On profinite spaces of orderings (Q1940194)

In this paper, the author studies spaces of orderings \((X,G)\) in the sense of \textit{M.~A.~Marshall} [Spaces of orderings and abstract real spectra. Berlin: Springer (1996; Zbl 0866.12001)] where \(X\) is a set, \(G\) is a subgroup of \(\{ 1,-1\}^X\) that contains the constant function \(-1\), separates points in \(X\) and satisfies some further axioms. \(X\) can then be considered as a subset of the character group \(\chi(G)\) by identifying \(x\in X\) with the character \(G\to\{ 1,-1\}:a\mapsto a(x)\). One can define morphisms between spaces of orderings and inverse limits by suitably defining inverse systems of such spaces. A profinite space of orderings is one that can be gotten as an inverse limit of finite spaces of orderings. If \(k\) is a formally real field and \(X_k\) denotes the set of all orderings of \(k\), and \(G_k\) denotes the multiplicative group \(k^*/(\sum k^2)\), the classes of nonzero elements modulo sums of squares, then \(G_k\) can be identified with a subgroup of \(\{ 1,-1\}^{X_k}\) as follows: for every \(a\in k^*\) one maps its class in \(k^*/(\sum k^2)\) to \((\overline{a}(\sigma))_{\sigma\in X_k}\) where \(\overline{a}(\sigma)=1\) or \(-1\) according as \(a\in\sigma\) or not. Thus, \((X_k,G_k)\) becomes a space of orderings. The first main result in the present paper is that \((X_{\mathbb{Q}(x)}, G_{\mathbb{Q}(x)})\) is profinite where \(\mathbb{Q}(x)\) is the rational function field in one variable over \(\mathbb{Q}\). An explicit description of the elements of \(X_{\mathbb{Q}(x)}\) can be found in [\textit{M. Dickmann, M. Marshall} and \textit{F. Miraglia} [Ann. Pure Appl. Logic 132, No. 1, 27--49 (2005; Zbl 1080.03017)]. The second main result is a proof of a theorem due to \textit{V. Astier} and \textit{M. Tressl} [Arch. Math. Logic 44, No. 1, 77--95 (2005; Zbl 1099.03027)] stating that if a space of orderings \((X,G)\) is an inverse limit of an inverse system of spaces of orderings, such that the so-called pp conjecture holds for each space in that system, then the pp conjecture also holds in \((X,G)\). While the original proof uses model theory, the present proof only uses the theory of spaces of orderings itself plus some elementary topology. Both proofs, however, make use of a key lemma due to Marshall. The paper concludes with a few consequences and remarks, including that the above provides a new proof that the pp conjecture holds in \((X_{\mathbb{Q}(x)}, G_{\mathbb{Q}(x)})\), originally due to Dickmann, Marshall, Miraglia [loc. cit.].