Geometrical mappings of Grassmannians of linear spaces satisfying the exchange axiom (Q866800)

The author considers \textit{linear spaces (= incidence spaces)} \( \Pi := \Pi(P,\mathfrak f L) \) (i.e. \(\forall p,q \in P, \;p \neq q \;\exists_1 L \in \mathfrak f L \;: \;p,q \in L \) and \( \forall L \in \mathfrak f L \;: \;|L| \geq 2\)), satisfying the exchange axiom. Therefore for \( \Pi \) and for \( \mathfrak f U \) the set of all subspaces of \(\Pi \) the notions \textit{base} and \textit{dimension} can be introduced. The author makes the assumption that \( n := dim \Pi \) is finite. For \( k \in \{0,1,\dots,n-1\} \) and \( T \in \mathfrak f U \) let \(\mathfrak f G_k(\Pi) := \{U \in \mathfrak f U \;| \;dim U = k\} \) be the set of all \(k\)-dimensional subspaces of \(\Pi \) called \textit{Grassmannian} and \( \mathfrak f G_k(T) := \{X \in \mathfrak f G_k(\Pi) \;| \;X \subseteq T \) or \( T \subseteq X \}\). If \( S,T \in \mathfrak f G_k(\Pi) \) with \(dim (S \cap T) = k - 1 \) then \(S\) and \(T\) are called \textit{adjacent}. If \(B\) is a base then the subset of \(\mathfrak f G_k(\Pi)\) spanned by points of \(B\) is called the \textit{base subset} of \(\mathfrak f G_k(\Pi)\). The author shows: Let \( \mathfrak f X \subseteq \mathfrak f G_k(\Pi) \) be an \textit{A-subset} (i.e. \( \forall S,T \in \mathfrak f X , S \neq T \;: S \) and \(T\) are adjacent), then there exists a subspace \( U \in \mathfrak f U \) such that \(\mathfrak f X \subseteq \mathfrak f G_k(U) \). If \( 0 < k < n-1 \) and \(U \in \mathfrak f G_{k-1}(\Pi) \) then \(\mathfrak f G_k(U) \) is a maximal A-subset. If \( \Pi := (P,\mathfrak f L) \) and \( \Pi' := (P',\mathfrak f L') \) are given with \( 3 \leq n := dim \Pi = dim \Pi' \) then an injection \( f \;: \;P \;\to \;P' \) is called an \textit{embedding} if \(\forall L \in \mathfrak f L, \exists L' \in \mathfrak f L' : \;f(L) \subseteq L' \) and \( \forall L' \in \mathfrak f L' \) there is at most one \(L \in \mathfrak f L \) such that \( f(L) \subseteq L'\). If moreover \(f\) maps independet subsets to independent subsets then \(f\) is called \textit{strong}. Then: Let \( f : \;\mathfrak f G_k(\Pi) \;\to \mathfrak f G_k(\Pi')\) be a bijection such that \(f\) and \(f^{-1}\) map base subsets to base subsets. If \( n > 2k + 1\) then \(f\) is induced by a collineation of \( \Pi \) to \(\Pi'\). If \( k = 1 \) and \(f\) is not induced by a collineation of \( \Pi \) to \(\Pi'\) then \( n = 3 \), \( \Pi \) and \(\Pi'\) both are projective spaces and \(f\) is induced by a collineation of \( \Pi \) to the dualspace \(\Pi'{}^\ast\) of \(\Pi'\) . Let \( f \;: \;\mathfrak f L \;\to \;\mathfrak f L' \) be an injection sending base subsets to base subsets and let \(n > 3\), then \(f\) is induced by a strong collineation of \( \Pi \) to \(\Pi'\).