Complementing cappable degrees in the difference hierarchy. (Q1428038)

The authors study r.e.\ and d.r.e.\ degrees (which are called c.e.\ and d.c.e.\ in the paper). A degree is called r.e.\ if it contains an r.e.\ set and d.r.e.\ if it contains a set which is the difference of two r.e.\ sets. The authors show that the following statements are equivalent for any nonzero r.e.\ degree \(a\): (1) \(a\) is cappable. That is, there is a nonzero r.e.\ degree \(b\) such that \(0\) is the only Turing degree below \(a\) and \(b\). (2) \(a\) has an d.r.e.\ complement \(c\). That is, \(0\) is the only degree below \(a\) and \(c\), \(0'\) is the least Turing degree above \(a\) and \(c\). (3) \(a\) cups with an d.r.e.\ degree \(c\) such that \(c\) is isolated by \(b\) and \(a\), \(b\) form a minimal pair. That is, \(0\) is the only Turing degree below \(a\) and \(b\), \(0'\) is the least Turing degree above \(a\) and \(c\), \(b\) is the greatest r.e.\ Turing degree below \(c\). Note that (3) is not a direct corollary from (2) since there are d.r.e.\ degrees which are not isolated by an r.e.\ Turing degree below. Statement (1) is an obvious corollary of (3). \textit{Guohua Wu} [J. Symb. Log. 67, No. 3, 1055--1064 (2002; Zbl 1023.03037)] proved that (3) implies (2). So the main result of the paper is that (1) implies (3).