A new approach to derivation (Q4779993)

A partial linear space is a pair \((\mathcal P,\mathcal L)\), where \(\mathcal P\) is a set of points and \(\mathcal L\) is a set of subsets of \(\mathcal P\), called lines, such that any line contains at least two points and any two distinct points lie on at most one line. A parallelism in \((\mathcal P,\mathcal L)\) is an equivalence relation on \(\mathcal L\) such that through each point there is exactly one line of each parallel class. A net is a partial linear space with a parallelism where any two non-parallel lines meet. Let \(\mathcal P'\subseteq \mathcal P\) and \(\mathcal L' \subseteq \mathcal L\). The pair \((\mathcal P',\mathcal L')\) is an affine subplane of the net \((\mathcal P,\mathcal L)\) if \((\mathcal P',\mathcal L')\) is an affine plane with incidence and parallelism inherited from \((\mathcal P,\mathcal L).\) A derivable net is a triple \((\mathcal P,\mathcal L,\mathcal B)\), where \((\mathcal P,\mathcal L)\) is a net and \(\mathcal B\) is a set of affine subplanes of \((\mathcal P,\mathcal L)\) such that the following conditions hold: (1) Each subplane \(B\in \mathcal B\) and each parallel class of \(\mathcal L\) have a common line. (2) For each subplane \(B\in \mathcal B\) and for each point \(p\in \mathcal P\) there is a line of \(B\) containing \(p\). (3) For any two different points \(p, p' \in \mathcal P\) that are joined by a line in \(\mathcal L\) there is exactly one subplane in \(\mathcal B\) containing \(p\) and \(p'\). \textit{N. L. Johnson} [ Abh. Math. Semin. Univ. Hamburg 58, 245--253 (1988; Zbl 0686.51002)]{} proved that every derivable net \((\mathcal P,\mathcal L,\mathcal B)\) can be represented in a \(3\)-dimensional projective space PG\((3,F)\) over a (skew) field \(F\) in the following way: The elements of the point-set \(\mathcal P\) of the net are the lines of PG\((3,F)\) that are skew to a fixed line \(L\). The line-set \(\mathcal L\) of the net contains all intersections between \(\mathcal P\) and the stars of lines not containing \(L\). The affine Baer subgeometry \(\mathcal B\) consists of all intersections of \(\mathcal P\) with the ruled planes not containing \(L\). And vice versa every such triple gives a derivable net. In the paper under review, an algebraic description of derivable nets is obtained starting from the geometric one given by N. L. Johnson. Indeed the following is proved: Let \((\mathcal P,\mathcal L,\mathcal B)\) be a derivable net. Then there are a (skew) field \(F\) and a \(2\)-dimensional left vector space \(U\) over \(F\) such that \(\mathcal P=\text{End}_F(U)\), the lines and Baer subgeometries are the sets \(\alpha \mathcal P +\beta\) and \(\mathcal P\alpha+\beta\), respectively, where \(\alpha \in \mathcal P \setminus (\mathcal P^*\cup \{0\})\) and \(\beta \in \mathcal P\). Vice versa in this way we always obtain a derivable net.