This article discusses the new microhistory of the 1970s and 1980s in terms of the concept of exceptional typical, and contrasts the new microhistory to old microhistory, in which the relationship between micro and macro levels of phenomena was defined by means of the concepts of exceptionality and typicality. The focus of the essay is on Carlo Ginzburg's method of clues, Walter Benjamin's idea of monads, and Michel de Certeau's concept of margins. The new microhistory is also compared with methodological discussions int he social sciences. In the mid-1970s concepts like the micro-macro link or the microfoundations of macrotheory were introduced in sociology and economics. But these largely worked in terms of the concepts of typicality or exceptionality, and this has proved to be problematic. Only historians have developed concepts that escape these and the older definitions of the micro-macro relationship; indeed, the 'new microhistory' can best be described in terms of the notion of 'exceptional typical'. The essay explores the meaning of this notion.