It is well known that the Einstein tensor is trivially 0 in any two dimensions spacetime: \begin{equation}\tag{1} G_{\mu \nu}^{(2)} \equiv 0. \end{equation} Thus, it is not possible to formulate the Einstein equation when the spacetime dimension is $D = 2$ and introduce the cosmological constant in the usual way: \begin{equation}\tag{2} G_{\mu \nu}^{(D)} + \Lambda_D \, g_{\mu \nu}^{(D)} = -\, 8 \pi G_D \, T_{\mu \nu}^{(D)}. \end{equation} This equation imposes $\Lambda_2 = 0$ and $G_2 = 0$ when $D = 2$.

So is there a way to introduce a non-trivial cosmological constant in 2 dimensional spacetimes? Is $\Lambda_2 \equiv 0$ really an unique necessity?


1 Answer 1


One thing to note first is that pure gravity (i.e. $\Lambda = 0$) is a topological theory. What this means is that on any manifold, the Einstein action $S_E = \int d^2 x \sqrt{|g|} R$ only depends on quantities on the boundary. In fact, due to the Gauss-Bonnet theorem, if your manifold is compact and has no boundary, then

\begin{equation} \tag{1} \int d^2 x \sqrt{|g|} R = n \in \mathbb{Z} \end{equation}

where $n = 2 - 2 g$ is the Euler characteristic of a genus g surface, and is an integer which is importantly independent of the metric's specific values. This manifests itself in the fact that in the equations motion for the $R$ term vanish identically, i.e. that (exercise)

\begin{equation} \tag{2} \frac{\delta \sqrt{|g|} R}{\delta g^{\mu\nu}} = 0 \end{equation}

So to compute the equations of motion for the cosmological constant, you get

\begin{equation}\tag{3} \end{equation} \begin{split} & \frac{\delta \Lambda \sqrt{|g|}}{\delta g^{\mu\nu}} \\ &= \Lambda \frac{\delta \sqrt{|g|}}{\delta g^{\mu\nu}} \\ &= -\Lambda \frac{1}{2} g_{\mu \nu} \sqrt{-g} = 0 \end{split}

so the equations of motion are $g_{\mu\nu} = 0$ identically everywhere.

It is worth mentioning that most places where people talk about gravity in 2D is 2D dilaton gravity, where the gravitational field is coupled to an additional scalar $\phi$ with an action looking like

\begin{equation} \tag{4} S_E = \int d^2x e^{\phi}(R + \Lambda + (\partial \phi)^2 ) \end{equation}

These types of actions have relevance to string theory, and in more recent developments, they are thought to arise as low energy theories of a quantum mechanical model known as the SYK model.

  • $\begingroup$ I think there's a missing $\Lambda$ in your equation, in the middle, which should imply $\Lambda = 0$ and not $g_{\mu \nu} = 0$. Also, please, consider adding a number to your main equations, using the command \tag{number} $\endgroup$
    – Cham
    Commented Aug 14, 2019 at 2:45
  • $\begingroup$ Thanks, fixed it. But Lambda is a constant that you stick in by hand, not a dynamical field. So it would mean that $g_{\mu\nu}=0$ (assuming that $\sqrt{g} \neq 0$) $\endgroup$
    – Joe
    Commented Aug 14, 2019 at 2:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.