Can you recover the manifold from the topological vector space of distributions?