Infinitesimals without logic (Q455512)

This article develops the theory of \textit{Fermat reals} as a framework for using infinitesimals in analysis and mathematical physics that does not require any nontrivial results of mathematical logic. A function \(x:\mathbb R_{\geq 0}\to \mathbb R\) is said to be a \textit{little-oh polynomial} if there are \(k\in \mathbb N\), \(r,\alpha_1,\ldots,\alpha_k\in \mathbb R\), and \(a_1,\ldots,a_k\in \mathbb R_{\geq 0}\) such that \(x(t)=r+\sum_{i=1}^k \alpha_i\cdot t^{a_i}+o(t)\) as \(t\to 0^+\). A Fermat real is then the equivalence class of a little-oh polynomial with respect to the equivalence relation \(x(t)=y(t)+o(t)\) as \(t\to 0^+\). The collection of Fermat reals forms an ordered ring \({}^\bullet \mathbb R\) extending the field of real numbers which contains so-called \textit{nilpotent infinitesmals}. (We emphasize ring here as the usual nonstandard extensions of the real numbers employed in nonstandard analysis are fields). The article discusses the structure of \({}^\bullet \mathbb R\) and then gives examples of its uses in mathematical physics.