Combinatorial characterization of pseudometrics (Q2204104)

Recall that a mapping \(f\) from a metric (or pseudometric) space \((X,d)\) to a metric (or pseudometric) space \((Y,\rho)\) is \textit{isometric} if \(\rho(f(x), f(y))=d(x,y)\) for all \(x,y\in X\). This classic concept has numerous generalizations. In this paper, the authors introduce \textit{similarity}, \textit{weak similarity}, and \textit{combinatorial similarity}. Furthermore, they expand the concept of combinatorially similar pseudometrics to the concept of combinatorially similar functions of two variables, as follows. For sets \(X\) and \(Y\), let \(\Phi\) and \(\Psi\) be mappings with the domains \(X^2\) and \(Y^2\), respectively. The mapping \(\Phi\) is \textit{combinatorially similar} to \(\Psi\) if there are bijections \(f:\Phi(X^2)\to\Psi(Y^2)\) and \(g:Y\to X\) such that \(\Psi(x,y)=f(\Phi(g(x), g(y)))\) for all \(x,y\in Y\).\par The authors give a characterizations of mappings which are combinatorially similar to pseudometrics, strongly rigid pseudometrics or discrete pseudometrics. For instance, in Section 3 they prove the following: Let \(X\) be a nonempty set. The following conditions are equivalent for every mapping \(\Phi\) with the domain \(X^2\): (i) \(\Phi\) is combinatorially similar to a pseudometric, (ii) \(\Phi\) is symmetric, there is \(a_0\in\Phi(X^2)\) such that \(\Phi\) is \(a_0\)-coherent, and the inequality \(\vert \Phi(X^2)\vert \leq\mathfrak{c}\) holds, where \(\mathfrak{c}\) is the cardinality of the continuum.