From my introductory chapter to the catalogue accompanying an exhibition of Oriental Games curated by the Asia Society, New York, 2004
In surveying the principal games of Asia it will be helpful to use a classificatory rather than a geographical approach, as its categories are more clear-cut. The most basic classification of games groups them into three broad divisions: physical games, strategic games, and gambling games. These divisions are said to reflect the principal channels through which the outcome is determined - whether by physical prowess, by mental application, or by chance.34
This, however, confuses the issue, as the three are not as clear-cut as they may appear. The essence of physical games is that they require the physical presence and participation of the player, unlike, say, Chess, which can be played on-line, by post, or by dictating moves to an amanuensis. Their outcome may involve a high degree of mental strategy. The forward planning and positional play of Snooker, for example, may justifiably be compared with that of Chess. It also involves chance by virtue of the Butterfly Effect referred to above. Strategic games, though by definition not primarily physical, may nevertheless be influenced by the physical constitution of the player (illness, tiredness, etc) and the physical circumstances of play (background noise, interruptions, etc), while chance is at least simulated, if not represented, by the unpredictability of an opponent's plans or reactions. Gambling games may involve an element of mental skill, such as taking cognizance of mathematical probabilities and matching one's stake accordingly. They may also be also physical to the extent dedicated players of dice games such as craps would not dream of letting anyone else roll the dice for them. It is certainly possible to play craps on-line, but throughout history the attraction of dice-play has been the pretence that a player can somehow exert control over the outcome by the mere fact of rolling them. The point is well made by Dennis Tedlock:
We might call it a "game of chance," which is what Culin calls similar games in this book,35 but that expresses the point of view of an observer. Meanwhile, the participants constantly think in terms of strategy, pitting their wishes against chance in momentary acts of magic, which is what we all find ourselves doing when we throw dice. / The second paradox of the Zuñi game of wooden dice is that, technically, it is not what Culin calls a "game of dexterity," and yet the players do try slightly different ways of handling the sticks, as if they could influence the outcome of a throw. [...] So if there is any dexterity here, it must remain on the side of magic [...]
Games of chance involving dice or other lots are therefore, in practice, largely physical games, requiring the active participation of the player for social, psychological and "magical" reasons rather than by their nature.
The word "game" is notoriously imprecise, encompassing, as it does, a number of related activities. In the case of physical games it overlaps with "sports", while so-called games of chance overlap to a large extent with "gambling" in the broadest sense of the term, including betting on the outcome of events entirely beyond the control of the player. As neither of these overlapped areas is the subject of this exhibition or essay, we will restrict our general survey to games of strategy, whether involving a significant or an absolute minimum of intrinsic elements of chance.
Strategic board games
The basis of strategic board games is that the players, normally two in number, place pieces on, and/or move them around, a flat surface marked with a geometric array of specific locations. They each move alternately; each move changes the overall scenario; and the winner is the first to achieve whatever the rules of the game define as a winning position, such as a particular location, the capture of all opposing pieces, or whatever. Moves comply with well-defined rules of play which exclude intrinsic chance factors such as dice. Each player therefore has "perfect information" as to all past moves, the current position, and all potential future positions. The perfect information one has of all previous positions is a matter of memory. That of all possible future moves, as governed by the rules of play, is something like the reverse of memory: it might be called forward visualization, or " projection". Strategy is by definition a plan of procedure based on this factor. How far forward you can see, and therefore plan, is a function of your own power of projection and of the strategic depth inherent in the game you play.
Strategic depth can be measured by the mathematical device of the strategy tree. A strategy tree (which is usually drawn upside down, and so looks more like a strategy seaweed) consists of nodes that branch out into further nodes. Each node represents a move by one player, and each branch leads to one or more nodes representing potential responses by the other. In programming an artificial intelligence it is possible to assign to each node a value derived from a combination of obvious strengths and weaknesses and a database of previously encountered positions, and to compute the best line of play from perhaps millions of alternative futures. If people were machines, such games would not be fun to play; in fact, people with over-developed powers of mental analysis are prone to a pathology36 that might be described as the strategic-game equivalent of an addiction to gambling, which is only better known because it is more common.
