×

Theory of linear operators. (Théorie des opérations linéaires.) (French) Zbl 0005.20901

Monografie Matematyczne 1. Warszawa: Seminarium Matematyczne Uniwersytetu Warszawskiego; Warszawa: Instytut Matematyczny PAN. vii, 254 S. (1932).
There are two zbMATH reviews for this item, first A.N. Kolmogorov’s review from 1933 and then A. Pietsch’s “Looking Back” review, written on the occasion of Banach’s 125th anniversary in 2017. For the Jahrbuch review by G. Aumann see [JFM 58.0420.01].
First review by A.N. Kolmogorov (Moscow) (1933):
Ein Buch, welches zum erstenmal die Theorie der linearen Räume und der linearen Operatoren in allgemeinen linearen Räumen vollständig, und deshalb nicht weniger klar und durchsichtig, darstellt. Außer dem bekannten Materiale enthält das Buch mehrere früher nicht veröffentlichte Resultate der neueren Untersuchungen des Verf.und seiner Mitarbeiter, in erster Linie von S. Mazur; über diese neuen Ergebnisse wird weiter ausführlicher referiert.
Einleitung enthält eine kurze Darstellung einiger für weitere Darstellung notwendigen Eigenschaften der Lebesgueschen und Stieltjesschen Integrale und der metrischen Räume (D-Räume); hier findet man auch mehrere Beispiele der linearen Räume. Kap. I, welches übrigens nur 5 Seiten enthält, ist den metrischen Gruppen gewidmet. Im Kap. II sind die allgemeinen Vektorräume wie die kommutativen Gruppen mit reellen Zahlen als Operatoren definiert; keine weiteren topologischen oder metrischen Eigenschaften des Raumes sind hier vorausgesetzt, diese allgemeine Theorie gestattet jedoch einige interessante Anwendungen. Wenn man in einem allgemeinen Vektorraume eine Metrik einführt und die Stetigkeit der Addition und der Multiplikation fordert, erhält man einen metrischen Vektorraum; im Kap. III betrachtet Verf.eine speziellere Klasse von F-Räumen, zu dieser Klasse gehören einige für die Anwendungen wichtige Räume, wie z. B. der Raum \((s)\) aller Zahlenfolgen und der Raum \((S)\) aller meßbaren Funktionen.
Im Zentrum der weiteren Darstellung steht aber der Begriff des normierten Raumes (Kap. IV); nur für diese Räume sind die wichtigsten Begriffe der Funktionale- und Operatorentheorie erklärt und untersucht. Die vollständigen normierten Räume (B-Räume) bilden den Gegenstand des Kap. V; Kap. VI enthält die Theorie der totalstetigen linearen Operatoren und der konjugierten Operatoren. Nach der Definition des Kap. VII sind eine Folge von Elementen \((x_i)\) und eine Folge von Funktionalen \((f_i)\) zu einander biorthogonal, wenn \(f_i(x_j)\) gleich \(0\) oder \(1\) ist, je nachdem \(i\neq j\) oder \(i=j\) gilt. Im Kap. VIII betrachtet Verf.die Begriffe der regulären Geschlossenheit, der schwachen Geschlossenheit der Funktionalemengen und der schwachen Konvergenz der Funktionale; die schwache Konvergenz der Elemente ist im Kap. IX untersucht. Im Kap. X sind die allgemeinen Grundlagen der Theorie der linearen Funktionalgleichungen dargestellt.
Die letzten Kap. XI und XII enthalten fast ausschließlich neue Resultate. Verf.nennt zwei lineare Räume äquivalent, wenn sie eineindeutig auf einander abgebildet mit der Erhaltung der Norm und der Vektorrelationen sein können; jeder separable B-Raum ist einem Unterraume des Raumes \((C)\) aller stetigen Funktionen äquivalent; mehrere andere Sätze über die Äquivalenz der Räume können wir nicht alle anführen. Wenn die Abbildung stetig ist, ohne notwendig die Norm zu erhalten, heißen zwei Räume isomorph; auf den Begriff der Isomorphie begründet Verf. seine Theorie der linearen Dimension (Kap. XII). Die dimensionalen Verhältnisse verschiedener bekannter Räume sind nur teilweise erklärt; es ergibt sich z. B., daß \(\dim(L^{(2)})<\dim(L^{(p)})\) ist, wenn nur \(1 < p \neq 2\) gilt (dabei bezeichnet man durch \(L^{(p)}\) den Funktionalraum mit \((\int |f|^pdx)^{1/p}\) als Norm). Dem Haupttext folgen mehrere “Bemerkungen”, sie enthalten interessante Verzeichnisse der ungelösten Probleme.


