The Origins of Chess

Edward Winter


The present article comprises the full texts of a discussion in Chess Notes during its final year as a magazine.


Louis Blair (Pittsburgh, PA, USA) writes:

‘Chess gimmicks of various kinds seem to be catching on in the USA. Lately, people have been selling videotapes that supposedly help improve one’s game. Jeremy Silman’s “Chess: A Winner’s Strategy” starts by saying dramatically, “Chess, the world’s oldest board game. It was being played at the beginning of the Persian civilization, six thousand years ago”.’


From V.D. Pandit (Bombay, India):

‘Jeremy Silman’s claim that “chess” was being played at the beginning of the Persian civilization six thousand years ago sounds – leaving aside gimmick – fascinating and absurd.

Though it is now sufficiently proved and accepted that the game of “chess” originated around 570 A.D., the origin of “chaturanga” – from which “chess” is descended – remains obscure.

As a sequel to the partition of India in 1947 all the important sites of the Indus Civilization went over to Pakistan. Hence arose the necessity of finding Harappan (Indus Civilization) sites within the present borders of India. A systematic survey of the Ghaggar (Sarasvati) Valley in the North and the Kathiawar peninsula in the South was undertaken by the Archaeological Survey of India between 1953 and 1958. The exploration resulted in the location of more than two dozen Harappan sites by Dr A. Ghosh in the Bikaner Division of Rajasthan and nearly 100 sites by Dr S.R. Rao and Dr P.P. Pandya in Gujarat (including Kutch and Kathiawar). Among them, Lothal is most important, especially because it was a port city contemporary with Harappa and Mohenjo-daro.

The site was discovered in 1954 and excavated on most scientific lines between 1955 and 1962. Lothal is situated at the head of the Gulf of Cambay at a distance of 80 kilometres south-west of Ahmedabad.

A fact of particular pride to chess lovers is that Dr S.R. Rao discovered some “chaturanga” pieces during excavations at Lothal. With carbon tests they are dated 2,500 B.C.

This proves that “chaturanga” was in vogue in India some 4,500 years ago. Jeremy Silman can use this figure for his videotape, and I am sure it will not reduce his/its gimmick in any way.’


From Bob Meadley (Narromine, NSW, Australia):

‘I should like to respond to V.D. Pandit’s remarkable letter in C.N. 1832, in which as a few throwaway lines he gives chess, or certainly its parent chaturanga, a birthdate of 2500 B.C. I am fairly sure that not many “chess lovers”, as Mr Pandit calls us, were aware of this find at the city of Lothal above Bombay. What Mr Pandit has done in C.N. is blow up completely H.J.R. Murray’s view of the birthdate of chess being 455-543 A.D. (page 1 of A Short History of Chess). Richard Eales, on page 17 of his book Chess The History of a Game, wrote: “The four-handed game certainly existed in India, but as there are no references to it earlier than the eleventh century it was more probably an offshoot of the two-handed game rather than the other way round.” It seems that Eales too did not know of Lothal. In an excellent summary on page 16, though, he gives information about Duncan Forbes’s views, which were laughed out of court when Murray’s A History of Chess came out in 1913. To quote: “Forbes reverted to the Indian origins of chess and to the ordinary board of 64 squares, but he held that ‘original’ chess was a game for four players, an idea first advanced by Captain Cox at the beginning of the century.” Forbes dated the four-handed game at 3000 B.C. and now through Lothal and radio-carbon dating, “chess” pieces are found dating to 2500 S.C. An absolutely stunning find, and it would be of interest to many chess enthusiasts if Mr Pandit could give us further information through C.N. on these carbon-dating tests – as to when they were carried out and their margins of error (if any). Certainly a photograph of these pieces and information on where they are today, together with some further informed analysis, is crying out for publication. Congratulations yet again on a scoop rivalling the Morphy and Sarratt letters, and one which, in my view, surpasses them.’

Hugh Myers (Davenport, IA, USA) writes:

‘In C.N. 1783 Louis Blair quoted J. Silman as saying that chess is the world’s oldest board game” and that it “was being played” in Persia circa 4000 B.C. Mr V.D. Pandit calls the latter idea “absurd”. I would add that such statements as Silman’s are unconscionable if they are made without presentation of proofs of error in accepted historical records.

There is written evidence of chess from the sixth century A.D. I don’t know of anything earlier. Allegedly older chess pieces have been found – although I haven’t heard claims for anything before the second century A.D. (except mistaken ones for pieces of Egyptian games which were not chess) – and they don’t prove anything. There are questions both of dating and of identification of the pieces.

