In this article (which is based on a paper delivered to the 1997 symposion of the IGK) I will give an overview on early Chinese hypotheses on the origins of Chinese Chess; we will then inspect references to Xiangqi  in earlier Chinese texts, and refer to other evidence to decide whether or not these hypotheses hold good.

As I regard chess in East Asia a fascinating field in itself, my paramount interest is to delve into the origins, the history, and the development of chess and chess-like games. Although I will refrain from speculations on the connections between Chinese Chess and the Indian and Persian chess-games, the reader is invited to draw whatever conclusions himself.

However, I hope I will succeed in making obvious that Chinese Chess and its forerunners have a sound claim to necessarily be allowed for by anyone dealing with the history of chess.


As it seems to be a time-honoured custom to explain the meaning of the word "Xiangqi" when dealing with Chinese Chess, I will say a few words on the subject as well. Throughout the history of the Chinese writing system, many of the characters were assigned new meanings, lost old meanings or gained additional meanings, and suffered shifts in the semantic contents of words written with that character. These changes may even have taken place in specific areas only, or only for a certain time, or with considerable lags in larger portions of the country. In certain professions such as the military words and characters were given highly specialized meanings. Thus it has become quite difficult to determine what a certain, especially a rare, character might have meant in a given text. It is important to know the background of a given writer to decide whether a specialized meaning of a word or character could have been intended.  The Chinese word "Xiangqi" is written with two characters of whom the first, xang , nowadays denotes 'elephant; portrait; phenomenon; ivory; stellar configuration, omen; acting, playing; official interpreter', the second, q , denotes 'chessman; chess or similar games; foundation'. Q usually refers to the game as a whole (board and pieces), and is often used to write words denoting board-games, as Weiqi 圍棋, Tanqi 彈棋 &c.

To further complicate matters, Xiangqi is not the only possible word to label a chess-game. Just to mention a few possibilities: to signify board-games as a whole or certain board-games the words q , , , y , , b , , x , d , alone or in combination, with perhaps a supplementary d or xao (great rsp. small) were all in use. Since tracking all these references down is near impossible or has failed to provide any satisfying result until now, I will concentrate on occurrences of the Chinese words Xiangqi 象棋 and Xiangxi 象戲.

By the way, as I think that almost all the translations for "Xiangqi" that have been proposed up to now (e.g., Elephant Chess, Ivory Chess, Symbol Chess) are not unreasonable, but the explanations given arent really satisfying. So Id rather stick to Xiangqi or "Chinese Chess".


One might ask whether the xang in Xiangqi has anything to do with the game-piece of that name. It has often been found strange that a relatively unimportant piece might have lent its name to the game. I propose that the name of the game derives from older sources we are going to deal with in a few moments, but that actually something different was intended. In most of the Xiangqi sets we find the "elephants" either marked xang or xang . The second character signifies amongst other meanings chancellor, minister, great councilor, the word was used from earliest times on as the title of a high-ranking official. The two words are homophones from about 600 AD and near homophones from about 1000 BC. Perhaps the Xang in the name of the game refers to the symbolic pieces moving around the board, while the piece Xang / originally was a Grand Minister. So maybe this was forgotten in later times, and the Xang survived only as a variant writing to make it easier to distinguish the pieces in well-used sets of pieces. This is nothing more than an idea yet - but who knows?


In the elder Chinese literature five hypotheses on the origins of Chinese Chess feature prominently The list follows Zhou Jiasen 周家森  and Li Songfu 李松福 . Ordered according to the antiquity they ascribe to Chinese Chess these hypotheses are :


1. An origin in the age of the legendary Shennong 神農 (trad. reigned 2737-2697 BC), as proposed by the Yuan (1206- 1368) monk Nianchang 念常 (1282-1342?) in his "Fozu lidai tongzai 佛租歷代通載 (Buddha in passing generations and all the years)",     x Shnnng y r yu xing chn wi xang Tng xanggo Nu Sengr yng ju m sh z jia po di zhi wi ji y(.) 昔神農以日月星辰為象唐相國牛僧孺用車馬士卒加砲代之為()

"In olden times Shennong used the sun (ri ), the moon (yue ), the stars (xing ), and the planets (chen ) as symbols (xiang ); the Tang - Minister of State (xiangguo 相國) Niu Sengru 牛僧孺 used chariots (che ), horses (ma ), scholars (shi ), soldiers (zu ), and catapults (pao ) to replace these as utensils in the game."