The best-known and most widespread games of strategy are Go/Wei-qi, Chess, Checkers/Draughts, and alignment games. All are demonstrably of Asian origin, though today the most internationally favoured form of Chess is that of European development, while Checkers, at least in its black-and-white chequered form, is also distinctively European. The antiquity of Wei-qi and its attribution to the China have already been remarked upon. Checkers derives from the Moorish game of Alquerque, of which a description appears in the Alfonso MS of 1283. Alquerque is played on the points of a linear grid, and is one of dozens, if not hundreds, of variants involving different designs and numbers of pieces and played throughout northern Africa, the Middle East, India, and south-east Asia as far as Indonesia. Variants and relatives of Alquerque are still being numerously recorded across a broad sweep of Africa and southern Asia. Murray describes about three dozen in varying degrees of detail, and Lhôte another dozen or so. Their wide variability and diffusion suggest considerable antiquity, which is underlined by the appearance of a diagram incised in a roofing slab of the ancient Egyptian Temple at Kurna. The European contribution was to transfer its play from a linear grid to the dark squares of a chessboard and (perhaps, but not undisputedly) to introduce the promotion of "men" to "kings".37
Chess
Chess. The origin of Chess is usually (but also not undisputedly) attributed to northern India some time before the first undisputed mention of Chaturanga around 600 CE. Although strategic board games were known to classical European antiquity, Chess eclipsed them by virtue of being played with a number of pieces differentiated as to movement and capture and perhaps by its distinctive objective of winning by capturing a key piece - the king - rather than a whole army of commoners as in Alquerque/Draughts, which leads to frequent draws. The attribution is, however, disputed on the grounds that the many varieties of Chess played throughout Asia differ in too many fundamentals to have shared a common ancestor as late as 7th-century India, particularly aberrant examples being the Chinese game of Xiang-qi and the Japanese Shogi. Others include Changgi (Korea), Sittuyin (Burma), Makruk (Thailand), and Mongolian and Cambodian Chess. Travelling westwards, Chaturanga became Chatrang in Persia (Iran) and Shatranj in the Arabian world of Islam. Thanks to the superior playing and analytical skills of Arabian Chess masters, and especially to their voluminous writings, Shatranj could be regarded as the first truly international form of Chess - indeed the only one - from 800 to the development and expansion of the modern European game from 1500 on.
Asymmetric strategy games
A more diffuse class of strategy games, as represented by the European Fox & Geese, is of vaguer provenance. Murray and others refer to them as hunt games, but as the hunting theme is by no means universal they are better described as asymmetric capturing games, in that one player has one or a small number of powerful pieces while the other controls a larger number of weaker pieces. Typically, though not invariably, the latter seeks to win by hemming in or capturing the singleton, while the former endeavours to prevent this by escaping or capturing too many of the larger force make this possible. Asian games of this type typically pit one or more tigers or leopards against herds of cattle or flocks of sheep, while the Chinese equivalent features rebels against a general, and those of northern Europe (such as Hnefatafl) some sort of rabble against a king and his elite bodyguard. Ancient games of this type are commoner in northern than southern Europe, suggesting some route of cultural transmission well north of the Silk Roads route, though of as yet unknown direction. It is conceivable that asymmetric war games gave rise to those of the Chess type by equipping both sides with one powerful piece and one army in order to secure symmetry and hence equality.
