scientific article; zbMATH DE number 15589

Authors Theo Grundhöfer, Helmut Salzmann

Publication date 25 June 1992

compact projective planes homogeneity of planes locally compact quasifields and nearfields shift planes topological automorphism group topological double loops topological projective planes topological ternary fields translation planes

Mathematics Subject Classification ID

Nonassociative division algebras (17A35) Infinite automorphism groups (20B27) Other topological algebraic systems and their representations (22A30) Non-Desarguesian affine and projective planes (51A35) Translation planes and spreads in linear incidence geometry (51A40) Topological linear incidence structures (51H10)

Review text

[For the entire collection see Zbl 0704.00017.] Via coordinatization, every topological projective plane gives rise to topological ternary fields and double loops, extensive generalizations of topological fields. In spite of the vast variety of arbitrarily wild topological double loops, fairly weak and reasonable assumptions on the topology or the underlying geometry have stringent consequences for the algebraic and topological structure and even allow a complete classification of topological double loops and ternary fields with special properties. As to the latter, in particular within the last decade, considerable and significant progress was made. The second author managed to complete this outstanding and extensive project (1962-1987) on characterizing the classic planes via their collineation groups: Let \(P\) be a compact connected projective plane coordinatized by a ternary field \(K\) of finite positive topological dimension \(\ell\), then deep algebraic topology yields \(\ell=2^ m\leq 8\); if the topological dimension of the group \(\Sigma\) of all continuous collineations of \(P\) is greater than \(g:=m^ 3+3m+4\), then \(K\) is the field of the reals, the complex numbers, the quaternion field, or the real octonion algebra; and if \(\dim \Sigma=g\) then \(P\) belongs to an infinite but well-known list of projective planes. R. Löwen even showed that any compact connected plane admitting a transitive group of automorphisms is classical (1981), and the first author was able to extend F. Kalscheuer's classification of locally compact connected nearfields to the nonlogical, planar Dickson nearfields constructed from locally compact fields (1987). The article under review contains a fairly complete, readable account of these and related results. Accompanied by a detailed bibliography, some effort has been made to illustrate the matter by proper examples and to give meaningful sketches of proofs elucidating the main ideas. Besides its benefit for the expert as a valuable reference work this paper opens the theory of locally compact double loops and ternary fields to a broader public, and may be called outstanding, at least with respect to this latter function. In detail, the paper is organized as follows. The first section is devoted to the notion of topological double loops and their properties: Any locally compact double loop \(D\) is a separable, complete metric space, \(\sigma\)-compact but not compact, and either connected (and then locally connected) or totally disconnected (and then homeomorphic to the Cantor set minus a point). The next section deals with the concept of topological ternary fields and topological projective planes. In particular, for locally compact ternary fields \(K\), Grundhöfer's conditions for \(K\) coordinatizing a topological (compact) projective plane \(P\) are given. Note that Problem 2.4, whether a topological ternary field always yields a topological projective plane, has been answered in the negative by \textit{E. Eisele} [Abh. Math. Semin. Univ. Hamb. 60, 257- 264 (1990; Zbl 0721.51016)]. The third section introduces the groups of continuous automorphisms \(\Sigma\) of \(P\) and \(\text{Aut }K\) of \(K\), which are locally compact in the compact-open topology, provided \(P\) is compact. The sections 4-7 treat double loops which satisfy some of the algebraic laws of field. Examples for linear ternary fields are given, isotopy is discussed, and Grundhöfer's conditions are translated to Cartesian fields in section 4. Since every quasifield \(Q\) is a vector space over its kernel \(K\), linear algebra is applied in section 5 to study topological quasifields. The vector space structure of locally compact connected quasifields is covered (\(K\) is \(\mathbb{R}\), \(\mathbb{C}\), or \(\mathbb{H}\), and \(\dim_ \mathbb{R} K\in\{1,2,4,8\}\)), the Haar measure is used to construct an absolute value on locally compact quasifields \(Q\), the compactness of \(\text{Aut }Q\) is proven, and finally some constructions are given. Locally compact nearfields and their classification by Kalscheuer and Grundhöfer, heavily based on Lie-theoretic arguments, are the concern of section 6 (see above), and the well-known classification results of Pontryagin, van Dantzig and Jacobson on locally compact (skew) fields are presented in section 7. The following sections 8-14 are solely devoted to the connected case. Section 8 covers some properties of locally compact connected double loops \(D\): if the topological dimension \(\ell\) of \(D\) is finite, then \(\ell\mid 8\) and the one-point compactification of \(D\) is homotopy equivalent to the sphere \(S^ \ell\) (Löwen 1983), if \(\ell\leq 2\) then \(D\approx \mathbb{R}^ \ell\), if \((D,+)\) is associative then \(D\approx \mathbb{R}^ \ell\), and if the multiplicative loop \(D^*\) is associative then it is isomorphic to \(\mathbb{R}^*\), \(\mathbb{C}^*\) or \(\mathbb{H}^*\). The next two sections deal with the automorphism groups of ternary rings and projective planes, present Löwen's result on planes with transitive automorphism groups, and culminate in Salzmann's classification of compact planes admitting automorphism groups of sufficiently large dimension (see above). The remaining sections 11-14 discuss the planes \(P\) with \(\dim\Sigma\leq g\) separately for the possible values of \(\ell\). In the case \(\ell=1\), the topology is given by an order which allows to construct a wide variety of planes by ``bending lines. For \(\dim \Sigma=g=4\), \(P\) is a real Moulton plane, and for \(\dim \Sigma=3\) all (infinitely many) planes are known (recent results of Groh and his pupils on the \(\dim \Sigma=2\) case are not yet included). In the case \(\ell=2\), one has \(\dim \Sigma=g=8\), if and only if \(P\) is a translation plane. All these planes and all flexible translation planes with \(\dim \Sigma=7\) are determined (Betten). In the case \(\ell=4\), all planes with \(\dim \Sigma=g=18\) can be coordinatized by semifields and are known (Hähl). The planes with \(\dim \Sigma=17\) are either translation planes or topological analogues of the Hughes planes and can also be described explicitly. Finally, in the case \(\ell=8\), all planes with \(38\leq \dim\Sigma\leq g=40\) have been determined (Hähl).

Recommendations

Cited in

(25)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3972230)