George Bernard Shaw is alleged to have remarked that when a historian had to rely on one document he was safe, but if there were two to be consulted he was in difficulty, and if three were available his position was hopeless. We like to think that Shaw was one-third right and that our position as historians of the Cultural Revolution would indeed be hopeless if we had only three documents to rely on. Here we merely want to say a few words about our far more numerous sources— their provenance, strengths, and weaknesses—and how we have put them to use.
Our starting point, as we began work on this book, was not Mao's famous "blank sheet of paper," but our views and opinions shaped by the events themselves and the existing literature. To test those views and opinions, to refine them with an eye to writing a history of the Cultural Revolution that would make sense not only to ourselves but to our imagined readers, we set about consulting sources of every conceivable kind—and then some. We did this not merely in a search for information, but believing that there is in itself a virtue in using many different kinds of sources, that the inevitable bias in one kind is to some degree offset by the counterbias in another. While texts, as our notes and bibliography suggest, in the end remain our primary sources, we also conducted interviews with members of Mao's inner circle and the CCRG, the speechwriters and assistants of PSC members, and Red Guard leaders of the left and the right. We teased memories out of retired Western diplomats, watched contemporary newsreel footage, pored over old photographs, transcribed tape recordings, and even deciphered inscriptions on the reverse side of Mao-badges.
One source that we have found particularly informative and made much use of is the common (in China) so-called chronology of major events—called a dashiji when concerning institutions broadly defined and a nianpu when documenting the life of a person. Produced in abundance during the Cultural Revolution and in even greater numbers since, chronologies of major events serve as