There is plenty of substantial evidence that other board games are older. Boards for one were found in tombs at Ur, dated circa 3000 B.C. They were not chess boards, and we don’t know the rules of the game. Of games that are still played, however, Nine Men’s Morris (The Mill, or merels) goes back to at least 1400 B.C., and Wei Chi (Go) was played in the seventh century B.C.

Mr Pandit gives factual information about the important 1950s-60s Indian excavations of sites of the Harappa/Mohenjo-daro culture. I’ve been interested in archaeology as well as chess history, and in 1971 I first read about Mr S.R. Rao’s excavations at Lothal, where Mr Pandit says there were “chaturanga pieces” dated to 2500 B.C.

My understanding is that there were two Lothal cultures, Lothal A (2500-1500 B.C.) and Lothal B (1500-1000 B.C.). Lothal A was called “highly organized”, a port with a brick dockyard which was certainly ahead of its time. But it would be extraordinary if the only “chaturanga pieces” found there would be dated back to the town’s very beginning.

What is a “chaturanga piece”? What do those found at Lothal look like, and why were they so identified? In fact, what does Mr Pandit consider “chaturanga” to have been?

When Mr Pandit states that chess is “descended” from chaturanga, and that the supposed Lothal chaturanga pieces prove “that chaturanga was in vogue in India some 4,500 years ago” he is inviting the same description of “absurd” that he applied to Mr Silman’s statement.

Why? Because he is simply ignoring the conclusions of H.J.R. Murray, whose 1913 A History of Chess remains the basic source for chess history, and those conclusions should either be accepted or individually refuted. They should not be ignored.

Mr Pandit’s statements are nothing new. They are essentially the old Cox-Forbes theory (i.e. that chaturanga, a four-handed chess game, played with dice, goes back to circa 3000 B.C.). That theory was discredited by Murray. There is no proof of it. “Chaturanga” was used long ago as a word for army. The 64-square board was used long ago for a race game played with dice or cowries; it was nothing like chess. When chess was put on the 64-square “ashtapada” board it was logical to call it “chaturanga”, the army game. Written proof of a four-handed chess game, or chess played with dice, wasn’t seen until well after the sixth century A.D.

Anyone who wants to prove otherwise should refer to Murray’s History, and prove that he was mistaken.

There have also been claims that chess originated in China or Central Asia. My opinion is that so far no-one has proven that Murray was wrong when he said evidence indicates that it started in India in the sixth century A.D. Mr Pandit knows that. Our difference is my opinion that the chess of the sixth century A.D. was chaturanga, and that prior to that time there was no such thing as chaturanga except for non-chessic use of the word.

I wrote in more detail about early chess history in issue 34 of the Myers Openings Bulletin.’


From V.D. Pandit:

‘Hindu mythology believes that the game of Chaturanga – of which, I repeat, chess is the descendant – was invented by Mandodari, the Queen of King Ravana of Lanka (now Sri Lanka) to amuse her ever ambitious and warring husband. The recent excavations at Lothal (Gujarat) lead us to think that after all this does not seem to be a mere myth.

The first version of Ramayana (basically an epic of war between King Ravana and Lord Rama wherein Ravana was demolished by Lord Rama) was composed during 800 B.C. But scholars differ about the period of Ramayana. In any case, it is, however, older than 1000 B.C. But the pieces found at Lothal are from 2500 B.C. It seems possible then that the prosperous and civilized Dravidians, who ruled India, knew the game and Mandodari, a Dravidian Queen, did not actually invent the game but taught Ravana the game already in vogue. Newly wed Mandodari was not happy with Ravana’s habit of frequently going out on wars leaving her behind all alone. With a view to promoting acquaintance between the two, she taught him the game.

However, any newly-wed bride will not invent a game wherein two “intruders” will be required. (Chaturanga is played among four players, but can also be played between two.) So it is more logical that Mandodari knew this game and after seeing Ravana’s thirst for warfare she introduced the game to him so that he might be induced to curtail his war activities and to spend more time with her.

Mandodari should then perhaps be given credit for discovering not only Chaturanga but the Simulation Technique as well (which is so widely used these days in Operations Research).

As ancient Bharat (Hindustan = India) was the base for Chaturanga (and chess), I strongly feel that anybody who does not possess thorough knowledge of Sanskrit language and Hindu philosophy will be hampered in his task of writing an authentic history of chess.’

Hugh Myers replies:

‘After being told that I shouldn’t criticize a system for ranking chessplayers because I’m not a mathematician, now I’m told that knowledge of Sanskrit and of Hindu philosophy are prerequisites for writing about chess history.

