1: \begin{abstract}

2: We prove that solution operators of elliptic obstacle-type variational inequalities 

3: (or, more generally, locally Lipschitz continuous functions possessing certain pointwise-a.e.\ convexity properties)

4: are Newton differentiable when considered as maps between suitable Lebesgue spaces

5: and equipped with the strong-weak Bouligand differential 

6: as a generalized set-valued derivative. It is shown that this Newton differentiability allows

7: to solve optimal control problems with $H^1$-cost terms and one-sided pointwise control constraints by 

8: means of a semismooth Newton method. The superlinear convergence of the resulting algorithm 

9: is proved in the infinite-dimensional setting and its mesh independence 

10: is demonstrated in numerical experiments. We expect that the findings of this paper 

11: are also helpful for the design of numerical solution procedures for quasi-variational inequalities 

12: and the optimal control of obstacle-type variational problems. 

13: 

14: \end{abstract}