This article is concerned with history that is fuzzy in the sense of impressionistic rather than systematic, using “soft” rather than “hard” data and concerned more with “lumping” than with “splitting.” It argues that there have been at least four phases in the two centuries of conflict between precise and fuzzy historians. In the first phase, in the nineteenth century, precise history, firmly based on documents, was defined, by Leopold von Ranke and the Rankeans, against an older fuzzy or “conjectural” history. In a second phase, at the beginning of the twentieth century, a new fuzzy history was defined against “positivist” history, by Karl Lamprecht in Germany, Lucien Febvre in France, and George Trevelyan in England, among others. In a third phase, in the middle of the twentieth century, practitioners of quantitative history (sometimes known as the New Economic History or more generally as “cliometrics”) condemned all nonquantitative historians as fuzzy. In a fourth phase, from the 1970s onward, a still newer fuzzy history, comprising historical anthropology, microhistory, and the New Cultural History, was defined against quantitative history by scholars such as Georges Duby in France, Carlo Ginzburg in Italy, and Robert Darnton in the United States. In short, there has been a gradual move from more or less unself-conscious imprecision to self-conscious antiprecision.