As I can always express an local point on a manifold to be local inertial frame, the Christoffel symbol is always zero at that particular point. In Carrol's Spacetime and Geometry, Eq (3.128) shows the derivation of Riemann tensor in terms of metric at a local inertial frame where the Christoffel symbol vanishes. However, why its derivatives do not vanish? Derivative of zero is still zero.
-
4$\begingroup$ $x$ is zero at $x=0$ but its deivative is not not zero there. Same with $\Gamma$. $\endgroup$– mike stoneCommented Mar 6, 2021 at 17:51
-
$\begingroup$ Just because a function vanishes at a point doesn't mean the derivative vanishes. eg: $f(x)=x$, or $f(x)=\sin x$ or $f(x)=\tan x$ or $f(x,y)=(x,y)$ all vanish at the origin, but none of their derivatives vanish. Derivative depend on the values of the function in an open neighbourhood of the point. If the function is identically zero in an open neighbourhood, then all the derivatives will vanish identically in that open neighbourhood. $\endgroup$– peek-a-booCommented Mar 6, 2021 at 17:52
1 Answer
The existence of Riemann normal coordinates means one can find coordinates at a point $p$ such that the metric can be written as $$ g_{\mu \nu}(p) = \eta_{\mu \nu} \quad , \quad g_{\mu \nu , \lambda}(p)= 0 \ , $$ so we also have $\Gamma^{\lambda}_{\mu \nu}(p) = 0$. This does not mean that $g_{\mu \nu , \lambda \sigma}(p)= 0$, so in general $\partial_\rho \Gamma^{\lambda}_{\mu \nu}(p)\neq 0$.
As the other comments say, a function can be $0$ at a point but have non-zero derivative; similarly, its derivative vanishing at a point doesn't mean its second derivative is nonzero, etc.
See section 3.3.1 of these lecture notes for details of how the existence of normal coordinates can be shown or for more details in general.
-
$\begingroup$ I thought the symbol is just coefficient of basis vectors and treated that as a scalar instead of a function. Thanks everyone for answering my question. $\endgroup$– Simon219Commented Mar 6, 2021 at 18:13
-
$\begingroup$ It is a coefficient, but the vectors change from point to point. $\endgroup$– PedroCommented Mar 6, 2021 at 18:47
-
$\begingroup$ All these quantities are fields defined at each spacetime point $p$. $\endgroup$– EletieCommented Mar 6, 2021 at 18:48