematically precise, computationally accessible, and physically rich. Its key results—emergence of a semiclassical (de Sitter) universe, fractal and dimensional reduction at short scales, and a robust critical phase structure—demonstrate that macroscopic spacetime geometry can arise dynamically from Planckian quantum fluctuations in a background-independent path integral [2401.09399, 1302.2173, 1302.1093].

The phase structure and critical phenomena are robust against variations in spatial topology and technical parameters, indicating that the emergent continuum physics is not an artifact of lattice implementation. CDT provides a controlled setting for probing the ultraviolet completion of gravity, serving as a laboratory for asymptotic safety and for testing scenarios in which the low-energy limit supports general relativity or Hořava–Lifshitz gravity [2212.11043, 1302.2173]. The systematic mapping of universality classes, scaling exponents, and effective continuum actions remains a central direction for resolving outstanding questions regarding the continuum limit and the dynamical restoration (or breaking) of spacetime symmetries.

Source: https://www.emergentmind.com/topics/causal-dynamical-triangulations