Mancala
Mancala games, notably absent from Europe, are widespread in a broad band stretching from western Africa through the Middle East and onwards to China. The whole family, with a vast number of regional and local varieties, is often referred to as constituting the national game of Africa. Asian representatives are still being explored, but there is nothing as yet to suggest an Asian priority. Alignment games, such as the European Nine Men%s Morris (Mill, Merels, etc) and its pencil-and-paper equivalent Tic-tac-toe, are, similarly, of too ancient and universal a distribution to suggest a specifically Asian provenance, except in so far as a majority of the earliest civilizations are found in Asia. Go-ban and Renju are distinctively Japanese, but relatively modern forms. Simple blockade games, in which each player seeks to manoeuvre his pieces in such a way as to prevent his opponent from moving, are widespread throughout Asia in a form typified by the Indian game of Do-guti.
Race games
Turning from strategy games to games of mixed chance and skill, we find these divided into two broad classes: race games, which are board games with pieces moved in accordance with the throw of dice or other randomizers, and card-domino-tile games, which, though necessarily played on a flat surface, require no board with positional markings. Actually, both classes include a variety of games ranging from a high degree of skill to virtually none at all, thereby forming an overlap with gambling games.
The most skilled race games are those of the Backgammon family, best referred to by the generic name of Tables. What makes them games of high skill is that the player controls a large number of pieces, typically fifteen as in Backgammon, and therefore is usually presented with a wide range of options at each turn (depending on the roll of the dice), making it possible to plan strategically ahead, at least on a contingency basis. Tables is equally well known in classical European antiquity, ancient China, and all stations between, making it difficult to determine priority. If ancestors are sought in a geographically central area, attention will attach to the Royal Game of Ur. Home of the biblical Abraham and capital of the Sumerian Empire in the third millennium BC, this city state's remains lie some 105 miles WNW of Basra in modern Iraq. Ur is also the home of the oldest complete set of board-gaming equipment ever found. The Royal Game does not look like Tables, but shares with it a bilaterally symmetrical layout with the two players facing each other across the shorter dimension of the board and moving pieces in such a way as to interact with one another by sharing pathways.
Virtually identical with Backgammon is the middle eastern race game Nard, first mentioned in the Babylonian Talmud (300-500 CE). The Middle Persian romance Chatrang-Namak (c.650-850 CE), which retails the story of how Chess entered Persia from India and led to the invention of Nard in return. The Chinese game of Shuan-liu, "Double Sixes", certainly resembles Nard, and may date from the fifth century. The Arabs adopted Nard following their 7th-century conquest of Persia, and it became popular throughout the Moslem world, spreading northwards in Georgia, to the Kalmuks of Central Asia, the steppes north of Astrakhan and into the Deccan, but failing to displace games of the Pachisi family in the affections of the Hindus. Shuan-Liu also spread from China throughout the orient, including Korea (Ssang-ryouk), Japan (Sunoroku or Sugoroku), and elsewhere. Nard also travelled westwards, entering Europe through the Moorish conquest of Spain and in the baggage of crusaders returning from the east.
Whether the ancient Chinese game of Liubo bears any relation to early forerunners of Tables is a tantalizing question. This prestigious game of the Han period (206 BCE-220 CE) was said by a contemporary historian to have been played as early as the 12th century BCE, in the reign of King Wuyi. Information gathered from the survival of complete sets, from contemporary illustrations, passing references, and, especially, from a ceramic tomb sculpture in the British Museum of two men playing the game, tell us practically everything there is to know about it apart from that most vital component, the rules of play. There appear to have been two forms of the game, one played with and one without dice.38 Another relatively strategic race game is Pachisi, justly regarded as the Indian national board-game, though also widespread in western Asia. Pachisi is well-known in appearance, having inspired the 19th-century children's game of Ludo (British) and Parcheesi (American); but these are relatively tame compared with the real thing. Although each player moves only four pieces, compared with the fifteen of Nard, the fact that Pachisi is played by four, in two opposing partnerships of two each, adds a notable element of strategic and interactive skill.
