Draughtsmen and Dice
The collection of Islamic ivory table pieces in the Ashmolean Museum consists of two draughtsmen (nos 13, 14; figs 34, 35), two rectangular dice (nos 15,16; figs 36, 37), with the numbers on the opposite faces 1-6, 2-5, and one cubic die (no.17; fig38), with the modern number pairings. The cubic die is obviously modern, of the type commonly found also in Europe, and discussion will therefore be concentrated on the other four pieces. The problems are not ones of identification, but rather of dating and of establishing for which game or games they were used.

Draughtsmen (34)
If scholarship on Islamic chess pieces is still at an early stage, for Islamic draughtsmen and dice it is almost non-existent. Draughtsmen have retained practically the same shape, with almost identical decoration, in every period and every culture. Roman ivory or bone draughtsmen, (35) for example, are not dissimilar to Coptic (36) and Islamic ones (see Appendix II). The decorations employed are concentric circles containing rings of concentric circles, or small circles with a dot in the middle. Such circles are employed decoratively both for draughtsmen and dice from the Mohenjo Daro era of the 3rd millennium BC (37) down to Roman times (38) and beyond to Coptic (see note 36), medieval (39) and modem Islamic and European periods. This simple form of draughtsman has been used for various board games such as alea, tablas and nard. (40)

(p.124)

Dice (41)
Equally, dice are used for a wide variety of games of chance (figs 39, 40)(42) and, in the Islamic context, it is very difficult to differentiate those that use cubic dice from those that use rectangular dice. Apparently the rectangular die is a direct evolution of the talus, the Roman knuckle-bones of sheep and goats. According to Culin 'Among the Greeks and Romans numerical values were attributed to the four long sides, the two pointed ends not being counted. The two broad sides, respectively convex and concave, counted three and four, while of the narrow sides, the flat counted one and the indent six.' (43) He then deals with the Turks, Arabs and Persians, specifying that 'The Arabic name for bones is kab (dual, kabatain, plural, kabat) (44) meaning 'ankle', referring to their source. Two bones are now commonly used - one from the right and the other from the left leg of a sheep. I regard them as the direct ancestors of cubical dotted dice, the name of which in Arabic is the same as that of the bones. The dice used in Arabic countries are made in pairs, and the most popular and universal game is with two dice, kabatain. (45) There is evidence that the rectangular die could have been used for the four-handed chess game, (46) and possibly for the oblong chess game (47) and nard (a race game).(48)

(p. 125)

The game is played by four people, having eight men each: King, Bishop Knight, Rook, and four Pawns. It is clear that both cubic (1-6, 2-5 3-4) and rec- tangular (1-6, 2-5) dice can be used. Each number relates to a specific piece or pieces, which must be moved when that number comes up. If the dice are rectangular, therefore carrying only four numbers, each number must correspond to two kinds of piece. But if they are cubic, with six numbers, the same rule still applies because the two extra faces are considered equivalent to two of the set of four faces, thus implying that the rectangular dice are the original ones and the cubic dice only a substitute for them.

(p.126)

Ordinary Chess
In the Alfonso manuscript we are informed that the ordinary game of chess may be played with dice to 'speed it up',(55) and continues with the explanation of how to move the pieces according to the number thrown. Since it speaks of six numbers, it is clear that a cubic die is to be used. There is also literary evidence for chess played with dice in Italy in the eleventh century, although neither the form of chess is specified, nor the kind of dice.(56) It may also be noted that the Alfonso manuscript refers to the use of a seven-sided die for the great chess (twelve by twelve cells).(57)

Oblong Chess
Among the games deriving from ordinary chess is oblong chess, an account of which is found in al-Masudi's Muruj al-dhahab ('Lands of Gold', tenth century). (58) This was played with the help of dice on a board of four by sixteen squares, with pieces of the ordinary game. In a twelfth-century Arabic Kitab al-shatranj ('Book on Chess') in Istanbul (59) it is explained that the same kind of dice as those for nard should be used, but without specifying whether cubic or rectangular. This manuscript and others are stated to derive from a work by al-Adli (c.850) from which it would follow that the game was already the subject of a specialized literature in the middle of the ninth century. The Kitab al-shatranj also includes the earliest recorded Islamic mention of the use of dice to determine the moves of a form of chess.(61)

