The U.S. Census Bureau’s language statistics, like those of other national statistical agencies, are put to a wide range of public and private uses, but many users may be unaware of how or why these statistics are produced. After presenting a taxonomy of census language questions and their policy uses from around the world, this article examines how and why the U.S. Census Bureau asks about language. Specifically, I explain the federal policies that led to the development of the current three-part language question, analyze how policy has changed since it was introduced, and synthesize the survey methodological research that has been used to evaluate the validity and reliability of the resulting data. I conclude with a discussion of a recent qualitative study that investigated the basis of respondents’ answers to the language question, as well as the criteria that respondents use when evaluating the English-speaking ability of other household members.