Who Invented Xiangqi? Legends, Evidence and History

The legends

Chinese writers have credited the game to a series of famous names. A Yuan-dynasty Buddhist monk, Nianchang, wrote that the mythical ruler Shennong made a game that used the sun, moon, stars and planets as symbols, and that the Tang chancellor Niu Sengru later replaced them with chariots, horses, scholars, soldiers and catapults. The Northern Song scholar Chao Buzhi traced the game to the Yellow Emperor, who was said to have driven fierce beasts into battle, and suggested the game took its name from the elephant, the strongest of them.

Other traditions point to King Wu, founder of the Zhou dynasty about three thousand years ago. The Ming writer Xie Zhaozhe recorded that story while admitting the game might only go back to the Warring States period. A folk tale even names Xiang, the idle half-brother of the sage-king Shun, as the inventor.

The legend best known to players involves Han Xin, the general who helped Liu Bang win the war between Chu and Han. One folk story says Han Xin, imprisoned by Empress Lü, devised the game to play with his jailers, with the Hong Canal as the border between the armies. The Chu River and Han Border inscription on the board keeps that period in view, but there is no early evidence for the story.

Historians are blunt about these accounts. The German chess historian Peter Banaschak, who surveyed them, concluded that the Shennong and Yellow Emperor stories were later inventions of the eleventh and twelfth centuries AD, meant to give the game an ancient pedigree, and that the King Wu story probably confuses him with a much later ruler, Emperor Wu of the Northern Zhou.

An old word for other games

Part of the confusion comes from the name. The characters 象棋 appear in Chinese texts more than two thousand years old, including a poem in the Chu Ci anthology and the first-century BC collection Shuo Yuan, where an adviser tells Lord Mengchang of Qi that in his leisure he might fight at xiangqi. Neither text says what the game was. Commentators have usually read the Chu Ci line as describing ivory pieces for liubo, a game played with throwing sticks, and in early Chinese the character 象 could simply mean ivory.

In AD 569 Emperor Wu of the Northern Zhou completed the Xiang Jing, a manual for a game called xiangxi. The surviving preface says the game represented heaven and earth, yin and yang, the seasons and the eight trigrams, and later sources mention shuffling the pieces. The chess historian H. J. R. Murray concluded that this was an astronomical game and that chess in China later took over its name.

The Tang evidence

The first text that clearly describes a war game with chariots, horses and a general is a ghost story. Niu Sengru, a Tang-dynasty chancellor of the ninth century, is credited with a collection of supernatural tales called Xuanguai lu, a title rendered in English as Accounts of Mysteries and Monsters. The original book is lost, and its stories survive in later compilations. One of them, the tale of Cen Shun, is set in the first year of the Baoying era, AD 762.

In the story, Cen Shun, a poor scholar, moves into a disused mountain house belonging to his mother's family. At night he hears war drums and sees armies of armoured horsemen only a few inches tall, preparing for a battle between a kingdom called Golden Elephant and its enemy. A military adviser recites a verse that explains the moves. The heavenly horse flies slanting and stops after three, the general moves sideways, the chariot drives straight ahead and never turns back, and six armoured soldiers advance in order.

Eventually his relatives dig up the floor of his room and find an ancient tomb. Inside is a game board set with pieces of gold and bronze, and it becomes clear that the adviser's verse described the moves of xiangxi. Historians call the game in this story Baoying xiangqi, after the era in which it is set.

It is not the modern game. Banaschak points out that the six soldiers, one more than in xiangqi, show this is not the present-day game, and that the moves can be inferred from the text but not with certainty. No cannon is named outright, although flying arrows and stones in the battle might hint at catapult or archer pieces. Even so, he regards Niu Sengru's tale as the first real source on Chinese chess.

The chaturanga connection

Most chess historians place xiangqi in the same family as chess, shogi, janggi and Thai makruk, all descended from chaturanga. That Indian war game is first clearly recorded in the sixth century AD, during the Gupta period, and its name refers to the four divisions of an army: elephants, chariots, cavalry and infantry. Murray set out this view in his history of chess in 1913, and it remains the predominant opinion.

Some of the evidence sits on the board. Xiangqi's elephant moves exactly two points diagonally, the same move given to chaturanga's elephant, although the xiangqi piece can be blocked and cannot cross the river. Playing on the intersections rather than inside squares is usually explained as a borrowing from Go, which was already ancient in China.

A minority of scholars argue the reverse, that a Chinese game came first and travelled west. The best-known modern version is David H. Li's book The Genealogy of Chess (1998), which claims Han Xin developed the game in the winter of 204 to 203 BC from the older game liubo. Banaschak's review concluded that its main hypothesis rests on virtually nothing, and Chinese scholars such as Yang Yinshen and Zhang Ru'an have pointed out that liubo, the Northern Zhou game and the Tang game differ greatly in rules, pieces and boards.

The Song-dynasty game

Evidence becomes plentiful under the Song dynasty, from 960. Banaschak notes that from Song times on there are many texts and archaeological finds of boards and pieces, and that the game may already have had its present form. Song writers also experimented with it. The statesman Sima Guang devised a Seven States version, and Chao Buzhi wrote rules for an enlarged game.

By the end of the Northern Song and into the Southern Song, roughly the twelfth and thirteenth centuries, Chinese sources describe the familiar arrangement, with the general confined to its nine-point palace. Surviving Song pieces include a ceramic set held by the Luoyang Museum and bronze pieces shaped like coins, marked with characters for the general, advisor, elephant, horse, cannon, chariot and soldier. The Southern Song encyclopedia Shilin Guangji, compiled by Chen Yuanliang, contains the two oldest complete recorded games and the oldest known composed problem, Two Dragons Leaving the Sea.

The cannon deserves a note of its own. Its older character, 砲, has the stone radical, and when the piece was added it referred to a stone-throwing catapult. Banaschak translates it that way, and both English and Chinese encyclopedia entries gloss it the same way. The character 炮, with the fire radical and now used for one side's cannon, carries the later gunpowder meaning. Both are pronounced pào.

What historians agree on

Strip away the legends and a fairly clear outline remains. There is no known inventor. Words that look like xiangqi are old, but they referred to other games or simply to ivory pieces. A chess-like game with a general, chariots and horses is first described in a Tang story set in 762. Most historians connect that game to chaturanga, while a minority argue for a purely Chinese origin. And by the Southern Song the game had essentially reached the form played today, river, palaces, cannons and all.

What the legends do show is how completely the game was absorbed into Chinese culture. People wanted it to belong to their oldest heroes, and the Chu and Han names written across the river are a lasting reminder of that wish.

The board that took shape under the Song is the one on XiangqiPlay today. You can browse 5,000 engine-proven mate puzzles, or create a free account and play the computer at any of its eight levels, from Easy to God.