Wu Zetian and the Trouble With Hostile Sources

Wu Zhao — known to history by her posthumous title, Wu Zetian — is the only woman to have ruled China as emperor in her own right. She took the title in 690 under a new dynastic name, Zhou, and held it until a palace coup forced her abdication in 705. Almost everything narrative we know about her was written by men, centuries later, inside a tradition that regarded a female sovereign as a category error. That makes her one of the best cases in world history for a skill every reader of biography needs: how to use a source that is clearly against its subject without either swallowing it or flipping it.

The record, and when it was made

Start with the dates of the documents rather than the dates of the events, because that comparison does most of the work.

Wu governed in the second half of the seventh century. The principal narrative sources are the Jiu Tangshu (Old Book of Tang), compiled in 945; the Xin Tangshu (New Book of Tang), completed in 1060; and Sima Guang's Zizhi Tongjian, the great chronological history finished in 1084. The earliest of those is roughly two and a half centuries after her death. The latest is nearly four.

None of them is a neutral chronicle. The Zizhi Tongjian in particular was written as an instructive work for rulers — its title translates as something like "comprehensive mirror in aid of governance" — and its purpose shapes its selection. Scholarship on Sima Guang's project treats it as a work with a historical argument, not a container of facts, which is the right way to read any great compilation.

There is also a second-order problem. Compilers of later dynastic histories worked from earlier state records, court diaries and veritable records, most of which do not survive independently. So we are usually reading an eleventh-century editor's use of a tenth-century compiler's use of records that have since been lost. That is a long chain, and the general principle described in primary and secondary sources applies at every link in it. The broader framework for judging this kind of claim about a person is laid out in how we know what historical figures actually did.

What the hostility looks like in practice

The most notorious accusation against Wu is that she killed her own infant daughter and arranged for the blame to fall on the empress she was working to displace. It is repeated constantly in popular retellings. It reaches us through the later compilations rather than through any contemporary document, and the accounts do not agree with one another.

That combination — a charge that is most vivid in the latest telling and absent or unclear in the earliest surviving one, with no contemporary corroboration — is the signature of an accretion rather than a report. Historians do not therefore declare her innocent. They say the evidence cannot bear the weight the story puts on it, which is a different and more honest claim.

The same caution applies to the general picture of a court soaked in poisonings and purges. Purges certainly happened; the political violence of the period is well documented and she was a participant in it. But a hostile tradition tends to convert ordinary factional brutality — which every Tang reign had — into a specific personal pathology when the person at the centre is a woman. Modern Tang scholarship, including Andrew Eisenberg's work on factional politics in Gaozong's court, reads her rise as an episode inside a struggle between court factions rather than as an eruption of individual character.

Reading a hostile source without inverting it

The tempting correction is to reverse every negative claim and produce a feminist hero. That is a second distortion, and it borrows its structure from the first: it still treats the source as a reliable guide, just with the sign changed.

Three moves are more useful.

Separate what the source had no motive to invent. Institutional details, appointments, dates, building projects, grants and titles are usually recorded because a bureaucracy recorded them, not because they make anyone look good or bad. They survive comparatively well.

Ask what the charge is doing. In a Confucian dynastic frame, disorder at the top requires a moral explanation. A female emperor is disorder by definition. So the tradition needs her to be monstrous, and the stories that do that job are precisely the ones that grow.

Look for material evidence that is not narrative at all. Inscriptions, sculpture, coinage, temple foundations and administrative documents were produced at the time and by different hands. Wu's patronage of Buddhist art is a case in point: the Longmen Grottoes outside Luoyang, which UNESCO describes as holding the largest and most impressive collection of Chinese stone carving from the late Northern Wei and Tang periods, contain work from her era that art historians have studied as evidence in its own right. Sculpture does not moralise about its patron.