Abstract
Closure is the idea that what is true about a theory of truth should be true (and therefore expressible) in it. Commitment to closure under truth motivates non-classical logic; commitment to closure under validity leads to substructural logic (nontransitive or noncontractive). These moves can be thought of as responses to revenge problems. With a focus on truth in mathematics, I will consider whether a noncontractive approach faces a similar revenge problem with respect to closure under provability, and argue that if a noncontractive theory is to be genuinely closed, then it must be free of all contraction, even in the metatheory.