2. An origin in the age of the legendary Huangdi 黃帝 (trad. reigned 2697-2597 BC), the Yellow emperor, as proposed by Zhao Buzhi 晁補之 (1053-1110) of the Beisong 北宋 (Northern Song, 960-1126) in his "Guang Xiangxi ge xu 廣象戲格 序 (Rules for wide Xiangqi: Foreword)" ,

Xangx bingx y Hungd zhi zhn qu mngshu ywi zhn xang shu zhi xong y g x bing ery xangx mng zhi (.) 象戲兵戲也黃帝之戰驅猛獸以為陣象獸之雄也故戲兵而以象戲名之()

"Xiangxi 象戲 is a game of strategy; Huangdi in his wars used fierce animals in his battle array; as Elephants (xiang ) are the strongest of wild animals, the game is called Xiangxi after this strategy."

Shennong and Huangdi are two of the Chinese cultural heroes who taught the Chinese, Xiangqi were thus ascribed an origin in the third millennium BC, if we placed these cultural heroes before the first dynasty.


3. An origin in the age of Zhou (1122-249 BC) Wuwang 武王 (reigned1122-1115 BC), in the time of this last campaigns against the tyrant Shang Zhou (reigned 1154-1122 BC), as proposed by Ming - (1368-1644) time Xie Zaihang 謝在杭 (=Xie Zhaozhe 謝肇淛, 1567-1624) in his "Wuzazu 五雜俎 (Investigations on the five categories of things)" , thus ascribing Chinese Chess an origin in the late 12th century BC.

Xangq xiangchun wi Wwng f Zhu sh zo j brn y zhngo bingjiazh lu gi sh yu zhng chezhn y 象棋相傳為武王伐紂時作即不然亦戰國兵家者流蓋時猶重車戰也

"Xiangqi, according to tradition made by King Wu of Zhou in the time of his final campaigns against Shang; if that is not so, at least it became popular among military personnel in the time of the contending realms, as in this time chariot warfare was still important."


4. An origin in the time of the contending realms (475-221 BC). This was proposed in Hu Yinglins 胡應麟 (1551-1602) Bicong 筆叢 (Brush notes, a kind of essays) , and the "Qianqueju leishu 潛確居類書 (Encyclopaedia of hidden and real conditions)" , which was compiled by Chen Renxi 陳仁錫 (1581-1636), thus ascribing Xiangqi an origin in the third century BC,

Yong Mnzhou wei Mngchngjun zxa yn ju z du Xangq y Zhn'go zhi sh y gi Zhn'go yng bing g sh rn yng zhnzheng zhi xang wi qsh y 雍門周謂孟嘗君足下燕居則鬥象棋亦戰國之事也蓋戰國用兵故時人用戰爭之象為棋勢也

"Yong Menzhou 雍門周 said to Mengchangjun 孟嘗君: Mylord, if you are at leisure, play Xiangqi; thus it was a thing from the time of the contending realms. Because in the strategy of the contending realms the people of this time used elephants just as in the board game strategy (qishi 棋勢)."

The prince Mengchang mentioned here was a well-known man who lived during the times of the last Zhou-king; thus the admonition quoted here would point to a date in the late third century BC.


5. An origin in the time of Beizhou (Northern Zhou, 557-589) 北周 Wudi 武帝 (reigned 561-578), as proposed in the "Taiping yulan 太平御覽 (Grand mirror of the Taiping era)" (completed in 982) under the heading Xiangqi .

