So if we have a curve $\gamma(t): I \rightarrow M$ and riemannian connection $\triangledown$, i'm trying to calculate $\triangledown_{\gamma'} \gamma'$
Where $\gamma$' is the velocity vector of $\gamma$. Write $\gamma' = v^i \delta_i$, using Einstein notation. Then we have:
$\triangledown_{v^i \delta_i} \gamma'$
$=v^i\triangledown_{\delta_i} \gamma'$ since $\triangledown$ is linear over $C^{\infty}(M)$ in it's lower variable
$=v^i\triangledown_{\delta_i} v^j \delta_j$
$=v^i(\delta_i v^j \delta_j+ v^j \triangledown_{\delta_i} \delta_j$
$=v^i\delta_i v^j \delta_j+ v^iv^j \triangledown_{\delta_i} \delta_j$
$=(v^i\delta_i v^j \delta_j+ v^iv^j \Gamma_{i,j}^k) \delta_k$
This is supposed to equal
$=(\gamma''^k+ v^iv^j \Gamma_{i,j}^k) \delta_k$
where $\gamma''^k$ are the component functions of the acceleration of $\gamma$.
Can somebody please find where I went wrong or how to finish? Thanks!
What's misleading here is that you have $v^i(t)$ and $\partial_i(\alpha(t))$, so you're losing track of $t$ derivatives and the chain rule. (For some reason, you're writing $\delta$ instead of $\partial$.)
I don't like the notation, but they're writing $\gamma''{}^k$ to mean $(v^k)'$. For any function $f$ defined along $\alpha(t)$, we pretend that it is defined on a neighborhood of the curve so that $\partial_if$ makes sense; then the chain rule tells us that $f' = (f\circ\gamma)'(t) = (\partial_i f)v^i$, since $\gamma'(t) = \sum v^i\partial_i$. In your case, this gives $$(v^k)' = v^i\partial_i v^k,$$ which is the term you were trying to get.