Manighetti, M. (2016). Computational interpretations of Markov’s principle [Diploma Thesis, Technische Universität Wien]. reposiTUm. https://doi.org/10.34726/hss.2016.37906
Markov's principle is a statement that originated in the Russian school of Constructive Mathematics and stated originally that if it is impossible that an algorithm does not terminate, then it will terminate. This principle has been adapted to many different contexts, and in particular we are interested in its most common version for arithmetic, which can be stated as given a total recursive function f, if it is impossible that there is no n for which f(n) = 0, then there exists an n such that f(n) = 0. This is in general not accepted in constructivism, where stating an existential statement requires one to be able to show at request a witness for the statement: here there is no clear way to choose such an n. We introduce more in detail the context of constructive mathematics from different points of view, and we show how they are related to Markov's principle. In particular, several realizability semantics are presented, which provide interpretations of logical systems by means of different computational concepts (mainly, recursive functions and lambda calculi). This field of research gave origin to the well known paradigm often called Curry-Howrd isomorphism, or also propositions as types, that states a correspondence between proofs in logic and programs in computer science. Thanks to this the field of proof theory, that is the metamathematical investigations of proofs as mathematical objects, became of interest for computer science and in particular for the study of programming languages. By using modern research on the Curry-Howard isomorphism, we will obtain a more refined interpretation of Markov's principle. We will then use this results to investigate the logical properties of systems related to the principle, and introduce a proof transformation technique to interpret constructively some non-constructive proofs of arithmetic.
