The issue of local vs. global computation is currently debated in several subfields of linguistics. Against this background, I consider the mechanisms determining scalar implicatures. While implicatures have traditionally been analyzed as the result of a global mechanism (Grice 1978), Chierchia (2001) recently advocated a local view. I advance arguments against Chierchia's position, and present new evidence in support of the global view. More specifically, the argument has two parts to it: 1) Chierchia claims that the global view lacks an account for certain groups of data including disjunction. I show that the global view provides an elegant solution to the disjunction puzzle, which will be based on two independently motivated modifications: (i) an adjustment to the definition of scales used for disjunction, and (ii) a reassessment of the epistemic status of implicatures. 2) I argue that only the global view makes the correct prediction in cases with universal quantification. To this end, I argue for a novel semantics of tense. I propose that present tense is semantically vacuous, and that constraints on its interpretation arise from scalar implicatures. The results of this study have important consequences for other subfields of linguistics. If correct, they demonstrate that the conceptual motives offered by those who pursue strictly local accounts cannot be absolute --- implicature computation shows that our linguistic knowledge includes global mechanisms. References: Chierchia, Gennaro. 2001. "Scalar Implicatures, Polarity Phenomena, and the Syntax/Pragmatics Interface." Unpublished Manuscript, University of Milan Bicocca. Grice, Paul. 1978. "Logic and Conversation." In: P. Cole and J. L. Morgan, "Speech Acts", 41-58. New York: Academic Press.