In my course we have just derived the momentum operator in the position basis to be the first order-order derivative of the position. This was achieved by using a first order-order taylor expansion:
$$ \psi(\boldsymbol r-d\boldsymbol r)\approx\psi(\boldsymbol r)⁻\nabla\psi(\boldsymbol r)\cdot d\boldsymbol r $$$$ \psi(\boldsymbol r-d\boldsymbol r)\approx\psi(\boldsymbol r)-\nabla\psi(\boldsymbol r)\cdot d\boldsymbol r $$
The same derivation can also be found on wikipedia.
Now my question is whether the final result, $\boldsymbol p = -i\hbar\nabla$, is an approximation? What if we were to use a second order term in the taylor expansion, would that change anything?