Harvey Friedman's research on the foundations of mathematics (Q1073004)

As might be expected from the crass errors concerning - almost every aspect of - foundations in the introduction, this very uneven collection of essays conveys a wholly inadequate impression of the (actually exceptional) quality and interest of H. Friedman's work. It started about 20 years ago, after the 20 preceding years since the end of World War II had established new aims for foundations (and mathematical logic generally) in the light of prewar research; such as mathematical significance and informal rigour (at two extremes), illustrated by simple, though usually convincing examples. Particularly in model theory (Malcev, A. Robinson etc.) and recursion theory (Higman) relations were established with main stream mathematics, and not only corners like Borel sets. Mainly outside those areas Friedman has pursued the early, simpler examples with such imagination and determination that any shortcomings in the results can be attributed to the choices of aims themselves; not to clumsy or half-hearted execution. Thus his results can also be used to examine foundational idea(l)s, some of which are very dubious. For example, several of the more detailed results, involving the number of symbols in formal proofs, are artifacts w.r.t. the epistemological \((=cognitive)\) phenomena suggested by the terminology, since the formalizations considered are not even remotely faithful enough for such uses. The same applies with knobs on to - popular interpretations of - reverse mathematics, illustrated originally by the (flashy) relation between mathematical \(\epsilon_ 0\)-induction and logical reflection principles for formal arithmetic. The longer the lists of such relations, as on pp. 147-148, the more compelling the (philosophical) question: Where, if anywhere, do they contribute to effective knowledge? - Readers are particularly warned against the first and the last essay. Samples. On p. 1, Friedman's work is assigned to 'the' tradition of the unlikely trio: Frege, Russell and Gödel. Now the first two, with very different ideas of 'exact mathematical definitions' (p. 1, l. 8) never touched metamathematics on principle, while in his most famous work Gödel solved only metamathematical problems (formulated explicitly and made famous by Hilbert to boot). Furthermore Gödel was successful, as he himself stressed in retrospect, without needing new methods; unlike Cohen or, on a comparable level, Friedman; contrary to p. viii, 1. 10 inner models are not 'due' to Gödel, but familiar from Euclidean models of non-Euclidean geometry. On p. 399, the possibility, well known for nearly 40 years, of relating the rate of growth of bounding functions for \(\Pi^ 0_ 2\) theorems and derivability in formal systems (pedantically, together with true \(\Pi^ 0_ 1\) theorems), is presented as a 'development' of Gödel's (incompleteness) theorems. This is totally uninformed or thoughtless, both w.r.t. content and to method of proof; the common element is simply the idea of some kind of incompleteness. The general cluelessness is documented on p. 399, where it is not realized that a proof of (formal) undecidability of Fermat's Last \((\Pi^ 0_ 1)\) Theorem would establish it. As a result the really quite basic significance of the relation with rate of growth is never revealed. It provides a rational means of selecting a formal system by reference to some given \(\Pi^ 0_ 2\) problem (for which one happens to want to know about its bounding function). The problem arises because, by general incompleteness, there is no convincing general criterion for selecting any particular system of arithmetic. [Contents: Preface (p. v); Introduction (pp. vii-xii); Biography of Harvey Friedman (p. xiii); \textit{A. Nerode} and \textit{L. A. Harrington}: The work of Harvey Friedman (pp. 1-10); \textit{L. J. Stanley}: Borel diagonalization and abstract set theory: recent results of Harvey Friedman (pp. 11-86); \textit{S. G. Simpson}: Nonprovability of certain combinatorial properties of finite trees (pp. 87-117); \textit{R. L. Smith}: The consistency strengths of some finite forms of the Higman and Kruskal theorems (pp. 119-136); \textit{S. G. Simpson}: Friedman's research on subsystems of second order arithmetic (pp. 137-159); \textit{C. Steinhorn}: Borel structures for first-order and extended logics (pp. 161-178); \textit{C. Smoryński}: Nonstandard models and related developments (pp. 179- 229); \textit{D. Leivant}: Intuitionistic formal systems (pp. 231-255); \textit{A. Ščedrov}: Intuitionistic set theory (pp. 257-284); \textit{J. C. Shepherdson}: Algorithmic procedures, generalized Turing algorithms, and elementary recursion theory (pp. 285-308); \textit{J. C. Shepherdson}: Computational complexity of real functions (pp. 309-315); \textit{A. J. Kfoury}: The pebble game and logics of programs (pp. 317-329); \textit{R. Statman}: Equality between functionals revisited (pp. 331-338); \textit{R. E. Byerly}: Mathematical aspects of recursive function theory (pp. 339- 352); \textit{C. Smoryński}: ''Big'' news from Archimedes to Friedman (pp. 353-366); \textit{C. Smoryński}: Some rapidly growing functions (pp. 367- 380); \textit{C. Smoryński}: The varieties of arboreal experience (pp. 381-397); \textit{G. Kolata}: Does Gödel's theorem matter to mathematics? (pp. 399-404); Harvey Friedman's publications (pp. 405-408).]