Zhou Wd zo Xangx (.) 周武帝造象戲()

"Zhou Wudi created Xiangxi",

the "Wuyuan 物原 (Source of Things)"  of Ming Meng Qi 夢頎 explains:

Zhou Wd zo Xiangqi (.) 周武帝作象棋()

"Zhou Wudi made Xiangqi".

We should bear these five hypotheses -the Shennong hypothesis, the Huangdi hypothesis, the Zhou Wuwang hypothesis, the contending realms hypothesis, and the Beizhou Wudi hypothesis- in mind while we go on to inspect further textual and additional evidence that might or might not support the hypotheses put forth by these earlier Chinese scholars.


The earliest still extant text in which we find the combination of the characters xang and q is the Chuci 楚辭 (Poems from Chu) , a corpus of poems purportively deriving from the third or second century BC, but finally arranged (and edited?) only in the early second century AD. The poems themselves might indeed have been composed during the later contending realms period. The Zhaohun 招魂 (Calling back the soul) , a long poem by Song Yu 宋玉, contains the characters xiang and qi. Chinese commentators do agree that a kind of game is referred to but they usually assume that liubo 六博 is meant.

B b Xangq yu Lub xie (.) 篦蔽象棋有六簙些()

The castor shrubs hide the Xiangqi, but there still is the Liubo! (or: there it is, the Liubo! (?))

From this sentence it cannot be decided for sure whether one game (Liubo alone) or two games (Liubo and Xiangqi) are meant. If Xiangqi or one of its predecessors were meant, this would point to an origin in the contending realms period.


A somewhat later reference  is found in the Shuo yan 說宛 ('Collection of explanations', 'Collection of persuasions (shu yun 說苑)', 'Garden of Happiness (yu yun 說苑)') that has been composed in the first century BC. It was presented to the throne in 17 BC by Liu Xiang 劉向 (79-8 BC). Here as well we cannot decide what game the text actually alludes to, as we do not have any extra-textual reference. There remain some doubts about the actual translations of this passage, as an inspection of a larger portion of the text makes other meanings not impossible. The translation given here was chosen as it conveys a hint on chess.

r chn y (.) yn z du Xingq r w Zhng n (.) 而諂諛()燕則鬥象棋而舞鄭女() and flatter (.) If you have leisure, then fight at Xiangqi or dance with the women from Zheng(.)

This is in fact the passage quoted in the "Hu Yinglin bicong" and "Qianqueju leishu", the date of origin of the "Shuo yuan" thus backing the hypothesis that Xiangqi stems from the period of the contending realms.

What is noteworthy, anyway, is that the combination of the characters Xang and Q is quite old. Without difficulty we can trace it back to at least the the late century BC. A still earlier origin is not at all excluded definitely. However, we do not know for sure what kind of game was referred to by this word. It seems not improbable that anyone facing the task to name a new game in later times might have been inspired by these passages.


A game with a name similar to Xiangqi can be found in the 6th century AD. This is the Xiangxi 象戲 (symbol game (?)), a board game said to have been invented by emperor Wu (r. 561-578) of the later (Northern-) Zhou ()- dynasty (557-581). This claim is usually backed by quotes from Linghu Defen 令狐德棻 (583- 661) who wrote the (Hou) Zhoushu ()周書 (Book of the (later) Zhou Dynasty), Wei Zheng 魏徵 (580-643), author of the Suishu 隋書 (580-617), and Li Yanshou 李延壽 (612?-678?) writer of the Beishi 北史 (History of the Northern Dynasties, covering the time 386-618). All agree that in 569 Emperor Wu composed the Xiangjing 象經 (Classic of the symbol game) to explain the game. The Zhoushu writes in the annals of the emperors :

w ye d zh Xangjing chng j bi lao jangshuo 五月帝制象經成集百僚講說 "[4th year of the Tianhe 天和-period.] 5th month. The emperor finishes the making of the Xiangjing and gathers his officials to expound it."

