There Are No Maximal Low D.C.E. Degrees.
Rod Downey,Liang Yu
DOI: https://doi.org/10.1305/ndjfl/1099080209
2004-01-01
Abstract:We prove that there is no maximal low d.c.e degree. IntroductionA natural extension of the notion of a computably enumerable (c.e.) set is that of a d.c.e.set which is a set obtained as the difference of two c.e. sets A = W -V .Equivalently, a d.c.e set A is a set for which there exists a computable function f (x, s) so that A(x) = lim s f (x, s), f (x, 0) = 0, and ∀x|{s, f (x, s) = f (x, s + 1)}| ≤ 2. As well as being interesting in their own right, the d.c.e.Turing degrees can be studied both to give insight into the c.e. Turing degrees and into the 0 2 degrees.The investigation of the present paper can be viewed as contributing to all three of these goals.The uppersemilattice of the d.c.e.degrees is not elementarily equivalent to that of the c.e. degrees by Arslanov [1] and Downey [5].Perhaps the most striking difference between the d.c.e.degrees and the c.e. degrees comes from the following two theorems.Theorem 1.1 (Sacks [9]) The c.e. degrees are dense.Theorem 1.2 (Cooper, Harrington, Lachlan, Lempp, Soare [4]) The d.c.e.degrees are not dense.Indeed, there is a maximal d.c.e.degree a.That is, a < 0 ′ , and there are no d.c.e.degrees b with a < b < 0 ′ .Also notice that density properties allow us to compare the d.c.e.degrees and the 0 2 degrees.By an unpublished result of Lachlan, there are no minimal d.c.e.degrees, yet Sacks constructed a minimal 0 2 degree.Actually there is a very interesting theme here that "toward 0" the d.c.e.degrees are like the c.e.