AND
ALBERT0 POLICRITICourant Institute
IntroductionThe decision problem for a given class C of formulas of set theory asks for an algorithmic procedure to test, for any given formula + in C , say with n free variables, whether there are sets xl,* * a , x , satisfying +. Therefore the decision problem for C is solvable when an algorithm M, say a deterministic Turing machine, can be found, accepting formulas in C as inputs, such that the following holds: [lo]), and various classes of formulas for which the decision problem has a positive solution (i.e., is solvable) have been found. In the present paper, which evolved from work on the incompleteness theorem done by the first author in [ll], we shall provide bounds on the possibility of finding such positive solutions.To see how negative results on the solvability of the decision problem can be obtained, let us note that presumably the only way we have of finding, for a given C , an algorithm M and establishing that (*) above in fact holds, is by making use of the currently accepted mathematical principles and way of reasoning. Therefore, as the above mentioned works and common (meta)mathematical experience suggests, if the decision problem for C has a positive answer, then we expect to have a description of M and a proof of ( * ) in a system of axiomatic set