Public opinion research is not an area that has received much attention from applied linguists. But language lies at the heart of the procedures used to define, elicit, and report opinions, whether through such methods as polling, interviews, and focus groups, or through the less obvious channels of vox pop interviews, letters to the editor, radio phone-ins, or public hearings. In this paper I consider ways in which work in language studies--by linguists, conversation analysts, and social psychologists--might help us understand, improve, or question these procedures, and also give a new perspective on what opinions are. The relation of this new perspective to existing knowledge of academic and professional practitioners in the field may have parallels in other applications where applied linguists encounter experienced professionals and large organizations. Institutions of opinion feed back into political processes, so it is important that we consider them, not only as researchers and professionals, but also as citizens.