0
$\begingroup$

Let $(M,g)$ be a Riemannian $n$-manifold, and $\varphi: M \to \Bbb R$ be a smooth map. Define another metric tensor by: $$\widetilde{g} = g - {\rm d}\varphi \otimes {\rm d}\varphi$$I know that $\widetilde{g}$ is Riemannian if $\|{\rm grad}_M\varphi\| < 1$ and Lorentzian if $\|{\rm grad}_M\varphi\| > 1$ $(\|\cdot\|$ induced by $g$, of course$)$.

To see the second case it is easy, it suffices to evaluate $\widetilde{g}$ in ${\rm grad}_M\varphi$ itself - it will be a timelike vector.

Is there an easy way to check that $\widetilde{g}$ will be positive definite? I don't quite know if there's a result about this, but I feel that computing the metric on ${\rm grad}_M\varphi$ should be enough. I'm not sure how to exactly compute $\det \widetilde{g}$ here... Thanks.

$\endgroup$
1
$\begingroup$

Take a vector field $v$, and compute $\bar g(v,v)$. You need to show this is strictly positive when $v \ne 0$. But $\bar g(v,v) = g(v,v) - (d\varphi(v))^2$. Isn't this enough to make it work?

$\endgroup$
  • $\begingroup$ Oh yes, it is. Thanks, I don't know what I was thinking. $${\rm d}\varphi(v)^2 = g({\rm grad}_M\varphi, v)^2 \leq g(v,v).$$ $\endgroup$ – Ivo Terek Jul 15 '15 at 0:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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