What are mathematical theories? What mathematical objects should correspond to this informal concept? The classical and most important answer to these questions is: Theories formalized in first order logic. But this answer has also some undesirable features. One of theme is the dependence of such theories upon the language or the choice of primitive concepts, whereas a slightly deeper view would identify theories interpretable in each other.

The purpose of the present memoir is to investigate further, to survey the former work and to point out a number of open problems about local interpretability.