Ordered exponential fields (Q1594594)

This volume contains a wealth of recent material on ordered (usually non-archimedean) exponential fields, only a small amount described below. The strategy is to study closely the interaction of an exponential function with the natural valuation (for which the valuation ring is the set of maximal elements). In Chapter 1 it is shown that an exponential function \(f\) on an ordered field \(K\) induces a \textit{left exponential}, a \textit{middle exponential} and a \textit{right exponential} with respect to the natural valuation \(v\). In addition, every exponential function induces a mysterious map, a `group exponential', on the value group; namely, an order-isomorphism from the set of values of the group to the set of negative elements. This yields also a `contraction mapping' on the value group. The model theory of contraction groups has been developed by F.-V. Kuhlmann, and is described in an appendix. From these results, it is shown that if \(K\) is a countable non-archimedean field with a divisible multiplicative group of positive elements, then any exponential on the residue field \(\overline{K}\) lifts to an exponential on \(K\) precisely if the value group is isomorphic to a Hahn sum indexed by \({\mathbb Q}\) of copies of \((\overline{K},+,0,<)\). In Chapter 2 the author introduces two axiom schemes for an exponential \(f\) on an ordered field \(K\), namely the growth axiom (GA) (which says \(a\geq n^2 \Rightarrow f(a)>a^n\)) and the Taylor axiom (T) (that \(|a|\leq 1 \Rightarrow |f(a)-E_n(a)|<|a^{n+1}|\), where \(E_n\) is a partial Taylor sum). If both hold, \(f\) is said to be a (GAT)-exponential. Results are given in the countable case on lifting (GAT)-exponentials from the residue field. The schemes (GA) and (T) are translated into valuation-theoretic language. By a theorem of Ressayre, the theory of the real exponential is axiomatised by the theory of the restricted real exponential, and (GAT). Chapter 3 includes a general account of convex valuations, the associated rank (the ordered set of convex valuation rings other than the natural one), and its exponential analogue `exponential rank'. This encodes the growth rate of an exponential, and it and (GA) are encoded using order-automorphisms of the value set of the value group. In Chapter 4, the main results are that if \(k\) is an archimedean ordered field and \(G\) a non-trivial ordered abelian group, then the power series field \(k((G))\) has no exponential (and without the archimedean assumption, there is no exponential compatible with the valuation). However, non-Archimedean ordered exponential fields which are countable unions of power series fields are constructed (with (GA), after a twist by a value group automorphism). This yields non-archimedean models of the theory of the reals with restricted analytic functions and exponentiation, and exponentials with many different exponential ranks. Chapter 6 contains an introduction to Hardy fields and connections with o-minimality. Hardy fields provide non-trivial examples of non-Archimedean exponential fields. It is shown that an o-minimal expansion of the reals with exponentiation is exponentially bounded precisely if the exponential rank of the Hardy field is 1. Finally, a new proof is given of Hardy's conjecture that the compositional inverse of the function \(\text{ log}(x)\text{ log}_2(x)\) is not asymptotic to an LE-series. A new proof is also given that the restriction to \((1,\infty)\) of the Riemann zeta-function is not definable in the real field with restricted analytic functions and exponentiation.