The text I'm reading claims that in an infinitesimally small region of the manifold, we can always find observers that locally see the metric as either Minkowski, $\eta_{\mu\nu}$, or Euclidean, $\delta_{\mu\nu}$. The observer seeing a Minkowski metric makes sense to me, if we look at a small enough region the curvature is small enough that the metric reduces locally to that of flat space, at least that's what I think. But finding the metric to be Euclidean implies $diag(1,1,1,1)$, unless I am wrong about that. My question is in what physical situation could an observer see a Euclidean metric?

  • $\begingroup$ Which text? The first "or" should be an "xor": The manifold is Euclidean xor Lorentzian. Not both. Possible duplicate: physics.stackexchange.com/q/129187/2451 $\endgroup$
    – Qmechanic
    Mar 31, 2020 at 15:15
  • $\begingroup$ Hmmm. What's the text? I guess they may mean that if all velocities are relatively small, then we get Galilean relativity. $\endgroup$
    – PM 2Ring
    Mar 31, 2020 at 15:16
  • $\begingroup$ The notes for my GR course, perhaps text was a poor choice of wording, the exact quote is "For this, let us assume that we start with an infinitesimally small region of our manifold. In this region one can find a set of ‘privileged’ observers that locally see the metric as trivial (the LIF for GR). This means that in this coordinate system the metric is diagonal, with +1 or −1 components (two examples are δµν or ηµν) and has vanishing first derivatives" $\endgroup$
    – Charlie
    Mar 31, 2020 at 15:17
  • $\begingroup$ I think that the text could be referring to that the spacetime metric is locally Minkowskian $ diag(1,-1,-1,-1) $ and then the projected 3D space-type for an instant $ t=t_0 $ is euclidean $ (1,1,1) $ $\endgroup$ Mar 31, 2020 at 15:22
  • $\begingroup$ Maybe, the notes have consistently used the (-,+,+,+) signature though. Also the indices on the Kronecker delta metric are spacetime indices which unless the notation has been confused suggests the entire metric is the identity. $\endgroup$
    – Charlie
    Mar 31, 2020 at 15:26

1 Answer 1


He can't see a Euclidean metric, except by restricting to space.

I think the writer did not mean what he appears to say, as quoted in the comment. I think he only intended to give two examples of diagonal metrics, not to give two examples of metrics which an observer can actually see.

  • $\begingroup$ That makes more sense, so there is definitely no local frame in which the metric is the 4d identity? $\endgroup$
    – Charlie
    Mar 31, 2020 at 16:06
  • 1
    $\begingroup$ Absolutely no possible example. Remember, it is a fundamental principle that the laws of physics are the same for all observers. It follows that any inertial observer can see a locally Minkowski metric. The metric signature refers to the sign of the eigenvalues; we must always have one for time and three for space, and opposite signs for time and space. $\endgroup$ Mar 31, 2020 at 16:13
  • $\begingroup$ Great, question answered, thank you :) $\endgroup$
    – Charlie
    Mar 31, 2020 at 16:14

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.