Ambiguities of fundamental concepts in mathematical analysis during the mid-nineteenth century (Q2391790)

The ambiguities of mid-nineteenth century analysis are illustrated here through an account of E. G. Björling's reception in 1852 of Cauchy's 1840 work on expansion of complex-valued functions in power series. The focus is on the question of the differentiability conditions for a function to be so representable. Cauchy stated that the function and its first derivative had to be continuous. (A few years later he questioned whether the continuity of the first derivative might be superfluous, but soon returned to confirm his original condition.) Björling maintained that it must have a continuous second derivative. Meanwhile, in 1846, A. H. E. Lamarle had asserted that Cauchy needed a certain periodicity condition in addition to continuity and that the continuous first derivative was superfluous. Björling argued that Lamarle's way of defining continuity explained why Lamarle felt the need for the additional condition. The author makes the point, however, that both mathematicians appeared to agree on what a continuous function at a specific point is. A key to understanding this difference between Cauchy and Lamarle is identified in a footnote by Björling, namely the difference in the way each defines the logarithm function with respect to complex variables. Noting that Cauchy refers to the possibility of a function being both continuous and discontinuous, it is clear that the term `function' is not being used in the present-day sense. Cauchy apparently takes function to mean simply the algebraic expression which may be characterized in different ways (e.g. continuous or discontinuous) depending on the domain chosen. This paper is mainly devoted to showing the resulting complications in discussions between mathematicians in the absence of agreed-upon definitions, especially that of function. The respective arguments of the mathematicians are analyzed and compared here in a clear and interesting way. This whole episode is presented as an example of the change in nineteenth-century analysis from being intuitional, empirical, and formula-centered, to conceptual, abstract and concept-centered. But the paper also claims to make the argument that Björling had a tendency to sometimes consider mathematical objects in a naturalistic way, `that he considers mathematical definitions as descriptions of entities rather than conventions'. Evidence is presented for this and it seems to be implied that it distinguishes him from the other players in the account, but a comparison is not explicitly made.