Traditional research in bilingualism has consistently found that switching languages is effortful, placing demands on neural systems of cognitive control. This finding runs counter to most bilinguals' intuitive experience. We review a body of recent work showing that, in fact, when bilinguals switch languages voluntarily, both the behavioral cost of switching and the associated recruitment of cognitive control areas are greatly reduced or completely eliminated. This suggests that switching languages is not inherently effortful, but rather, particular communicative demands may make it costly. The new evidence also challenges the basic premise underlying the bilingual advantage hypothesis. We articulate a more nuanced version of it, in which the advantage is limited to bilinguals who frequently switch languages based on external constraints. Language-Switching and Traditional Ways to Model It Commanding more than one language is the norm for the majority of the world's population [1,2], and multilingualism has been notably increasing in recent years [3]. A remarkable feature of multilingual individuals is their ability to quickly and accurately switch back and forth between their different languages. Thus, multilingual individuals not only command each of their languages independently, but they are additionally able to alternate and coordinate the rules that govern each language. What are the mechanisms that allow bilingual individuals to effectively do this? How are these systems organized to achieve seamless transitions from one language to the other? Highlights Results of traditional language switching tasks have suggested that language switching is behaviorally effortful and requires increased involvement of cognitive control areas. Recent evidence has challenged this conclusion, showing that when bilinguals are allowed to switch languages freely, this cognitive effort and behavioral cost are either greatly reduced or completely eliminated.