Did the Tui Bei Tu really foretell Chinese history?

What the Tui Bei Tu claims

The Tui Bei Tu (推背圖), or Back-Pushing Chart, is a Chinese prophecy text traditionally credited to two Tang dynasty court astronomers. It pairs a small painted scene with a four-line verse for each of roughly sixty turning points in history, and generations of readers have matched individual verses to events from the fall of the Tang dynasty onward. The claim does not hold up under scrutiny. The earliest official record of the book names no author, a Song dynasty emperor had its sequence scrambled on purpose, and no two surviving copies agree on how many images the book contains or what their verses say.

Where to read the Tui Bei Tu

There is no single original to read. Surviving copies disagree on the number of images and on the verses. One manuscript is held by the Fu Ssu-nien Library of Academia Sinica in Taiwan, which describes it online, and a widely circulated version is transcribed on Wikisource (both are in the sources below).

Where the Tang dynasty attribution comes from

Tradition credits the text to Li Chunfeng (602-670), a Tang court astronomer versed in astronomy and yin-yang theory, and Yuan Tiangang, a physiognomist from Chengdu. Both men also appear in stories about foretelling the rise of Wu Zetian. The Fu Ssu-nien Library at Taiwan's Academia Sinica, which holds one manuscript copy, retells the book's origin story: Li kept writing prophecies until Yuan pushed him from behind as a warning not to reveal too much of heaven's secrets, which gave the work its name.

That origin story is a later legend rather than a contemporary record. The oldest official listing of the book, in the bibliography chapter of the official History of Song, names only a title, Tui Bei Tu in one scroll, and records no author at all, according to a 2003 catalog study by the Sinological Institute Library at Leiden University. Centuries separate Li Chunfeng's Tang dynasty career from the first surviving mention of a book that would later carry his name.

What the images and verses look like

Each numbered plate in a typical copy pairs a small ink-and-color scene with a four-line, seven-character verse written in riddling, allusive language: split characters, cyclical zodiac terms and oblique references to rulers rather than plain statements. In the Academia Sinica copy, one opening pairs the death of the imperial consort Yang Guifei at Mawei Slope with the Huang Chao rebellion, later Tang events that readers count among the book's fulfilled prophecies.

The Song dynasty ban and the reshuffled order

According to the Southern Song writer Yue Ke's Cheng Shi, quoted in the Leiden library's 2003 study, the founding Song emperor, Taizu, banned prophecy books soon after taking the throne, fearing they would mislead the public. Copies had already circulated for centuries, and too many households owned them to confiscate. When a minister reported that prosecutions for owning them were piling up, the emperor changed course: officials kept one verified copy for the court's own reference, scrambled the sequence of the rest, mixed them with other writings, and let roughly a hundred garbled versions circulate alongside the handful of genuine ones.

The plan worked as intended. Readers could no longer tell which copy, if any, preserved the original order, and the study notes that later owners largely stopped trying to verify or preserve authentic versions. Every copy that survives today descends from that confusion, which is one reason no two editions number their images the same way.

Why no two copies agree

A 1973 survey by the German sinologist Wolfgang Bauer, cited in the same Leiden study, compared seven hand-copied manuscripts held in Tokyo, Berkeley, Harvard and Taiwan, and found they contained anywhere from sixty to sixty-seven images each. The Leiden library's own three copies compound the problem: one, once owned by the Dutch sinologist J.J.M. de Groot, has sixty-five images; a second, from the collection of the diplomat and novelist Robert van Gulik, has sixty-six; and even where two copies number a picture the same way, the wording of its verse is not always identical.

How rulers and rebels used it

The book's political life continued long after the Song. A 2021 paper in the Shanxi Normal University Journal on the late Qing and Beiyang periods describes officials and rebels reading the same vague verses to fit their own moment: supporters of the 1911 Revolution, officials choosing reign names for Yuan Shikai's short-lived monarchy, and ordinary people trying to judge whether Chiang Kai-shek held the mandate of heaven. Shanghai publishers marketed the book aggressively despite repeated bans, shaping public opinion when ordinary readers had little other access to information about decisions made at the top.

Why the "hits" don't hold up

The Leiden study's own summary of the verses is blunt: they are equivocal and ambiguous in meaning, forecasting the rise and fall of unnamed rulers without pinning down a date, a name or a place. Verses that vague can be matched to almost any later event, and generations of copyists and commentators had centuries to add, drop or rewrite lines once they already knew how history had turned out. A prediction that can be fitted to an outcome after the fact is not a foretelling of it.

Questions people ask

Verdict

Real or Legend files the Tui Bei Tu as debunked. A book does exist, its attribution to Li Chunfeng and Yuan Tiangang is centuries old, and it has shaped Chinese political culture since at least the Song dynasty. But the evidence a real prophecy would need is missing: no author is named in the earliest record, the images were scrambled on imperial order within a few centuries of that record, and the surviving copies disagree with each other on how many verses exist and what they say. What remains is a text vague enough, and old enough, to be read into whatever has already happened.