Abstract
Ssu-ma Ch'ien's (145?-86? B.C.) Shih chi (Historical Records) is one of the most influential of Chinese histories, but its organization reflects a historiography quite different from that of traditional Western history. Ssu-ma divided his account of the past into five overlapping sections: basic annals (of dynasties and emperors), chronological tables, treatises, hereditary houses (of the feudal lords), and biographies. One result of this fragmented arrangement is that stories may be told more than once, from different perspectives, and these accounts may not be entirely consistent. From a Western perspective this would seem to indicate a certain disregard for the truth, but in many Shih chi passages Ssu-ma Ch'ien demonstrates a passionate concern for accuracy. In this article I examine in detail one typical set of multiple narrations-the five versions of Wei Pao's defection in c. 205 B.C. - and argue that in some ways Ssu-ma's conflicting accounts reflect the past more accurately than the unified narrative we expect from Western histories. Although Ssu-ma's methods might seem amenable to the constructivist theories of Louis Mink and Hayden White, in the end this type of analysis is inadequate to explain a work which is rooted in a non-Western tradition of historiography. Ssu-ma Ch'ien's own conception of history recognized the limitations of historians and evidence, held out the possibility of multiple interpretations, and focused on moral insight. It is a mix unfamiliar to Westerners, but it does provide a coherent picture of Ssu-ma Ch'ien's historical methodology, and it may serve as an interesting example for modern historians who seek to escape traditional modes of historical writing.