The Beishi  reports the matter in identical wording in the entries on the annals of the Zhou emperors. The Suishu reports the matter in the biography section, in the biography of Lang Mao 郎茂 :

sh Zhou Wd wi Xangjing (,) Gaoz cng rng wi Zheng yue (:) rn zh zhi su wi y (,) gn tiand (,) dng gushn (,) r Xangjing duo jiu f (,) jiang hy zh zh 時周武帝為象經高祖從容謂徵曰人主之所為也感天地動鬼神而象經多糾法將何以致治

"When Zhou Wudi made the Xiangjing, the now deceased emperor casually asked Zheng: 'What is the place of the ruler of mankind, should he unify heaven and earth, should he move the ghosts and spirits, like in the Xiangjing with many binding rules, how shall I govern?'"

The existence of this book is further corroborated by its mentioning in the bibliographical chapters of the histories of later dynasties, namely the Sui (581-617) and the Tang (618-907) dynasties. It is interesting to note that the first gives the title among the military works, the second gives the title among the artistic works.


Although the Xiangjing itself is not extant anymore, the preface written by Wang Bao 王褒 (flourished between 552-581) has been handed down to us . From this preface we know that this game was thought to represent phenomena of heaven and earth, the principles of Yin and Yang , the passing of seasons, the eight trigrams, divination, music, filial piety and loyalty, proper rites, the order of government, and orderly conduct. All this is linked in certain schools of Chinese thought. But still it is difficult to perceive how all this could have been represented in a board game. Nevertheless the preface states that there were pieces that were moved on the board, and we are told that military thinking and strategy played a role in the game. Thus we are entitled to count Xiangxi amongst the forerunners of modern Xiangqi.

A somewhat less tangible source on Xiangxi is a poem  ("Xiangxifu 象戲賦, A fu poem on Xiangxi") composed by the general Yu Xin 庾信 (513-581). In somewhat obscure language he appraises how all and everything is displayed in this game in its appropriate proportions. In the letter accompanying the presentation of the poem to the throne he repeats his praise for the emperor who succeeded in representing the order of the world in this game. From its content we can deduce that it must been composed in or shortly after 569.

We find further references to the Xiangjing and the game described therein in the following years. In the Jiande 建德 -Era (572-577) of Zhou Wudis reign a certain Yang Jian 楊堅 writes a book on history and criticizes the game . In the biography section of the Jiu Tangshu 舊唐書 (Old Annals of the Tang dynasty) by Liu Xu 劉昫 (887-946)  it is mentioned that the official Lu Cai 呂才 (biographical data not available) is called in the year Zhenguan 貞觀 3 (629) of Tang Taizongs 唐太宗 reign to explain a quote from a book whose title is given as "Zhou Wudi sanju xiangjing 周武帝三局象經 ("Zhou Wudis three games in the Xiangjing")", but is probably the selfsame Xiangjing. This quote reads: Tizi x m 太子洗馬, literally "the crown prince washes the horses", but in fact meaning "the crown prince shuffles the pieces".


The eldest extant reference to a game that more closely resembles modern Xiangqi can be found in the "Xuanguai lu 玄怪錄 (Tales of the obscure and peculiar)"  by the Tang Minister of State Niu Sengru 牛僧孺 (779-847), a collection of tales of the supernatural. He makes most of his stories appear as if they had previously been orally transmitted.

The content of the story "Cen Shun 岑順"  in a nutshell is that the impoverished scholar Cen Shun takes up to live in an old house that belongs to one of his relatives. In a dream he is commissioned as military advisor by a messenger. The following nights he helps in the defeat of attackers from a foreign kingdom. When his relatives notice that he has changed they entice him to tell what has happened. They dig up the floor of the room he was sleeping in and find a set-up Xiangxi board in an old grave. These occurrences are dated to the year Baoying 寶應 1, that is 762 AD.


