Exoteric homomorphisms (Q1097924): Difference between revisions

Let R be a commutative reduced ring. A ring homomorphism \(f: R\to S\) is called exoteric, if the following condition holds: If two finitely generated ideals of R have the same annihilator in R, then their images under f have the same annihilator in S. An ideal of R is exoteric, if it is the kernel of an exoteric homomorphism. These concepts have been investigated under other names by \textit{G. Picavet} [Math. Scand. 46, 23-53 (1980; Zbl 0491.13003)] and \textit{T. P. Speed} and the author [J. Austral. Math. Soc. 13, 1-6 (1972; Zbl 0227.13002)]. An example of an exoteric homomorphism is given by \(R\to N(R)\), where N(R) denotes the epimorphic hull in the sense of the reviewer [Comment. Math. Helv. 43, 378-401 (1968; Zbl 0165.053)]. The exoteric ideals of R turn out to be the restrictions of the ideals of N(R) and also of those of the maximal quotient ring \(Q_{\max}(R)\). It is shown, that N(R) can be represented as the ring of global sections on the space of exoteric prime ideals endowed with the constructible topology. Using exoteric homomorphisms and ideals, the author characterizes rings such that \({}_ RQ_{\max}(R)\) is flat (e.g.: R has this property iff every prime exoteric ideal is a torsion ideal) and rings (in particular rings of continuous functions) containing no finitely generated dense ideals. The latter class is also investigated in a non-commutative setting.