“Looking Back” review by A. Pietsch (Jena) (written on the occasion of Banach’s 125th anniversary in 2017):
The theory of Banach spaces is an important part of functional analysis. Its original purpose was to provide abstract tools for solving integral equations. In the course of time the range of applications grew rapidly; see the chart on p. 550 of [A. Pietsch, History of Banach spaces and linear operators. Boston: Birkhäuser (2007; Zbl 1121.46002)]. Based on preliminary work of D. Hilbert, M. Fréchet, F. Riesz, E. Helly and others, Stefan Banach prepared the ground in his thesis [S. Banach, “Sur les opérations dans les ensembles abstraits et leur application aux équations intégrales”, Fundam. Math. 3, 133–181 (1922; JFM 48.0201.01)]. At the suggestion of M. Fréchet [Les espaces abstraits. Paris: Gauthier-Villars (1928; JFM 54.0614.02), p. 141] a complete normed linear space is now referred to as a Banach space (originally: espace de M. Banach, where M. stands for Monsieur), a terminology that Banach turned into espace du type (B).
Banach’s Théorie des opérations linéaires appeared as the first volume in the famous series monografie matematyczne, Warszawa 1932. Basically, it is a French translation of [Teoria operacji, Tom I, Operacje liniowe. Warszawa (1931)]. There are, however, several improvements. For example, Chapitre XII, with the title “Dimension linéaire” was added, and we have a substantially extended version of the valuable Remarques written in collaboration with Stanisław Mazur.
This excellent book has become a milestone in the history of mathematics. In the opinion of J. Dieudonné [History of functional analysis. Amsterdam etc.: North-Holland Publishing Company (1981; Zbl 0478.46001), pp. 142–143] “it had on Functional Analysis the same impact that van der Waerden’s book had on Algebra two years earlier.”
The following quotation from [Éléments d’histoire des mathèmatiques. Paris: Hermann (1960; Zbl 0129.24508), p. 217] shows N. Bourbaki’s high esteem for Banach’s work:
“La publication du traité de Banach sur les “Opérations linéaires”, marque, pourrait-on dire, le début de l’âge adulte pour la théorie des espaces normés. Tous les résultats dont nous venons de parler, ainsi que beaucoup d’autres, se trouvent exposés dans ce volume, de façon encore un peu désordonnée, mais accompagnés de multiples exemples frappants tirés de domaines variés de l’Analyse. \(\;\dots\;\) De fait, l’ouvrage eut un succès considérable, et un de ses effets les plus immédiats fut l’adoption quasi-universelle du langage et des notations utilisés par Banach.”
This monograph is written in a modern style; large parts could be taken as a basis for today’s lectures. Of course, it contains the cornerstones of Banach space theory: the (real) Hahn-Banach extension theorem (Chap. II, §2), the open mapping, bounded inverse, and closed graph theorem (Chap. III, §3) as well as the principle of uniform boundedness (Chap. V, §1). A more detailed description of the contents can be found in A.N. Kolmogorov’s Zentralblatt review from 1932 above.
However, due to the state of the art, two fundamental features are missing:
(1) Though N. Wiener [“Note on a paper of M. Banach”, Fundam. Math. 4, 136–143 (1923; JFM 49.0213.01)] had stressed the importance of the complex case, Banach treats only real spaces. Hence, even in spectral theory (Chap. X, § 3), the parameter \(h\) in \(x - h U(x)\) is restricted to \(\mathbb R\). We also mention that, in his famous work [“Über lineare Funktionalgleichungen”, Acta Math. 41, 71–98 (1918; JFM 46.0428.02), imprimé le 5 décembre 1916], F. Riesz consistently used complex scalars. Already in his book [Les systèmes d’équations linéaires à une infinité d’inconnues. Paris: Gauthier-Villars (1913; JFM 44.0401.01), p. 117] he observed that “la substitution \(A_\zeta = A( I - \zeta A)^{-1} \) montre le caractère d’une fonction holomorphe en \(\zeta\).”
H. Heuser [Funktionalanalysis. 2nd edition. Stuttgart: Teubner (1986; Zbl 0653.46002)] wrote in his chapter “Ein Blick auf die werdende Funktionalanalysis”, p. 660:
“Banach findet nie aus den reellen B-Räumen heraus, ungeachtet der Rieszschen und Wienerschen Fingerzeige, und läßt so ohne Not ein fruchtbares Feld unbeackert liegen.”