Niu Sengru explicitly lists units of cavalry (tianma 天馬), a general (jiang ), chariots (ju ), infantery (jia ), and gives clues to catapults (pao ) and archers (gong ). The existence of these last two types of pieces may be inferred from his mention of arrows and stones flying hence and forth. He does neither elaborate on the number of pieces nor the size of the board, with one exception: he states that the six soldiers (liu jia 六甲) advance in proper array. If we take this as a clue on the number of pawns in the game it becomes clear that this game is not the same as present-day Xiangqi. Furthermore he gives some hints at the move of certain pieces. It is stated that the General moves horizontally into the four (cardinal) directions; the chariots can only advance and do never retreat; the cavalry moves three measures aslant; the pawns move one step ahead. To make it absolutely clear: these moves can be deduced from the text, but not with certainty. The narration goes on to tell of heaps of soldiers rushing east and west, north and south. Thus one could as well come to think that most of the pieces could move into any direction. To distinguish the game described by Niu Sengru from present-day Xiangqi on the one hand and Beizhou Wudis Xiangxi on the other hand this older game is called 'Baoying Xiangqi' in works on Chinese Chess history.


There are a few more references to Xiangxi in the dynastic histories, in literary sketches (bj 筆記), and in poems. From these references we can conclude that Xiangxi was a game not unknown to at least part of the Chinese populace, more specifically the formally educated 'literati'.


In connection with Tang- and Song -time (960-1280) Xiangxi we face a puzzling problem, whose implications I have not yet uncovered fully.  This problem involves the so-called Suzhou zhijn qnqshuhu t 蘇州織錦琴棋書畫圖 ("Silk-brocade picture of qin-lute, game-board, books, and painting-scroll from Suzhou") , which is in unison dated to the period between the late Tang and early Song time, that would be the time of approximately the 10th century. This piece of silk-brocade of whom I have only seen small-size drawings shows a set of the four treasures of the scholar or symbols for the four lesser arts. Playing the qin, playing Weiqi or Xiangqi, the ability to do calligraphy, and the ability to paint are the four arts an allround scholar had to master in addition to his knowledge of the four classics (i.e. the Great Learning (Daxue 大學), the Doctrine of the Mean (Zhongyong 中庸), the Analects of Confucius (Lunyu 論語), and the Book of Mencius (Mengzi 孟子)). There are other sets of four that group together things, books, or modes of behaviour that literati had to master. All the items shown are given in good detail: a five-stringed qin, a tidily rolled painting, a neat stack of books in which the bindings can be seen, and a 8 by 8 black-and-white squares game-board. The qi-games meant are in all instances I know of either Weiqi or Xiangqi. We can definitely exclude Weiqi on the simple reason that the game-board shown simply is not a Weiqi-board. So it ought to have been a Xiangqi- or Xiangxi-board. Naturally this board features none of the hallmarks of present-day Xiangqi: no river, no nine-castle (jiugong 九宮), no symmetrical set-up of pieces on the board. My first idea was that this might be a fake made up in later times, but from the drawings alone I could not disprove anything. The other items shown definitely fit the time to which this piece of silk-brocade is ascribed, ie. the number of strings on the lute, the method in which the books appear to be bound, and the way in which the painted hand-scroll is rolled. Usually in this kind of pictures originality is neither required nor expected, so one could reasonably suppose that quite a usual kind of game-board was shown on that piece of brocade. A somewhat more determined evaluation will have to wait until I know more about this piece of silk-brocade.

According to Li Songfu 李松福  two phrases from Wang Baos preface to the Xiangjing, the phrases

san yue yinyng qi yue ba gu y ding ji wi 三曰陰陽七曰八卦以定其位 " the third is: Yin and Yang, the seventh is: the eight trigrams arranged in their proper position "

and a phrase in Yu Xins Xiangxifu as well possibly point to an 8 by 8, black-and-white Xiangxi-board.

