A combinatorial geometric structure on the space of orders of a field. I (Q1367591)

It is an easy observation and has been known for quite some time that the set of orderings \(X(K)\) of any formally real field \(K\) is a subset of a \(GF(2)\)-vector space, namely of the set of all half-orderings of \(K\). This latter set may be identified with the group of quadratic characters \(\chi(K^*/K^2)\) of the multiplicative group \(K^*\) of \(K\), which carries a natural topology (as the dual of a discrete group) and is a vector space over \(GF(2)\) (as a group of exponent two). Both structures on \(X(K)\) are mirrored within Murray Marshall's abstract theory for spaces of orderings. In view of this easy and nearly folkloristic fact, it is somewhat astonishing that, after introducing and studying the appropriate and of course canonical closure operator on \(X(K)\), the author of the paper under review needs nearly four pages in order to show that his concept makes \(X(K)\) a matroid representable over \(GF(2)\). Matroids are mere generalizations of subsets of vector spaces and therefore usually have less structure than \(X(K)\) has. So, at the first angle, the approach of the author to view \(X(K)\) as a matroid seems not to be very promising. However, matroid theory focuses at the combinatorial properties of not necessarily closed geometries, and therefore supplies the appropriate language to describe \(X(K)\) as a subset of a vector space. Indeed, further results of the author, such as the identification of SAP spaces as free geometries, the determination of fans by the combinatorial properties of their associated geometries, or the characterization of spaces of orderings with reduced stability index at most two as unimodular matroids, prove that matroid theory yields the right vocabulary to study spaces of orderings. (In the opinion of the referee, the cited results should immediately go over to Marshall's abstract setting.) Finally, the approach of the author provides us with a new, rich and well studied class of binary, not necessarily finite matroids.