Whether we like it or not, “political correctness” is a phenomenon that has characterized languages and discourses in the Western world for the past five decades. Born in the US in the 1970s and imported in Europe in the early 1990s, “political correctness” is associated with a set of practices, including language innovation and use, aimed at promoting respect, inclusiveness and equity for all sections of the population. “Political correctness” is interesting from a linguistic viewpoint because the controversy it is inextricably tied to hinges on complex theoretical issues about language, namely (1) the relationship between language and reality, (2) the relationship between form and meaning and (3) the role of context in shaping language. Research on “politically correct” language has been conducted in critical discourse analysis and semiotic theory (Andrews 1996; Fairclough 2003; Suhr & Johnson 2003), as well as in pragmatics (Bolinger 1980; Lakoff 2000; Hughes 2010; Halmari 2011) and sociolinguistics (Allan & Burridge 1991, 2006; Cameron 1995). The general consensus emerging from those fields is that such language does not involve a mere relabeling of unchanged concepts but rather a process of reframing or “looking at [a given reality] from a different angle” (Cameron 1995: 147). Although the notions of frames, perspectivization and construal in general are central to cognitive linguistics (Langacker 1991; Fillmore 1982; Croft & Cruse 2004), “politically correct” language has not yet been explored by the field. My main research goal is to fill this gap. Drawing on corpus data, I explore the reframing processes at work in specific English “politically correct” expressions using the tools of cognitive linguistics. I argue that “politically correct” language rests, by and large, on a few combinable semantic features (individuality, agency, dynamicity/punctualness, gradability, contingency) conventionally associated with particular forms, while an additional element of vagueness triggers context-based inferencing.