In DoCarmo’s Riemannian Geometry book, an affine connection $\nabla$ on a smooth manifold $M$ is defined to be a mapping :$\Xi(M)\times \Xi(M)\rightarrow \Xi(M)$, where $\Xi(M)$ is the set of all smooth vector fields on $M$.
Now DoCarmo wants to clarify this definition with a theorem containing the following fact: ... c) if $V$ is a vector field along the curve $c$, and if $V$ is induced by a vector field $Y\in \Xi(M)$, then the covariant derivative of $V$ along $c$ denoted by $\frac{DV}{dt}$ equals $\nabla(\frac{dc}{dt},Y)$, where $\frac{dc}{dt}$ is the velocity field of the curve $c$.
But the domain of $\nabla$ does not include $(\frac{dc}{dt},Y)$, since $\frac{dc}{dt}$ does not belong to $\Xi(M)$. What is the justification?