Approaches to analysis with infinitesimals following Robinson, Nelson, and others (Q1750317)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Approaches to analysis with infinitesimals following Robinson, Nelson, and others |
scientific article |
Statements
Approaches to analysis with infinitesimals following Robinson, Nelson, and others (English)
0 references
18 May 2018
0 references
The paper under review is an extensive survey (172 bibliographic entries) of what is now known as ``nonstandard'' analysis (NSA), which allows for infinitesimals and infinitely large numbers. As the authors claim, most mathematicians adhere to what they call the ``CD + II mindset'': If in a model of ZF the set \(\mathbb{N}\) of natural numbers is defined as the smallest set which contains 1 and being closed under induction, then the set of naive or counting numbers is not necessarily equal to \(\mathbb{N}\). A fact attributed already to \textit{T. Skolem} (1933). Thus it is consistent to have infinite elements in \(\mathbb{N}\). Of course, the naive numbers are then not a set within our model of ZF. Reeb's ``Claim \(Q\)'' states this fact as: ``The naive integers don't fill up \(\mathbb{N}\).'' Nowadays, models of NSA are obtained via higher-order ultra-products \({}^*\mathbb{R}\) of a standard model \(\mathbb{R}\) of the reals. Most importantly, \({}^*\mathbb{R}\) satisfies the transfer principle. E.g. rules, laws, procedures valid for \(\mathbb{R}R\) are transferred to \({}^*\mathbb{R}\). E.g. the function \(\sin: \mathbb{R}\to \mathbb{R}\) is extened to \({}^*\sin: {}^*\mathbb{R}\to {}^*\mathbb{R}\) and familiar laws, like \(\sin(n\pi)=0\) hold for infinite \(n\in\mathbb{N}\). An axiomatic approach to NSA was proposed by \textit{E. Nelson} in his book on [Radically elementary probability theory. Princeton, NJ: Princeton University Press (1987; Zbl 0651.60001)]. Besides the membership relation ``\(\varepsilon\)'' a unary predicate ``\(\operatorname{st}(x)\), the set \(x\) is standard'' is added. Nelson added to the axioms of ZFC three axioms IST, and called his system ``internal set theory''. I stands for ``Idealization'', S for ``Standardization'' and T for ``Transfer''. The authors state and discuss these axioms in detail for a ``bounded'' variant of IST In BST ``every set is an element of some standard set'' while in IST ``there is a set that contains all standard sets as elements''. In IST or BST, the set \(\mathbb{R}\) of real numbers is per se nonstandard, infinitesimals are already there. This makes IST especially attractive. The authors discuss ``The hyperfinite vibrating string'' and ``Hyperfinite Brownian motion in this setting''. The ultrapower construction makes NSA a highly nonconstructive branch of mathematics. The authors explain how Browers argument for rejecting ``\(A\) or (not(\(A\)))'' may lead naturally to NSA. Here \(A\) is an as-yet unproven conjecture which has been verified for any large number. But then (not(\(A\))), if it exists, cannot be constructed in the physical world. ``Browers intuitionism motivates the existence of non-standard numbers especially in IST.'' The article contains lots of further material related to NSA and leaves no stone untouched. The reviewer highly recommends the article to anybody who wants to get an overview about the foundations of NSA.
0 references
axiomatisations
0 references
infinitesimal
0 references
nonstandard analysis
0 references
ultraproducts
0 references
superstructure
0 references
set-theoretic foundations
0 references
multiverse
0 references
naive integers
0 references
intuitionism
0 references
soritical properties
0 references
ideal elements
0 references
protozoa
0 references