Other noteworthy race games include the Chinese game of Promotions, the Korean national game of Nyout, and an extinct Arabian astrological game described in The Alfonso manuscript of 1283. At the complete non-skilled - or, rather, fate-determined - end of the spectrum comes the Asian original of Snakes & Ladders (Chutes & Ladders), known from India, Nepal, and Tibet, and occurring in Vaishnaite (Hindu), Jain, and Muslim versions. Each player has only one piece to move, and no choice of play. It is a game of moral instruction, in which the player pursues a sort of Pilgrim's Progress from the lower levels, associated with earthly desires and vices, to the higher, associated with increasingly more ethical and spiritual values, eventually reaching Vishnu, or Nirvana, or an equivalent peak experience. Given its great variety of sizes, designs, and relative scales of values, this family of games is thought to have originated as early as the thirteenth century CE. From the existence of examples lacking the snake-and-ladder short cuts it is possible that they were originally absent, and introduced at a later date to make the game more exciting. The numbers and relative proportions of snakes and ladders vary considerably, but in design remain fairly constant, with few local exceptions and departures.
Cards, tiles and dominoes
Cards, tiles and dominoes are basically the same type of gaming equipment. Whether made of paper, card, bone, ivory, or plastic, they consist of flat objects of which one side is marked with numbers or symbols significant to the play while the other is blank. They can therefore be held in such a way as to enable players to conceal the identity of their pieces until they are played. Hence, as the Italian scholar Cardano observed in 1564, "There is a difference from play with dice because the latter are open, whereas play with cards takes place from ambush for they are hidden".39 Chinese priority in the invention of paper and printing reinforces the sparse documentary evidence ascribing their origin of card games to China perhaps as early as the ninth century CE. Chinese cards, however, differ considerably from those known to the west. The principal types are domino cards, each end of which carries a pattern of spots representing the roll of two dice; money cards; and chess cards. The latter are marked with chess pieces and are apparently of 19th-century origin. Money cards are of two types. One contains suits designated coins, strings of coins, and myriads of strings, each being ten times the value of the preceding suit; the other represents a further development with an additional suit designated tens of myriads. Ma-jiang, known to the west as Ma Jong, is another 19th-century invention played with what amounts to an extended set of four three-suited packs shuffled together. Only in the west is the game necessarily played with tiles: in Chinese communities they are more often made of paper like ordinary cards. Similar remarks apply to dominoes, which only reached the west in the 18th century. Most western domino games are positional, in that play consists in matching ends together, for which purposes solid tiles are essential. Chinese domino cards, however, are also made of card or paper, being mainly used for "fishing" games (resembling the western game of Casino or Cassino) and trick-taking games, like ordinary money cards or western playing-cards. Japanese and Korean cards - in the generic sense of the word - are equally varied, but evidently related to or inspired by Chinese originals.
Substantially different, however, are the playing-cards of India and western Asia. Whereas Chinese cards display a hierarchy of suits containing unequal numbers of cards, those of India and the Islamic world consist of suits all equal in value and containing an identical series of numerals (usually 1 to 10) and either two or three court cards headed by a king. Indian cards, often but not invariably circular, are of three main types: an eight-suited deck used mainly by Muslims; the Hindu Dasavatara deck with ten suits, each representing an incarnation of Vishnu; and a 12-suited, 96-card deck bearing thematic reference to the Râmâyana epic. They are used for two classes of games: Ganjifa, based on trick-taking, and Naksha, a numerical gambling game more akin to Twenty-one/Blackjack. There is no written evidence for the existence of Indian playing-cards earlier than the 16th century.
What we now regard as the standard international deck of cards is of European development dating from the 14th century. Ironically, they have supplanted and expunged from the Middle East the ancient playing-cards of western Islam from which they are thought to derive. Evidence adduced from specimens in the Topkapi Sarayi Museum in Istanbul show the Islamic models to have consisted of four suits, namely swords, polo sticks, goblets, and coins, each containing numerals one to ten and headed by a king and two viceroys. Exactly how these relate to Indian (and imperfectly attested Persian) playing-cards are a problem in search of a solution, as the evolutionary relationships between eastern and western cards remain maddeningly obscure.
