In the spirit of B. C. van Fraassen's view of science called Constructive Empiricism, we propose a scientific criterion to decide whether a concrete object is observable, as well as a coextensive scientific-philosophical definition of observability, and we sketch a rigorous account of modal language occurring in science. We claim that our account of observability solves three problems to which current accounts of observability, notably van Fraassen's own accounts, give rise. We further claim that our account of modal propositions (subjunctive conditionals included), which proceeds wholly within the framework of the semantic view on scientific theories, grounds his claim that such an account is possible without relying on 'inflationary metaphysics', notably without postulating an infinitude of different universes besides the universe we inhabit. We thus claim to solve a fourth problem: how to give a precise nominalist account of modal language in science tailor-made for Constructive Empiricism.