scientific article; zbMATH DE number 3966063
An intrinsic feature of Cantor's set theory, later axiomatized by Zermelo and Fraenkel, is its incompleteness. It cannot answer many questions about set - for example, whether there are Lebesgue-nonmeasurable sets of real numbers of whether the real continuum can be well-ordered, and if so, where on the scale of alephs its cardinality lies. In order to organize the realm of sets in a more complete way new principles have been invented, among them E. Zermelo's axiom of choice (1904), abbreviated as AC, and H. Steinhaus and J. Mycielski's axiom of determinateness (1962), abbreviated as AD. The book under review provides an introduction to those chapters of set theory which depend on either of the two principles, discusses some of their most remarkable consequences, and reports on their mutual relationship. The first part of the book is devoted to AC, the statement that every family of nonempty sets has a choice function. Of the theorems which depend on AC the author proves Zermelo's well-ordering theorem and Zorn's lemma. A corollary to the first theorem, that any two cardinals are comparable, warrants the conclusion that AC considerably simplifies the structure of cardinal numbers. In this context it is also mentioned that AC does not solve Cantor's celebrated continuum problem. As a corollary to Zorn's lemma the author proves what he calls Tarski's theorem that every proper filter extends to an ultrafilter. In this part a weaker version of AC, that every denumerable family of nonempty sets has a choice function, denoted as \(AC_{\omega}\), is also discussed. That \(AC_{\omega}\) suffices in many cases is illustrated by its use in the proofs of the following theorems: (i) Denumerable unions of denumerable sets are denumerable; (ii) if x is a limit point of a set X of real numbers, then there are sequences with members in X which converge to x; and (iii) every infinite set contains a denumerable subset. Turning to paradoxical consequences of AC the author constructs (iv) a Lebesgue-nonmeasurable set of real numbers, due to G. Vitali; (v) a nondenumerable set with no perfect subsets (i.e. nonempty closed subsets with no isolated points), due to N. Luzin; and (vi) he also proves a theorem, due to S. Ulam and A. Tarski, that every (Ulam-) measurable cardinal is strongly inaccessible. In this context the author mentions P. Cohen's independence result. No use of the axiom of choice is required in the study of natural numbers, denumerable groups, algorithms, programmes and similar objects. The theory of real numbers, real functions and mathematical analysis actually depends only on \(AC_{\omega}\). In its full extent AC is necessary only in the most abstract regions of set theory, topology, universal algebra and similar abstract theories. As the human mind is unable to perform infinitely many choices in the absence of an effective choice-making rule, AC must remain a noneffective principle of mathematical thinking. The second part of the book centers on the determinateness axiom, AD, according to which every set of infinite sequences of natural numbers is determinate in the sense that it determines a two-person game with a winning strategy for one of the players. As it stands, AD does not look very simple and intuitive. However, using expressions preceded by infinitely many quantifiers, AD can be given the following, more intuitive form of a generalized excluded middle principle: \[ (vii)\quad \exists \alpha_ 0\forall \alpha_ 1\exists \alpha_ 2\forall \alpha_ 3...((\alpha_ 0,\alpha_ 1,\alpha_ 2,\alpha_ 3,...)\in A)\vee \] \[ \forall \alpha_ 0\exists \alpha_ 1\forall \alpha_ 2\exists \alpha_ 3...((\alpha_ 0,\alpha_ 1,\alpha_ 2,\alpha_ 3,...)\not\in A), \] where A is a family of infinite sequences of natural numbers, and where the first [the second] disjunct of (vii) amounts to saying that in the game determined by A there exists a winning strategy for the first [the second] player. Many determinate sets can easily be constructed. However, AC implies the existence of indeterminate sets, no indeterminate sets have been found in the theory of sets without AC, one may consider AD as a rival to AC even though it has not as yet been proved that indeterminate sets cannot be found in the theory of sets without AC. The importance of AD is strengthened by the fact that it entails \(AC_{\omega}\) (the proof of this fact is included). Consequently, AD can be used, instead of AC, in the proofs of many important theorems such as (i)-(iii) above or the theorem on \(\sigma\)-additivity of the Lebesgue measure. AD entails a rich crop of important set-theoretic results. Among those proved in the book there are: (viii) Every nondenumerable set of real numbers contains a perfect subset, due to M. Davis; and (ix) every set of real numbers is Lebesgue measurable, due to J. Mycielski and S. Swierczkowski. Davis' result immediately entails that every nondenumerable set of real numbers has the cardinality of the continuum; thus AD provides a positive solution to Cantor's continuum problem. It follows that, unlike AC, AD makes the properties of the real numbers extremely regular. A different picture also results from AD on the matter of ultrafilters. A proof is provided in the book for the nonexistence of nonprincipal ultrafilters over the set of natural numbers. This negative result is followed by the proof of a positive result, due to R. M. Solovay, that \(\aleph_ 1\) is (Ulam-) measurable. Thus, AD not only eliminates paradoxical sets but also leads to some set-theoretic constructions which are out of reach for AC. As things are, it remains only to investigate the impact of AD on topology, something which is still largely unknown.
- Shadows of the axiom of choice in the universe \(L(\mathbb {R})\)
- \(AD+DC_{\aleph_ 1}\) is incompatible
- The Cantor function
- scientific article; zbMATH DE number 4014700 (Why is no real title available?)
- On arbitrary sets and ZFC
- scientific article; zbMATH DE number 3877932 (Why is no real title available?)
- The axiom of choice
- scientific article; zbMATH DE number 3904538 (Why is no real title available?)
- The axiom of determinateness implies ω_2 has precisely two countably complete, uniform, weakly normal ultrafilters
- The axiom of determinacy implies dependent choices in L(R)
- Axioms of symmetry: Throwing darts at the real number line
- Independence, randomness and the axiom of choice
- scientific article; zbMATH DE number 1201510 (Why is no real title available?)
- scientific article; zbMATH DE number 1302487 (Why is no real title available?)
- An infinity which depends on the axiom of choice
- Two Classical Surprises Concerning the Axiom of Choice and the Continuum Hypothesis
- Choice principles for special subsets of the real line
- scientific article; zbMATH DE number 195042 (Why is no real title available?)
- Canonical models for fragments of the axiom of choice
- The axiom of choice and inference to the best explanation
- Properties of the real line and weak forms of the Axiom of Choice
- Axiom of choice
- A nonconservativity result on global choice
- Dynamical conflict models. I. Language of modeling
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3734396)