Gödel's Dialectica interpretation is a fundamental tool for the extraction of computational content from proofs, and plays a central role in today's proof mining program. In the past decades, it has also been studied from the perspective of programming languages, and our contribution is in that direction. Specifically, we present Dialectica as a collection of rules in the style of Hoare logic, where Dialectica is now viewed as a language for specifying procedural programs that come with a forward and backward direction. This viewpoint captures the interesting dynamics of realisers extracted by the Dialectica interpretation, and we illustrate this by defining a generalised backpropagation semantics for a fragment of this language. We envisage this work as providing a base for several future developments, both theoretical and practical, which we outline at the end.

Motivated by recent applications of proof theory in probability, we introduce a novel computational interpretation of probabilistic $$\exists \forall $$ -formulas, called dependent learnability. This encompasses several important notions of quantitative stochastic convergence, where it represents a generalized version of the property – widely studied in probability and ergodic theory – that a sequence of random variables has bounded fluctuations. We study both deterministic and stochastic variants of this notion and relate these to other computational interpretations of $$\exists \forall $$ -formulas from the literature. In particular, we prove dependent learnability to be primitive recursively equivalent to the influential notion of metastability, which in conjunction with results from applied proof theory highlights that dependently learnable rates can be extracted from large classes of nonconstructive proofs of $$\exists \forall $$ -formulas. Furthermore, we present a primitive recursive algorithm for joining two (and thus finitely many) dependently learnable rates, which in particular proves to be considerably more mathematically intuitive than the corresponding functional for joining rates of metastability. Finally, we discuss our results in the light of game semantics.

We introduce an extension of first-order logic that comes equipped with additional predicates for reasoning about an abstract state. Sequents in the logic comprise a main formula together with pre- and postconditions in the style of Hoare logic, and the axioms and rules of the logic ensure that the assertions about the state compose in the correct way. The main result of the paper is a realizability interpretation of our logic that extracts programs into a mixed functional/imperative language. All programs expressible in this language act on the state in a sequential manner, and we make this intuition precise by interpreting them in a semantic metatheory using the state monad. Our basic framework is very general, and our intention is that it can be instantiated and extended in a variety of different ways. We outline in detail one such extension: A monadic version of Heyting arithmetic with a wellfounded while rule, and conclude by outlining several other directions for future work.

In this paper we study Littlewood's Tauberian theorem from a proof theoretic perspective. We first use the Dialectica interpretation to produce an equivalent, finitary formulation of the theorem, and then carry out an analysis of Wielandt's proof to extract concrete witnessing terms. We argue that our finitization can be viewed as a generalized Tauberian remainder theorem, and we instantiate it to produce two concrete remainder theorems as a corollary, in terms of rates of convergence and rates metastability, respectively. We rederive the standard remainder estimate for Littlewood's theorem as a special case of the former.

We give a computational interpretation to an abstract instance of Zorn's lemma formulated as a wellfoundedness principle in the language of arithmetic in all finite types. This is achieved through Gödel's functional interpretation, and requires the introduction of a novel form of recursion over non-wellfounded partial orders whose existence in the model of total continuous functionals is proven using domain theoretic techniques. We show that a realizer for the functional interpretation of open induction over the lexicographic ordering on sequences follows as a simple application of our main results.

Abstract During the last 20 years or so, a wide range of realizability interpretations of classical analysis have been developed. In many cases, these are achieved by extending the base interpreting system of primitive recursive functionals with some form of bar recursion, which realizes the negative translation of either countable or countable dependent choice. In this work, we present the many variants of bar recursion used in this context as instantiations of a parametrized form of backward recursion, and give a uniform proof that under certain conditions this recursor realizes a corresponding family of parametrized dependent choice principles. From this proof, the soundness of most of the existing bar recursive realizability interpretations of choice, including those based on the Berardi–Bezem–Coquand functional, modified realizability and the more recent products of selection functions of Escardó and Oliva, follows as a simple corollary. We achieve not only a uniform framework in which familiar realizability interpretations of choice can be compared, but show that these represent just simple instances of a large family of potential interpretations of dependent choice principles.

Abstract We carry out a proof-theoretic analysis of the wellfoundedness of recursive path orders in an abstract setting. We outline a general termination principle and extract from its wellfoundedness proof subrecursive bounds on the size of derivation trees that can be defined in Gödel’s system T plus bar recursion. We then carry out a complexity analysis of these terms and demonstrate how this can be applied to bound the derivational height of term rewrite systems.

We present a new variant of Gödel's functional interpretation in which extracted programs, rather than being pure terms of system T, interact with a global state. The purpose of the state is to store relevant information about the underlying mathematical environment. Because the validity of extracted programs can depend on the validity of the state, this offers us an alternative way of dealing with the contraction problem. Furthermore, this new formulation of the functional interpretation gives us a clear semantic insight into the computational content of proofs, and provides us with a way of improving the efficiency of extracted programs.

In this article we study Gödel's functional interpretation from the perspective of learning. We define the notion of a learning algorithm, and show that intuitive realizers of the functional interpretation of both induction and various comprehension schemas can be given in terms of these algorithms. In the case of arithmetical comprehension, we clarify how our learning realizers compare to those obtained traditionally using bar recursion, demonstrating that bar recursive interpretations of comprehension correspond to 'forgetful' learning algorithms. The main purpose of this work is to gain a deeper insight into the semantics of programs extracted using the functional interpretation. However, in doing so we also aim to better understand how it relates to other interpretations of classical logic for which the notion of learning is inbuilt, such as Hilbert's epsilon calculus or the more recent learning-based realizability interpretations of Aschieri and Berardi.

We use Gödel's dialectica interpretation to produce a computational version of the well-known proof of Ramsey's theorem by Erdős and Rado. Our proof makes use of the product of selection functions, which forms an intuitive alternative to Spector's bar recursion when interpreting proofs in analysis. This case study is another instance of the application of proof theoretic techniques in mathematics.