2
$\begingroup$

The covariant derivative of a vector $A^{\mu}$ at a point $x$ is defined as

$$D_z A^{\mu}=\partial_zA^{\mu}+\Gamma^{\mu}_{\rho\sigma}(x)\partial_{z}x^{\rho}A^{\sigma}$$

where Greek symbols are spacetime (dimension $d$) and $z$ is the index of a worldsheet coordinate $\sigma$.

Why doesn't the vector $A^{\mu}$ act as a scalar with respect to the $z$ derivative? How can I prove this?

Assume $A^{\mu}$ depends on $x^\mu(\sigma^b)$, i.e. $A^{\mu}(x^\mu(\sigma^b))$.

$\endgroup$
3
  • 1
    $\begingroup$ I edited your question to clarify it, but I'm not 100% sure I got it right. If I changed what you meant to ask, please edit it again to fix it. By the way, we'd really rather you not delete your question and post a new one. This time it's fine, but in the future you should edit the old question instead. $\endgroup$ – David Z Nov 26 '14 at 8:12
  • $\begingroup$ Possible duplicate by OP: physics.stackexchange.com/q/148384/2451 $\endgroup$ – Qmechanic Nov 26 '14 at 9:05
  • $\begingroup$ Thanks for the edit. I apologize for the mess with the previous post. $\endgroup$ – sol0invictus Nov 26 '14 at 9:20
2
$\begingroup$

It's the chain rule: $$ \partial_x f(y(x)) = \partial_y f(y(x)) \cdot \partial_x y(x)$$ Your vector field $A^\mu$ depends on the worldsheet coordinates only through the worldsheet coordinates $x^\mu$. Thus, when how $A^\mu$ behaves under an infinitesimal shift on the world-sheet you need to take into account how $A^\mu$ depends on $z$ - that's $\partial_z A^\mu$. But you also need to take into account how the worldsheet coordinates change - which is the connection piece.

$$D_z A^\mu = \underbrace{\partial_z A^\mu}_{\text{change in } A^\mu} + \underbrace{\Gamma^\mu_{\rho \sigma}(x) \partial_z x^\rho A^\sigma}_{\text{change of what "} \mu \text{" means.}}$$

$\endgroup$
1
  • $\begingroup$ So if I had a quantity $D_zX^{\mu}(\sigma)$ would it be just partial derivative with no connection. This X is the same x where my previous covariant derivative was taken $\endgroup$ – sol0invictus Nov 26 '14 at 9:20

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.