Abstract
The pair (A, Δ ), where A is a physical quantity (an observable) and Δ a subset of the reals, may be called an 'experimental question'. The set Q of experimental questions is, in classical mechanics, a Boolean algebra, and in quantum mechanics an orthomodular lattice (and also a transitive partial Boolean algebra). The question is raised: can we specify a priori what algebraic structure Q must have in any theory whatsoever? Several proposals suggesting that Q must be a lattice are discussed, and rejected in favor of the weak claim that Q must be a Boolean atlas.