Maybe, Banach would have answered, ‘Sorry, I had no complex Hahn-(B) theorem at my disposal’. Indeed, using the concept of opérations conjugées (associées) (Chap. VI, § 3), Banach was able to present the Riesz theory of compact operators (originally: vollstetig) in a dualized form (Chap. X, § 2, Footnote 1). Curiously enough, only one half of Schauder’s theorem [J. Schauder, “Über lineare, vollstetige Funktionaloperationen”, Stud. Math. 2, 183–196 (1930; JFM 56.0354.01)] is included (Chap. VI, Théorème 4): If an operator is compact, then so is its dual (but the converse is missing).
(2) Another serious gap concerns weak topologies. Though the basic tools from general topology were available around 1930, Banach worked only with (transfinite) sequences, which now look quite artificial. He had a good companion. In the preface to the revised version of his [Mengenlehre. Berlin: De Gruyter (1927; JFM 53.0169.01)] F. Hausdorff wrote:
“Es wird vielleicht bedauert werden, daß ich zu weiterer Raumersparnis in der Punktmengenlehre den topologischen Standpunkt, durch den sich die erste Auflage anscheinend viele Freunde erworben hat, aufgegeben und mich auf die einfachere Theorie der metrischen Räume beschränkt habe.”
In this spirit, a preliminary version of the later Alaoglu theorem is given only for separable Banach spaces; in the separable case the closed unit ball of the dual is metrizable in its weak topology, and toute suite de fonctionnelles linéaires dont l’ensemble des norms est borné contient une suite partielle faiblement convergente (Chap. VIII, Théorème 3). Nowadays, the general case is treated with the help of ultrafilters. In retrospect, the “birth of weak topologies” was a tough and lengthy process.
Many concrete Banach spaces are equipped with additional structures: partial orders and multiplications. Hence Banach lattices and Banach algebras became of interest. In my opinion, Banach’s book had no significant impact on these theories, which were independently developed by different schools: L.V. Kantorovich, G. Birkhoff, J. von Neumann, I.M. Gelfand, and C.E. Rickart, to name just a few.
The golden age of Banach space theory came to an abrupt end by the German occupation in 1939. Many Polish mathematicians were murdered in the following years. After the war, the needs of distribution theory stimulated the study of locally convex spaces, in particular nuclear spaces. So, for a while, Banach spaces were considered just as a special subclass in this general setting.
Luckily, the situation changed significantly by the efforts of succeeding generations. Here are a few keywords of modern Banach space theory: Dvoretzky’s theorem on spherical sections, Grothendieck’s inequality on absolutely summing operators, type and cotype, Radon-Nikodym property, \(s\)-numbers, interpolation theory, etc. A. Pełczyński’s plenary lecture [Proc. Int. Congr. Math., Warszawa 1983, Vol. 1, 237–269 (1984; Zbl 0575.46010)] provides an excellent impression.
However, step by step, an undesirable tendency could be observed: Many important problems were answered in the negative by the construction of sophisticated counterexamples, which show that the concept of a Banach space has its limits.
Supplementary literature: [R. Kałuża, Through a reporter’s eyes: the life of Stefan Banach. Basel: Birkhäuser (1996; Zbl 0849.01019); E. Jakimowicz (ed.) and A. Miranowicz (ed.), Stefan Banach: remarkable life, brilliant mathematics. Gdańsk: Gdańsk University Press; Providence, RI: American Mathematical Society (AMS) (2011; Zbl 1236.01023); K. Kuratowski, A half century of Polish mathematics. Oxford etc.: Pergamon Press; Warszawa: PWN – Polish Scientific Publishers (1980; Zbl 0438.01006); R. Duda, Pearls from a lost city. The Lvov School of Mathematics. Providence, RI: American Mathematical Society (AMS) (2014; Zbl 1303.01001); R. D. Mauldin (ed.), The Scottish Book. Mathematics from the Scottish Café, Boston etc.: Birkhäuser (1981; Zbl 0485.01013); 2nd edition (2015; Zbl 1331.01039); A. Pietsch, History of Banach spaces and linear operators, loc. cit.].
Banach’s monograph was reprinted by Chelsea in 1955 [Zbl 0067.08902] and in Vol. 2 of Banach’s Œuvres in 1979 [Zbl 0407.01009]; today it is freely accessible on the EuDML platform (see the link below). It was also translated into Ukrainian (Kiev, 1948), English [Amsterdam etc.: North-Holland (1987; Zbl 0613.46001)] and Russian (Moscow, 2001).

MSC:

46-02 Research exposition (monographs, survey articles) pertaining to functional analysis
Full Text: EuDML