I was doing a problem regarding field theory. I am given the following lagrangian density: $$\mathcal{L}=\frac{1}{2}\partial_\mu\phi_i\partial^\mu\phi_i-\frac{m^2}{2}\phi_i\phi_i$$ for three scalar fields. I want to determine the equations of motion for the field $\phi_i$. I used the Euler Lagrange equation:$$\frac{\partial\mathcal{L}}{\partial\phi_j}-\partial_\mu[\frac{\partial\mathcal{L}}{\partial(\partial_\mu\phi_j)}]=0.$$ My question is, is there a difference between using the Euler Lagrange equation the way I wrote above and the Euler Lagrange equation with an upper $\mu$ index: $$\frac{\partial\mathcal{L}}{\partial\phi_j}-\partial^\mu[\frac{\partial\mathcal{L}}{\partial(\partial^\mu\phi_j)}]=0~?$$ I ask this because by the end of my calculations, if I use the former equations , I get $$\partial_\nu\partial^\nu\phi_j+m^2\phi_j=0$$ while the latter equations give: $$\partial^\nu\partial_\nu\phi_j+m^2\phi_j=0.$$ Are both the Klein-Gordon Equation or just the last one?


2 Answers 2


As mentioned in the comments by WarreG and user2723984, since the full expression of the two in terms of the metric tensor is $$ \partial_\nu \partial^\nu=\partial_\nu g^{\mu\nu}\partial_\mu=\partial^\mu\partial_\mu, $$ the two equations are identical.

  • 1
    $\begingroup$ Isn't it a bit dangerous to write $\partial_\nu g^{\mu\nu}\partial_\mu$ which easily can be interpreted as $(\partial_\nu g^{\mu\nu}) \partial_\mu$? $\endgroup$
    – md2perpe
    Jan 11, 2019 at 20:51
  • 4
    $\begingroup$ $g_{\mu\nu}$ is constant, if it wasn't the equation wouldn't be correct either way. $\endgroup$
    – MannyC
    Jan 11, 2019 at 21:07

I would say the answer to your question is: it depends on the metric. Yes, because if your metric is Minkowski and Cartesian then $$\partial_{\mu}\partial^{\mu} = \partial_{\mu}\left(g^{\mu\nu}\partial_{\nu}\right)=\underbrace{\left(\partial_{\mu}g^{\mu\nu}\right)}_{0}\partial_{\nu}+g^{\mu\nu}\left(\partial_{\mu}\partial_{\nu}\right)=g^{\mu\nu}\partial_{\mu}\partial_{\nu}=\partial^{\nu}\partial_{\nu}.$$ But for example, if your metric is Minkowski and spherical? In this case $\partial_{\mu}g^{\mu\nu}\neq0$ and $\partial^{\mu}\partial_{\mu}\neq\partial_{\mu}\partial^{\mu}$.

As you can see, $\partial_{\mu}\partial^{\mu}$ and $\partial^{\mu}\partial_{\mu}$ are equivalent as long as you understand that you are dealing with Cartesian coordinates.


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.