Extension and interpretability (Q2658276)

This paper is focused on a structure of extension and interpretability which consists of finitely axiomatised consistent theories as objects and includes the relations of extension (a partial ordering) and interpretability (a partial pre-ordering). The benefit of this structure is that it enables us to consider theories from the perspective of degrees of interpretability and allows us to abstract away from the rigidity of a fixed signature. Within this structure, a notion of sameness of theories is developed, called \textit{i-bisimilarity} and defined in terms of mutual interpretability and extension. With this structure and notion of sameness established, Visser goes on to consider how properties of interest are captured from the perspective of this structure. For this discussion the notion of ``capital'' property (or relation) is introduced where a property is said to be capital if it is maximal i-bisimulation invariant. It is shown that locally faithful interpretability, local cointerpretability, local tolerance, decidability, and the finite model property are definable over the structure and are all capital properties (or relations). The verifications of these facts are routine and following from the established definitions. There is also a discussion of properties that are not capital, such as faithful interpretability and cointerpretability. This division between capital properties and those that are not capital provide a distinction between ``good'' and ``bad'' properties. From this distinction, it becomes possible to use ``bad'' properties to establish new, ``good'' properties. Overall the paper provides a detailed discussion of how properties are viewed from the perspective of the structure of extension and interpretability. It should also be mentioned that the paper includes an extensive appendix that gives further detail about the concepts that underlie the discussion throughout the paper. For the entire collection see [Zbl 1459.03003].