According to long-standing wisdom, relative clauses and tensed clauses are supposed to be *scope islands*. This means that scope-takers appearing inside a relative clause or a tensed clause are supposed to not be able to undergo Quantifier Raising in order to take scope outside of the clause. However, the original evidence motivating these claims is not compelling. Furthermore, it is easy to find abundant systematic naturally-occuring counterexamples. Therefore we need to rethink scope islands, both empirically and theoretically. In particular, many semantic analyses have been built over the years in part on the assumption that Quantifier Raising is clause bounded. Notable examples include focus computation, the semantics of indefinites, and functional relative clauses. If Quantifier Raising is not in fact clause bounded, we should reconsider it as a viable approach to handling these phenomena. In addition, we should resume the work of deciding what syntactic contexts *are* scope islands, which means at least figuring out a way to encode scope islands in a concrete formal grammar. Building on Barker and Shan 2014, Kiselyov and Shan 2014, and Kokke 2016, I propose a simple but flexible formal framework for encoding and enforcing scope islands.