My response is that Mr Pandit’s ignorance of scientific method disqualifies him from doing it. He proves nothing by making flat statements, no matter how emphatically he does it, such as “the game of Chaturanga – of which, I repeat, chess is the descendant”. He didn’t answer C.N. 1870’s questions: what do those alleged chaturanga pieces of 2500 B.C. look like? How were they identified? How were they dated? He still ignores H.J.R. Murray’s work. He talks about “prosperous and civilized Dravidians” and Sanskrit. What does any of that have to do with chaturanga in the Harappa culture of 2500 B.C.? Human remains from then are multi-ethnic; some are proto-Australoid or “Dravidian”, but mostly they are “Mediterranean” (Afghan types). Sanskrit, an Indo-European language, was brought to India by the Aryans who destroyed the Harappa culture many centuries after 2500 B.C. Hindu mythology is important and it deserves respect, but it can’t prove the age of chess. It was an oral tradition until modern times. Mr Pandit tries to back up his theories with Ramayana ... but: “An epic can be altered, brought up to date, recast and adapted to suit the changing tastes in heroic literature – the Indian epics, the Mahabharata and the Ramayana, are notable examples of this, their extant forms going back no further than the first few centuries A.D.” (Stuart Piggott in Prehistoric India). That’s a lot more recent than 2500 B.C. There is no documentary proof for Mr Pandit’s theory.’

A contribution from Bob Meadley:

‘The Lothal “chess” pieces are the most important historical item to be publicized in chess literature. It would also seem that these pieces have been radio-carbon dated, and if the dates as given are in any way sound, then the accepted evolutionary theory of chess is incorrect.

Richard Eales in his book Chess The History of a Game (page 19) placed the earliest texts relating to “chess” like games at 600 A.D. and he went on to write: “Before that, there is only archaeology, and conjecture ...”

There are a number of “chess” pieces that urgently require radio-carbon dating to fill in present archaeological finds. For example:

1. The Dalversin-Tyube find of an elephant and a zebu piece (ivory). Reputed to be second century A.D. (See Linder, Chess in Old Russia, page 18.)

2. The Venafro pieces as publicized by Chicco and Sanvito. Reputed to be third century A.D. (See the Dizionario enciclopedico degli scacchi, plate opposite page 370.)

3. The San Sebastián catacombs as publicized in 2. Reputed to be fifth/sixth century A.D. (Source as 2, same page.)

4. The “glass paste” pieces in the Cairo Museum. (See Eco Scacco, October/November 1978, by Chicco.) Reputed to be 1500 B.C. It is doubtful that these can be radio-carbon dated, but an examination cries out to be made.

As Richard Eales rightly puts it, “archaeology and conjecture”. We need to take steps to date the above four finds and then do some theorizing on the origin of chess. For example, number four above looks absurd. Place it alongside the Lothal pieces and it looks logical. With respect to all chess historians, we all seem to love a good debate at the expense of archaeological finds. There may well be no written evidence earlier than 600 A.D., but that does not mean that chess or its parents cannot be tracked back further as suggested.

I suspect that criticism of H.J.R. Murray lies at the base of the problem. His great work seems to be built on the premise that chess was invented in the sixth century A.D. Yet in the whole book this is probably the weakest link in his conclusions simply because he dismissed archaeological evidence. Even his later opinions almost 40 years after the great work was published in 1913 brooked no debate. Yet sometimes he was wrong. On page 86 of The Book of Chessmen, Alex Hammond, admittedly in hearsay, wrote that H.J.R.M. considered that the Lewis chessmen were from the years 1600-1650 A.D. Yet Michael Taylor, in a brilliant 1978 booklet of analytical deduction (sadly no radio-carbon dating) called The Lewis Chessmen, published by the British Museum, concluded that they dated from the period 1135-1170 A.D., almost 500 years earlier than Murray.

I would like to suggest to the chess public that an eminent body exists which could take steps to solve some of the above problems. I am in contact with one of its prominent members, David Hafler of the USA. The body is called “Chess Collectors International”, an organization devoted to the study and dissemination of chess art and history. It is one of the few chess associations that could bring pressure to bear on the various museums that hold some of these pieces in the spirit of pure research to try to bring some conclusions to this “talk-fest”: when and where did chess originate? Perhaps in the dying issue of C.N. some life can be breathed into this important issue.

We regret that we have received no further contribution from Mr Pandit. For other material on this debate, see the Quotes and Queries column in the December 1989 BCM.


Placed online on 24 February 2019.

To the Chess Notes main page.

To the Archives for other feature articles.

Copyright: Edward Winter. All rights reserved.