Arithmetizing proofs in analysis

From MaRDI portal





General methods of extracting numerical bounds from proofs in classical analysis (negative translation combined with functional interpretation) lead to the order of magnitude which is difficult to explain to a working mathematician. The author describes a supply of general tools which allow to extract much more reasonable bounds from proofs using notoriously non-constructive principles. These tools include the use of constructive \(\varepsilon\)-weakenings of classical principles. Many standard analytical principles (like sequential Borel Lemma, attainment of the maximum value of a continuous function etc.) are provable from WKL, the binary König's Lemma, which cannot even be expressed in ``polynomial systems, since this involves coding functionals for finite sequences. Such principles are treated using a certain schema \(F^-\) which is classically false. However, one can extract polynomial bounds from the proof using \(F^-\). On the other hand, the use of \(F^-\) can be replaced by its \(\varepsilon\)-weakening, and this \(\varepsilon\)-weakening is provable in a system for Kalmar-elementary functions. This shows that the bounds are classically true. A more complicated treatment (producing higher bounds) is needed for the theorems (like convergence of bounded monotone sequences of reals, the Arzela-Ascoli property, etc.) which are proved from instances of the arithmetical comprehension axiom. For the entire collection see [Zbl 0893.00029].




Cited in
(24)








This page was built for publication: Arithmetizing proofs in analysis

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4218526)