Abstract. We define G, a substructure of E Π (the lattice of Π 0 1 classes), and show that a quotient structure of G, G ♦ , is isomorphic to E * . The result builds on the ∆ 0 3 isomorphism machinery, and allows us to transfer invariant classes from E * to E Π , though not, in general, orbits. Further properties of G ♦ and ramifications of the isomorphism are explored, including degrees of equivalence classes and degree invariance.
IntroductionA Π 0 1 class may be defined as the collection of infinite paths through a computable subtree of 2 <ω , the complete binary-branching tree. Π The lattice of all Π 0 1 classes is called E Π , by analogy with E, the lattice of computably enumerable (c.e.) sets. The properties of E have been extensively studied (for a survey, see [19], chapters X and XV). Research on Π 0 1 classes and E Π is currently quite active, with many open questions (see [3] for a number of examples). However, relatively little is known about the orbits and invariant classes of E Π . The goal of the research presented here is to expand that knowledge, in particular by transferring information to E Π from E * , the lattice of c.e. sets modulo finite difference.A Π 0 1 class P is principal (or clopen) if there is a finite set F of nodes of 2 <ω such that an infinite path of the tree is in P if and only if it extends some σ ∈ F . Cholak, Coles, Downey, and Herrmann [7] showed that there were at most two nonisomorphic intervals of the form [P, 2 ω ] in E Π : those where P is principal and those where it is nonprincipal. Cenzer and Nies [4] showed that these are in fact distinct cases.Nies proceeded to define G = [P, 2 ω ] for P nonprincipal. It is via G that we will transfer information from E * to E Π , and many of Nies' unpublished results are reproduced in § §3-4. Several of the results are directly proved in the setting of Π 0 1 classes. However, although the goal is to transfer information to E Π , it is generally Prior to being investigated in this context, G (as a collection of ideals) arose as part of the study of the lattice of c.e. substructures of a computably presented model, an area suggested by Metakides and Nerode in a 1975 paper [15]. A number of papers emerged studying substructures of particular models, such as vector spaces, algebraically closed fields, and Boolean algebras (see Nerode and Remmel [16] for references). Remmel [17] and later Downey [9,10] generalized the work on specific structures to results about effective closure systems (M, cl), where M is a computable set and cl : P(M ) → P(M ) is an effective closure operator, a map with certain properties. For example, the operator could take a subfield of M to its algebraic closure within M . As part of this work the notion of equivalence modulo finite difference, as in E * , is extended to equivalence modulo "finitely-generated difference." That is, A = * B if there is a finite set X such that cl(A ∪ X) = cl(B ∪ X). Downey in particular gives a long list of examples of effective closure systems which includes the remark that in...