Bolzano, Cauchy and the ``New Analysis'' of the early nineteenth century (Q2540258): Difference between revisions

This paper deals with developments in mathematics and in the Académie des Sciences in Paris during the first quarter of the last century, and takes as its principal theme a conjecture concerning the publication in 1821 of Augustin-Louis Cauchy's `Cours d'Analyse'. During my researches into the development of mathematical analysis at this time I became convinced that the well-known anticipation of many of Cauchy's principal ideas in Bernard Bolzano's obscure 1817 Prague pamphlet `Rein analytischer Beweis ...' was not a coincidence, but that Cauchy had read Bolzano's paper, and drew on and greatly extended its principal ideas without acknowledgment. To deal first with the ideas: the common ideas involve the continuity of a function, necessary and sufficient conditions for the convergence of a series, and the use of upper limits as a technique in analysis. There are further direct points of correspondence, but beyond them there is a common style or approach in both works, which is subtly different from the analysis currently being practised, and especially from Cauchy's own earlier writings on the subject. These had been concerned with the use of \(\sqrt{-1}\) in the evaluation of definite integrals and with Fourier integral solutions to partial differential equations, and had clearly given Cauchy stimulus to examine anew the foundations of analysis; but there seems to be no hint of these new foundations in his work before the ``Cours d'Analyse'' appeared. The changes of direction in this book are startling, and mark the beginnings of the so-called ``arithmeticisation'' of analysis. I am not satisfied with this term as a characterisation of the ``new analysis'' of Cauchy and Bolzano, and in my paper I describe an historical tool called ``limit-avoidance'', in terms of which this new analysis is an early form of Weierstrass' analysis with which we are now familiar. The basic ideas and proof-methods are formulated in terms of the variable quantities avoiding limiting values by arbitrarily (and for Cauchy, infinitely) small amounts. Before Bolzano and Cauchy limit-avoidance was mostly absent (although arithmeticisation was used sometimes); after Bolzano and Cauchy it was the foundation stone; and with them the subject of mathematical analysis was really born, as the study of limiting values of expressions over an infinitesimaI continuum by means of limit-avoidance. The reason why Cauchy did not read Bolzano seems to me to lie in the intense atmosphere of rivalry which prevailed in Paris at this time. These rivalries have been little studied by historians: in my paper I describe a few of them which involved Cauchy, Fourier, Poisson and Laplace. I think that Cauchy was anxious to keep his discovery to himself and to develop mathematical analysis as his own subject. This is a controversial remark to make about Cauchy as a person, but seems to me fairly typical of both him and a number of his Parisian `colleagues'. With regard to the problem of Cauchy obtaining a copy of Bolzano's pamphlet, I found at least one plausible opportunity -- the Bibliothèque Impériale (now the Bibliothèque Nationale) began taking the Prague journal in which it appeared with precisely the appropriate volume. Bolzano and Cauchy did not meet at this time, but they did become acquainted in the 1830's, and I conclude my paper with Bolzano's reminiscences. This paper draws frequently on two books of mine, which are to be published by the M.I.T. Press: ``The Development of the Foundations of Mathematical Analysis from Euler to Riemann'', published in (1970; Zbl 0215.04401); and ``Joseph Fourier 1768--1839'', written in collaboration with Dr. J. R. Ravetz and appeared in (1972; Zbl 0245.01008).