On co-\(\kappa\)-Souslin relations (Q2266010): Difference between revisions

The author's main result is that if a \(\kappa\)-Souslin set defines a quasi-order on the reals (and continues to do so when a Cohen-generic real is adjoined) then there exists either (a) a perfect set of disjoint intervals or (b) a dense set of size \(\leq \kappa\). The method of proof resembles that used by the author in an earlier joint paper with \textit{L. Harrington} [Logic colloquium '80, Eur. Summer Meet., Prague 1980, Stud. Logic Found. Math. 108, 147-152 (1982; Zbl 0513.03024)] to prove the corresponding result about equivalence relations (where the alternatives are: (a) a perfect set of inequivalent elements or (b) a set of size \(\leq \kappa\) containing one representative from each equivalence class). The result is proved in the first section; later sections indicate some extensions that are possible - and one that isn't: the parenthetical clause (''and continues...'') cannot be deleted from the statement of the theorem.