What were the genuine Banach spaces in 1922? Reflection on axiomatisation and progression of the mathematical thought (Q2297367)

\textit{S.~Banach}'s Thesis [Fundamenta Math. 3, 133--181 (1922; JFM 48.0201.01)] can be regarded as the birth certificate of the theory of complete normed linear spaces. The main point of the paper under review is the observation that about one third of this fundamental work deals with ``genuine'' Banach spaces whose elements are measurable functions on some finite interval. In this special case, Banach requires additional axioms. Unfortunately, the formulation of axiom~VI, which says that the norm is lower semi-continuous with respect to the measure topology, contains a fatal misprint in the rendering by the author. In Banach's classic ``Théorie des Opérations Linéaires'' [Monografie Matematyczne~1. Warszawa: Seminarium Matematyczne Uniwersytetu Warszawskiego; Warszawa: Instytut Matematyczny PAN (1932; Zbl 0005.20901)], the complete metrizable linear spaces $S$ of measurable functions on \([0,1]\) as well as the related axioms are only mentioned in passing; see Introduction, \S\,7, and Chapter~V, \S\,5. It would be helpful to describe the role of axioms IV, V, and VI in the modern theory of function spaces. This article is a mixture of mathematics and philosophy (epistemology). The author profoundly speculates about Banach's understanding of the interplay between the different axioms. The real truth will remain unknown forever.