The fixed point theorem says that every computable transformation of programs has a fixed point, up to the function computed. A function that transforms every program into one computing a different function cannot be computable; this lesson studies the degrees of such fixed-point free functions. They coincide with the degrees of diagonally non-computable functions, which differ from the diagonal wherever it is defined, and with the degrees of effectively immune sets. Arslanov’s completeness criterion then shows that among c.e. sets, only the complete ones have this power.
3.1 DNC and fixed-point free functions
A total function is diagonally non-computable, or DNC, if for every such that . A total function is fixed-point free if for every .
No DNC function is computable: if were total, then . The set computes a DNC function, namely if and otherwise. The PA sets compute -valued DNC functions (Definition 4.1 of “PA degrees”). No fixed-point free function is computable, by the fixed point theorem (Fixed point theorem).
A set computes a DNC function if and only if it computes a fixed-point free function.
() Let be DNC. By the parameter form of SMN (Parameter form), there is a total computable function such that is the function defined only at the input , with value . Let ; it is -computable. If for some , then , contradicting the fact that is DNC. So is fixed-point free.
() Let be fixed-point free. By SMN, there is a total computable function such that the program , on input , computes and, if this converges with value , runs . So when . Let ; it is -computable and total, since it does not run the computation . If for some with , then , contradicting the fact that is fixed-point free. So is DNC.
The reductions are uniform. A DNC function avoids, at each , the single value . The next theorem shows that DNC functions can be used to avoid, uniformly, any finite c.e. set whose size is bounded in advance.
3.2 Effectively immune sets
An infinite set is effectively immune if there is a total computable function such that for every with . A c.e. set is effectively simple if its complement is effectively immune.
An effectively immune set is immune: an infinite c.e. subset would have more than elements. Effective immunity requires that a c.e. subset of reveals, through its index, a bound on its size.
For every set , the following are equivalent:
- computes a DNC function;
- computes a function such that whenever ;
- computes an effectively immune set.
Let denote a computable bijection between and , uniform in .
(1 ⇒ 2) Let be DNC. By SMN, there is a total computable function such that the program , for , on any input, waits for the -th element enumerated into (counting from ), and if it appears, outputs the -th component of the tuple ; otherwise it diverges. For , let
and . It is -computable. Suppose that and . Then , and is the -th element enumerated into for some . So the program , on input , outputs the -th component of , which is . This contradicts the fact that is DNC.
(2 ⇒ 3) Let be as in (2). Let be the partial computable function such that is the canonical index of the set of the first elements enumerated into , defined if and only if . We build an -computable strictly increasing sequence such that, for every ,
Suppose that are defined, for some . By SMN, uniformly in and in the finite sequence , one computes an index with
where the first part is empty for . This set has at most elements, a number computable from and . Let ; then . So , and for every with .
We check . Let with . If , by some element of is not among , and it is not either, since . If , the set has elements and does not contain ; it cannot be contained in the -element set .
Let ; it is infinite and -computable, since the sequence is increasing. If and , then and is finite, so for some , contradicting . So whenever : is effectively immune, with .
(3 ⇒ 1) Let be effectively immune, with bound . By SMN, there is a total computable with . Let , where is the canonical index of the set of the first elements of ; is -computable. If , then is a subset of with elements, contradicting the choice of . So , hence for every : is fixed-point free, and computes a DNC function by Theorem 3.2.
3.3 Arslanov’s completeness criterion
The halting set computes a DNC function. Arslanov showed that it is the only c.e. degree with this property.
A c.e. set is Turing complete if and only if it computes a DNC function.
If , it computes a DNC function. Conversely, let be c.e., with a c.e. approximation , and suppose that computes a DNC function. By Theorem 3.2, there is a functional such that is total and fixed-point free. Let if this computation converges, and otherwise; the function is computable. Let be the set of with .
By SMN, there is a total computable function such that the program , on input , searches for the least with and, if it finds one, runs . By the fixed point theorem with parameters (Fixed point theorem with parameters), there is a total computable function with for every .
With oracle , compute for each the least stage such that converges, with use , and , for . Such a stage exists, since is defined and eventually equals ; and the condition is decided by . Since is c.e., for every , so the same computation gives for every .
We claim that if and only if , which shows . Suppose that , and let be the least stage with . If , then the program runs , so , contradicting the fact that is fixed-point free. So , and , since . Conversely, .
The program waits for to enter and then copies the program that the current approximation of assigns to it. If entered after the approximation of has settled, would be a fixed point of . So, once knows that the approximation has settled, the entry of into is already decided. The c.e. hypothesis is necessary: there are low PA sets (Corollary 4.3 of “PA degrees”), which compute DNC functions without computing .
- A c.e. set of PA degree is Turing complete.
- (Martin) Every effectively simple set is Turing complete.
- The complement of a low simple set is immune but not effectively immune.
- A PA set computes a DNC function.
- If is effectively simple, then computes its complement, which is effectively immune; by Theorem 3.4, computes a DNC function, and Theorem 3.5 applies.
- A low simple set exists (Theorem 3.1 of “A low simple set”). Its complement is immune; if it were effectively immune, the set would be complete by 2, hence not low.
3.4 Exercises
Show that if computes a DNC function, then the degree of contains a DNC function.
Solution
Let be DNC. By the padding lemma, there is a computable strictly increasing sequence of indices of the nowhere defined function; its range is computable. Let if , and if . Then , and since . For , when ; for , diverges. So is DNC.
Show that there is a non-computable set, computable in , that computes no DNC function.
Hint
Build the set by finite extensions. For a functional and a string , either diverges for some and all , or the first convergent extension defines a total computable function, which agrees with the diagonal somewhere.
Solution
Build with oracle , alternating two kinds of steps. To make the set different from , extend by a bit different from if it converges, by otherwise; decides this. To defeat as a DNC function, given , ask whether there is such that for every , a question.
- If yes, find such an with and keep : every has , so is not total.
- If no, let be the value for the first found with convergence; is total computable. By the fixed point theorem applied to the function (an index of the constant function with value ), there is with constant equal to , so . Extend to the string found for . Then every has , so is not DNC.
The limit is computable in , is not computable, and computes no DNC function.
3.5 Notes and references
Arslanov (1981) proved the completeness criterion, in the form of fixed-point free functions; Martin (1966) had proved that effectively simple sets are complete. The equivalence between DNC and fixed-point free functions is due to Jockusch, Lerman, Soare and Solovay (1989), and the characterization through effectively immune sets to Jockusch. In the course on algorithmic randomness, DNC functions reappear through the complexity of the initial segments of a set.
- M. M. Arslanov, Some generalizations of a fixed-point theorem, Soviet Mathematics (Izvestiya VUZ. Matematika) 25 (1981), 1–10.
- D. A. Martin, Completeness, the recursion theorem and effectively simple sets, Proceedings of the American Mathematical Society 17 (1966), 838–842.
- C. G. Jockusch, M. Lerman, R. I. Soare and R. M. Solovay, Recursively enumerable sets modulo iterated jumps and extensions of Arslanov’s completeness criterion, Journal of Symbolic Logic 54 (1989), 1288–1323.
- C. G. Jockusch, Degrees of functions with no fixed points, in Logic, Methodology and Philosophy of Science VIII, North-Holland, 1989, 191–201.