In this talk I argue that two ways of scope-taking provided by the grammar—movement and focus alternative computation—are fundamentally incompatible with one another. Based on data from intervention effects in English questions, I show that movement cannot target a region in the structure in which focus alternatives are being computed. Instead, movement must target a position above or below such regions, or another scope-taking mechanism must be used. This proposal provides an empirical argument against higher-typed and variable-free semantics, which have been proposed to avoid such a theoretical incompatibility, and support for a simple-typed system with movement alongside alternative computation as scope-taking mechanisms. The proposal has far-reaching implications for a wide array of linguistic phenomena, including the nature of movement, focus, intensionality, and binding, as well as for the theory of intervention effects.