From Song times on textual evidence on a Chinese Chess similar or equal to present-day Xiangqi abound. We do not have only texts but numerous archaeological finds of game-boards and pieces. From all we know the game might already have had its present form. There have even been some variant forms of what became later the main line of Chinese Chess who gained a short-time popularity and then disappeared. These were Sima Guangs 司馬光 (1019-1086) Qiguo Xiangxi 七國象戲 (Seven States Chess), Zhao Buzhis 晁補之 Guangxiangxi 廣象戲, and perhaps the Daixiangqi 大象棋 (Greater Xiangxi).


When we set out, we undertook to inspect five early Chinese hypotheses on the origin of Chinese Chess.  These were 1. the Shennong hypothesis, 2. the Huangdi hypothesis, 3. the Zhou Wuwang hypothesis, 4. the contending realms hypothesis, and 5. the Beizhou Wudi hypothesis.

From lack of textual and archaeological evidence we can exclude the first and second hypotheses, that is, the Shennong and Huangdi hypotheses. These are clearly later-time inventions of the 11th and 12th centuries AD to introduce Xiangqi as an eon-old and time-honoured game, invented by two of the Chinese cultural heroes. The explanations and reasons offered are of such secondary nature that they bear no real value.

In the third hypothesis there seems to be a simple confusion of the well-known Zhou Wuwang, whose ascension to the throne marked the beginning of a new epoch in Chinese history, and the less-known Beizhou Wudi, whose reign brought no real change in Chinese History. It is quite possible that these two were confused, especially if it were true that oral tradition had it that Zhou Wuwang invented Xiangqi. It would have seemed only natural to call someone as important as Zhou Wu an emperor, d , instead of king, wng .

The fourth hypothesis, ascribing Xiangqi an origin in the time of the contending realms, in fact quoting from the Shuo yuan, is backed by the textual evidence from the Chuci, but we still lack extra-textual evidence. Nevertheless, the word Xiangqi is there, and we can assume with some certainty that a kind of game was referred to. Thus we cannot simply discard this hypothesis, even if we still do not know what kind of game was really meant. But since Chinese archaeology is still making progress, and finds of recognizable game-boards and pieces may come our way in the near future, we can hope that there will be undisputable hard facts on early Chinese Chess soon.

The fifth hypothesis as well possibly bears some truth. Even if not a single dynastic history does in fact report that Beizhou Wudi actually made Xiangxi or Xiangqi, it is at least stated that he made the Xiangjing. We could suppose that this Xiangjing was in fact a game and not a book, but there is no need to think so. There are quite a number of additional references to the game and the book, eg. Wang Baos preface, Yu Xins poem &c. Niu Sengru in Tang times can safely describe the Xiangxi-board allegedly found in the grave as "old". If the Beizhou-Xiangxi and the Baoying-Xiangxi were not identical, his description would still point to the fact that Baoying-Xiangxi was not regarded as anything recent.


The question arises whether one of these two games can be reconstructed. At the moment I do not think so, as too many factors are still unknown. While we have source material to enough to speculate on the moves of pieces in the Baoying-Xiangqi, we cannot do so in the case of the Beizhou-Xiangqi. It is still too early to judge on the size and appearance of the boards as we lack data. All reconstructions up to now should be regarded with utter caution. Not a single one of them takes all the evidence known until now into account.


Another unanswered question is which Chinese games might be connected to the early forms of Chinese Chess. We still have to connect Xiangqi and Xiangxi to Liubo 六簙, to Bolosaix 波羅塞戲, and Bosaix 波塞戲 (possibly both Backgammon-type games, but imported from India), to Tanqi 彈棋, and to Lingqi 靈棋. The sources for these games are still not fully analyzed, so preliminary reports, not to mention full-scale histories of Chinese board-games yet remain to be written.