A foundational delineation of computational feasibility

A principle directly pertinent to feasibility, which justifies the identification of P-time with feasible computing, is proposed. It is shown that the computable functions justified on the basis of positive quantifier-free comprehension are precisely the functions computable in deterministic polynomial time. This shows that the class P-time arises naturally from a foundational analysis of feasibility, and that terms using exponentiation can be justified as meaningful only under the admission of infinite sets as completed totalities.<<ETX>>