How did games begin?
The question is obvious, but, unfortunately, amenable only to speculation. The oldest civilizations to have left written records refer to games of various sorts without suggesting that any of them were novelties or out of the ordinary run of daily activities. Where written records are lacking, archaeological remains include game boards scratched on stone or clay and the occasional shaped piece that might have been used in gaming, not to mention obvious randomizing devices such as dice and astragals. Anthropologists who have studied still extant primitive societies find that most have games of one sort or another.40 It would seem that games, like languages, religion, and the arts, are natural products of the human urge to order and harmony, and originated far beyond the catchment area of even prehistorical research. "We are order-making creatures, like it or not", says Alex Randolph, one of the world's great games inventors.41 It is generally accepted that physical games developed earlier than gambling games, and that games of mental strategy later, if at all. Strategy games are rare among primitive societies, and become most prominent in relatively advanced civilizations, especially where administrative advancement tends to be secured by competition rather than by the casting of lots.
The most obvious supposition is that games arose by the adaptation to playful purposes of implements and activities originally devised for functional purposes. For example, the primary function of shooting arrows from bows in a hunting society is to secure food and perhaps to guard hunting grounds from rival tribes. A secondary purpose would be to practise the use of bow and arrow in order to improve one's skills. A subsequent one would be establish, by competition, who the best marksmen are in order to allocate more efficiently the various co-operative labours of hunting. Some archers would gain more pleasure from the exercise of skill itself, or from the kudos attaching to the best of them, than from fulfilling their original purpose. What began as the purely practical thus gradually evolves into the purely recreational. One might say that archery, by definition, has become more playful by becoming more abstract, thereby underscoring our previous comment on a natural tendency of games to be more characterized by abstraction than representation.
Another origin of games, or of some games at least, may have lain in that other natural tendency of human beings to representationalise their aspirations and activities. Cave paintings of hunting scenes have been the subject of various interpretations as to their probable purpose: to encourage a successful outcome by a process of sympathetic magic, for example, or perhaps to plan strategies or serve as a visual aid in identifying good target areas. It is possible that another type of visual aid to planning strategy would take the form of placing pieces representing prey and hunters (or enemy and warriors) on a flat surface charged with other pieces or markings representing topographical features, and moving them around to represent appropriate actions. A similar process of abstraction would then have resulted in the evolution of a game through a shift of emphasis from the practical to the recreational employment of the equipment involved. String games (non-competitive games of the Cat's Cradle type), which are found all over the world even in primitive societies that appear to lack any other type of formal game, may similarly illustrate the same abstracting tendency, as it is extremely unlike that string was originally invented for games and only later found to be really useful for tying things up.
The same process probably also underlies the derivation of games of chance from the casting of lots intended for what would originally have been considered practical and functional purposes such as divination, ritual, and decision-making. Doubt has been cast on this theory by reference to the awe in which the laity would have held those authorized for such practices, such as magicians, shamans, and the like (who also, no doubt, guarded their position and privileges jealously), or to a fear of upsetting the gods; but it is not difficult to see those driven by the natural gambling impulse as in some sense "playing the part of" such favoured authorities as they submit themselves to the outcome.
A survey of the games of Asia will not of itself provide an answer to the origins of games. But as we explore such ancient and distinctive contributions as playing-cards and dominoes, Go, Nyout, Pachisi, Nard, and the mysterious game of Liubo, we may come to realize that the whole of humanity is related not just by its urge to play, which is common to many creatures, but to play in accordance with those structured sets of rules that we call games.
In order to form a just estimation of the character of any particular people, it is absolutely necessary to investigate the sports and pastimes most generally prevalent among them. War, policy, and other contingent circumstances, may effectually place men, at different times, in different points of view; but, when we follow them into their retirements, where no disguise is necessary, we are most likely to see them in their true state, and may best judge of their natural dispositions.42]
