The History of Chess. H. J. R. Murray
Pawn has been lost (29), and probably also, though not explicitly stated, until the masterpiece of the file has been lost. Probably in such a case it is debarred from moving to the 8th rank. Promotion is to the rank of the master-piece of the file (25). But when a player has lost all his superior men save his Boat and one Pawn he may promote this Pawn on any square of the opposite edge to the rank of any piece, King included (30, 32).
The four-handed game would appear to have been played chiefly in Bengal, the North-West Provinces, and the Punjab. Sir William Jones’s authority, the Brāhman Rādhakant, told him ‘that the Brahmans of Gaur or Bengal were once celebrated for superior skill in the game, and that his father, together with his spiritual preceptor Jagannāth, now living at Tribeni, had instructed two young Brāhmans in all the rules of it, and had them sent to Jayanagar at the request of the late Raja, who had liberally rewarded them.’6
According to Raghunandana the four-handed dice-game was chiefly played on festivals like that of the full moon, when it is occasionally incumbent upon the worshippers to keep watch throughout the night. He states that on these occasions it was customary to relieve the tedium of the night with games of dice, and specially with chaturājī. I know of no living authority who has seen this game so played. None of the modern Indian chess-books which I have consulted mention the game as a living variety of chess, and the two which make any reference to it at all have obtained their knowledge of it from European works, and only include it for its historical interest. The Hindu Ram Chandra Pradan, in reply to questions from v.d. Linde in 1874 (v.d. Linde, i. 79), had never heard of this dice-game and declined to believe in its possibility.
On the other hand, a four-handed game of chess played without dice is still played in India. Ram Chandra Pradan told v. d. Linde that he had often seen this non-dice form played. The opposite players were partners, and chessmen of only two colours were used. It has been seen more recently in the Punjab at Naushahra, near Peshawar. Mr. J. Cresswell, who has recorded the fact,7 was shown the game at the conclusion of an ordinary game of chess which he had been watching. Three of the players were Muhammadans, the fourth a Hindu. They used the ordinary chessmen, dividing each colour between the allied players, and using the Farzīus (Counsellors, ‘Queens’) to supply the places of the two extra Kings required. The partners sat opposite one another, the game was played without dice, and there was no wager on the result, nor any value attached to the prisoners taken. He was informed that the game terminated
(1) when one side succeeded in capturing both of the opposing Kings;
(2) when one side succeeded in capturing all the opponent’s men excepting the Kings;
(3) when all four Kings were left bare; in which case the game was drawn.
On this occasion there was no exchange of captured Kings, no attempt to capture the partner’s King, and no promotion of Pawns was necessary. In the Autumn of 1909 I met a young Punjabi from Lahore who was in this country for purposes of education. He told me that shaṭranj was played in Lahore either as a two-handed or as a four-handed game; the two-handed game was the more usual.
Although these modern authorities speak of the use of the ordinary chessmen of the two-handed game being used, special sets for the four-handed game are not unknown. Mr. Falkener possessed a fine set in two colours, in which the Rooks are Boats, and has given a photograph of it in his Games, &c. (facing 119).
The modifications in the method of play which Mr. Cresswell describes appear to be natural ones after the removal of the dice and the abandonment of the method of scoring based upon the numerical values attached to the pieces taken. The game has gained in strategy, and the alliance between the partners is now straightforward. There is no longer any point in capturing the partner’s King, and each side can devote its entire energies to the task of winning without fear of treachery. Rules for Pawn-promotion probably exist, but from the nature of the game they can only seldom come into operation.
This is the game which in the Cox-Forbes theory is the primitive chess. Forbes discovered the seed from which our chess was to spring in the privilege that a player who gained his partner’s throne henceforward secured the sole conduct of the two armies. He considered that this manœuvre was an object of prime importance, and that it would often happen ‘that after some 20 or 30 moves, the contest remained to be concluded between two players only’. Moreover, he finds the use of the dice not only alien to the spirit of the game, but forbidden by the rigid law and religion of the Hindus. It is a small step to imagine that two players often sat down to chaturājī, and played it from the start without using dice at all. To unite the allied armies of red and yellow along one edge, to move the allied armies of black and green from their respective sides to the other edge, to replace two of the Kings by Viziers, are changes which appeared to Forbes with the advantage of the knowledge of the two-handed game, simple, obvious, and natural.
I feel bound to differ. Quite apart from the historical difficulties narrated in Chapter I, which appear to me to be insuperable, the transformation so glibly described seems to me unnatural, unlikely, and incredible. The value of the manœuvre by which the third and fourth players are eliminated seems exaggerated so long as the moves are dictated by dice, and the possibility of its successful accomplishment is much smaller than Forbes imagined. It will take a King seven moves at least to reach his partner’s throne, and he must move right down the front of the two opposing armies, exposed the whole way to attack and possible capture. The probability of seven fives turning up in the first 20 or 30 throws is extremely small. Again, undue weight is laid upon the religious and legal ordinances against the use of dice. Nothing is more certain than the continuous existence of gambling in India from the earliest times, and the two divinities, Siva and Parvati, are often depicted playing a dice-game. The theory of the final transformation I leave, as I believe it condemns itself.
APPENDIX
ATTEMPTS TO RECONSTRUCT THE FOUR-HANDED GAME
Of the two old descriptions of the game, that of al-Bērūnī contains most information as to the practical play, the Bengali poem being mainly concerned with advice to the player as to the considerations which should guide him in making captures or exchanges, and with a description of the different values of the various forms of victory. The rules governing the division of the stakes need not detain us now, except in so far as they suggest aims to be kept in view through the game, since they do not affect the broad question as to how the game was played. Both authorities agree in the initial positions of the forces, and in the moves of the pieces and the interpretation of the throws of the dice.
5 (including 1) K or P moves.
4 (including 6) Elephant (with move of our R) moves.
3 Horse (with move of our Kt) moves.
2 Boat or Rook (with move of Elephant in diagram no. 1, p. 59) moves.
Turning to al-Bērūnī’s account, we notice that he speaks of the use of two dice, though he does not explain how they were to be used. It only appears incidentally from his note on the Rook’s move that the dice are to be thrown simultaneously, although this would of course be the natural conclusion one would draw in any case. Nor is the method of interpretation of the throws at all clear. There would seem to be five possible ways of using the throws. These are: (1) The sum total of the pips might be taken and interpreted as laid down above. But this does not harmonize with the account of the Rook’s move, and of the 21 (or, supposing the two dice are distinguishable the one from the other, 36) combinations possible, 12 (21) give totals of 7 and upwards, and are unintelligible. (2) One die gives the piece to be moved, the other prescribes its move. But this again does not harmonize with the Rook’s move, and, besides, both King and Horse have more than six moves open to them in some positions, and the cubical die could not distinguish between more than six. (3) A combination of (1) and (2), which would involve the difficulties of both at the least. (4) Only one of the throws is to be used at the option of the player. This would reduce the number of unintelligible throws, and allow for the exercise of a certain amount of discretion. But again the Rook’s move is a difficulty, unless there is a special privilege attached to the throw of doublets. If so, 5 . 1 and 6 . 4, as meaning 5 . 5 and 4 . 4 respectively, would have to be counted as doublets. If both dice could