Nard (or Nardshir)
From the Chatrang-namak and, later, the Shahnama we leam that nard was invented in Persia.(62) Here we also find an account and a symbolic explanation of the game, which was to be transmitted to Arab writers. .Mm/was immediately successful among the Arabs, so that al-Adli wrote a treatise on it, which unfortunately is lost (see note 60).

Al-Yaqubi (tenth century), in his Tarikh,(63) gives the earliest Islamic account of nard.(64) Here we find, in addition to a description, the symbolic explanation of the game already present in the Chatrang-namak. This is important for our purpose because it includes the symbolism of the die. The board stands for the year. It has 24 points because there are 24 hours to the day. It is arranged in two parts, each with 12 points symbolizing the 12 months of the year, or the 12 signs of the Zodiac. The number of men (in Arabic called kilab, 'dogs') is 30, because there are 30 days to the month. The two dice stand for day and night. The faces are arranged 6-1, 5-2, and 4-3, so that the total of the dots on each pair of opposite faces is 7, the number of the days of the week and of the planets.(65) The same explanation may, we deduce, have been offered for the choice of numbers 1-6, 2-5 on the rectangular die.

(p.127)

(p.128)

What is important is not the sequence of the numbers, but the sum of the opposite faces.(66) This symbolism is further developed in the Arabic treatise on chess Kitab fi al-shatranj wa mansubatih wa malhih ('Book on the Game of Chess, its Positions and Subtleties'), which includes a section on nard, in the British Library.(67)

Regarding the origin of this symbolic explanation, Hyde (68) quotes notes on the Byzantine game oftabia from Greek authors, containing the germ of the astronomical explanation in the Chatrang-namak. Noldeke (69) suggested that the symbolism went back to a Neo-Platonic or Neo-Pythagorean source. The game of tabla, or taula, was probably identical with the Persian and Arabic nard. It is generally accepted that when nard reached the Byzantine empire it was given the name of tabulae, from the draughtsmen with which it was played, thus acknowledging that nard was played with draughtsmen.

Of all these games, nard-was the one most widely played and most successful in all Islamic periods and among all levels of the population. Similar to European backgammon, (70) it still enjoys great popularity in the Arab countries, where it is also called tawula. (71) As we have seen, there is evidence that draughtsmen were used, but it is not clear whether the rectangular dice were also employed. It is therefore likely that the draughtsmen in the Ashmolean Museum are pieces for nard, but we cannot be sure for which game the rectangular dice were used.

Dating the Ashmolean draughtsmen and dice
None of these pieces has been examined at the radiocarbon accelerator. Several were given by the Revd GJ. Chester in the second half of the nineteenth century, some acquired from Cairo. Only one piece is from an archaeological context: a draughtsman (no. 13) excavated at Fustat in a level datable to the ninth century.(72) The rectangular die (no. 15) is also from Fustat, (73) but a surface find. The other draughtsman (no. 14) and die (no.16) are similar in appearance to the datable draughtsman, but any attempt at dating by comparison is unreliable, because of the unchanging nature over centuries of the shape and decoration of draughtsmen and dice.

There is only one further general observation that may be relevant to our purpose: the rectangular die seems not to be used in the Arab countries in modern times. The archaeological evidence in Islamic contexts, so far, is only for a period between the ninth and the eleventh century (see Appendix II). But we need more archaeological evidence and a good analysis of the existing pieces before any reliable conclusions can be drawn. In the present state of research, therefore, a relative chronology for the dice is not possible, but it is to be hoped that these pieces will be subjected to radiocarbon examination, thus enabling a more accurate dating.

(p.130)