I'm in an introductory particle physics class, and in performing manipulations on the Dirac equation, my instructor expands the $\gamma^\mu \partial_\mu$ term as:

$$\gamma^\mu \partial_\mu = \gamma^0 \frac{\partial}{\partial t} + \vec{\gamma}\cdot \vec{\nabla}$$

where $\vec{\gamma} = (\gamma^1,\gamma^2,\gamma^3)$, but to my knowledge,

$$\gamma^\mu \partial_\mu = \gamma^\mu \eta_{\mu \nu} \partial^\nu = \gamma^0\frac{\partial}{\partial t} - \vec{\gamma} \cdot \vec{\nabla}$$

using the convention $\eta_{\mu \nu}=\operatorname{diag}(+,-,-,-)$. Am I missing something here?

  • 1
    $\begingroup$ you a forgeting an extra minus sign on the $\partial^\nu = (\frac{\partial}{\partial t}, - \nabla)$ $\endgroup$
    – Hydro Guy
    May 24, 2013 at 21:15

1 Answer 1


Yes. You are missing the fact that he is using the convention $$ \nabla = (\partial_1, \partial_2, \partial_3) $$ as opposed to $$ \nabla = (\partial^1, \partial^2, \partial^3) $$ The first convention is by far the most common in my experience.

  • $\begingroup$ I see. Now that I think about it, we didn't actually discuss the difference between $\partial_\mu$ and $\partial^\mu$. Is the only difference a minus sign in the spatial part? $\endgroup$
    – Gino
    May 24, 2013 at 21:22
  • $\begingroup$ Yes that's the difference in this case. This is a special case of a more general notational phenomenon used in lots od contexts called "raising and lowering indices" with a metric. $\endgroup$ May 24, 2013 at 21:26
  • $\begingroup$ Well, it is not only a convention. Lower indices represent covectors, and the derivative is naturally a covector. $\endgroup$ May 24, 2013 at 21:32
  • $\begingroup$ @PeterKravchuk Well, sure; it's a well-motivated convention that is consistent with other notational conventions used for vectors and covectors, but in this context the distinction is largely artificial anyway because of the tangent-cotangent isomorphism on semi-Riemannian manifolds. Also, besides all of this highbrow talk, at some point you have to choose a definition for the symbol $\nabla$, and you might think one definition is more natural than another, but you certainly could have picked the other one if you really wanted to. $\endgroup$ May 24, 2013 at 21:43
  • 1
    $\begingroup$ @Gino en.wikipedia.org/wiki/Raising_and_lowering_indices That should answer most of your questions. Also see idv.sinica.edu.tw/ftliang/diff_geom*diff_geometry(II)/3.04/cotangent_tangent.pdf and en.wikipedia.org/wiki/Killing_form (search for the phrase "raise and lower"). Also, consult any book on GR such as the first couple chapters of preposterousuniverse.com/grnotes $\endgroup$ May 24, 2013 at 22:03

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.