The treatment of negation in Heyting's intuitionistic logic has been criticized for including the principle that contradictions imply everything. As a result (and perhaps because, in applications such as arithmetic, what is given up in the logic is recovered from mathematical axioms), some have thought that Johansson's minimal logic, which on a natural axiomatization differs from Heyting's precisely in not making this assumption, represents a purer form of constructivism see Susan Haack, Deviant Logic (Cambridge: Cambridge University Press, 1974), pp. 101-102. In this note I will argue that, from the perspective of one form of constructivism, Johansson's logic can be seen as incorporating what, in other contexts, is a notoriously non-constructive principle. The late H. B. Curry's Foundations of Mathematical Logic (New York: McGraw-Hill, 1963), is written from the point of view of his formalistic constructivism, on which mathematics is seen as the study of formal systems. A formal system is thought of as, typically, containing some set of axioms, or primitive truths, and an assertion of the system is deemed true just in case it can be derived from these axioms. On this basis Curry motivates the rules of constructive positive logic, the negation-free fragment common to Heyting's logic and Johansson's. In introducing negation (p. 255) Curry suggests an extension of the notion of a formal system, on which a system can have, in addition to axioms, a set of primitive falsehoods, or counteraxioms. In analogy to the conception of truth in a formal system as derivability from its axioms, an assertion is to be thought of as false in a system (and its negation true) if one or another of the counteraxioms is derivable from it. The logic motivated by this conception of falsity is properly weaker that Johansson's, and until a better name is suggested, we may call it the logic of subminimal negation. It can be obtained from constructive positive logic by adding a symbol for negation and postulat-