ArticlePDF Available
Board Games Studies
CNWS PUBLICATIONS is produced by the Research School CNWS, Leiden
University, The Netherlands.
Editorial board: F.A.H.D. Effert; K. Jongeling; F.E. Tjon Sie Fat; W.J. Vogelsang
(editor in chief); W. van Zanten.
All correspondence should be addressed to: Dr. W.J. Vogelsang, editor in chief
CNWS Publications, c/o Research School CNWS, Leiden University, PO Box 9515,
2300 RA Leiden, The Netherlands.
Tel. +31 (0)71 5272987/5272171
Fax. +31 (0)71 5272939
Board Games Studies
Board Games Studies, Vol. 1. International Journal for the Study of Board Games -
Leiden 1998: Research School CNWS. - (CNWS publications, ISSN 0925-3084)
ISBN 90-5789-005-4
Subject heading: Board games.
Board Games Studies:
Typeset by Cymbalum, Paris (France) and Nelleke Oosten, Oegstgeest (Holland)
Cover design: Nelleke Oosten
© Copyright 1998, Research School CNWS, Leiden University, The Netherlands
Copyright reserved. Subject to the exceptions provided for by law, no part of this
publication may be reproduced and/or published in print, by photocopying, on
microfilm or in any other way without the written consent of the copyright-holder(s);
the same applies to whole or partial adaptations. The publisher retains the sole right
to collect from third parties fees in respect of copying and/or take legal or other action
for this purpose.
International Journal for the
Study of Board Games c n w s
The following affiliated institutes underwrite
the efforts of this journal and actively
exhibit board games material, publish or
financially support board games research.
Institut für Spielforschung und Spielpäda-
gogik, Salzburg
Address: Prof.-Dr. G.A. Bauer,
Hochschule Mozarteum, Schwarzstr., 24,
A - 5020 Salzburg (Austria)
International Institute for Asian Studies,
Address: Prof.dr. W.A.L. Stokhof, IIAS,
P. O. Box 9515, NL - 2300 RA Leiden
(The Netherlands)
Russian Chess Museum and Magazine
“Chess in Russia”, Moscow
Address: Natalya Ivanova/Yuri Averbakh,
Gogolevsky Blvd. 14, 121019 Moskwa
British Museum, London.
Address: Dr I.L.Finkel
London WC 1B 3DG, United Kingdom
Universiteit Maastricht, Department of
Computer Science, Maastricht
Address: Prof.dr. H.J. van den Herik,
P. O. Box 616, NL - 6200 MD Maastricht
(The Netherlands)
Corporate Sponsor
Spiel des Jahres e.V.
Patrons support the efforts of this journal
through continuous financial support. If you
wish to become a patron, please contact
CNWS by post, fax or E-mail. We hereby
thank all our sponsors for their generous
Sjaak Griffioen (Griffioen Design Puzzles and
Games), Niek Neuwahl, Jurgen Stigter,
Glenda Trew (Oware Society), Anneke Treep,
L.V. and P.V.
Editorial Board
Thierry Depaulis (FRA)
Vernon Eagle (USA)
Ulrich Schädler (GER)
Alex de Voogt (NL, Managing Editor)
Board Games Studies is an academic
journal for historical and systematic
research on board games. Its object is
to provide a forum for board games
research from all academic disciplines
in order to further our understanding
of the development and distribution
of board games within an
interdisciplinary academic context.
Articles are accepted in English,
French, and German and will be
refereed by at least two editors under
the final responsibility of CNWS,
Leiden University.
Editorial / Foreword
Ulrich Schädler,
Mancala in Roman Asia Minor?
Thierry Depaulis,
Inca Dice and Board Games
Vernon A. Eagle,
On a Phylogenetic Classification of Mancala Games, with
some Newly Recorded Games from the "Southern Silk
Road," Yunnan Province, China
Caroline G. Goodfellow,
The Development of the English Board Game, 1770 - 1850
Lieve Verbeeck,
Bul: A Patolli Game In Maya Lowland
Irving L. Finkel,
Edward Falkener: Old Board Games for New
E.R. Santos Silva, Jogos de quadrícula do tipo Mancala,
by Philip Townshend
A. van der Stoep, Over de herkomst van het woord damspel,
by Rob Jansen
M. Zollinger, Bibliographie der Spielbücher,
par Thierry Depaulis
Instructions to Authors
Research Notes
Notes de recherche
Book Reviews
Comptes rendus
Board games have played an important role as research objects in the sciences of
this century. At first, games and board games were studied from a historical
perspective. In 1944, Von Neumann and Morgenstern provided a basis for using
games and board games in the computer sciences and in economics, such as in the field
of game theory. Research on board games accelerated with research on chess, in
particular chess masters, which has proved fundamental in the cognitive sciences since
de Groot (1949), followed by Newell & Simon and others. Chess is still dominant in
most fields but slowly other championship games enter these fields as examples or tools
in research.
Only recently has research on board games other than chess been possible. Since
Thomas Hyde (1694) there are historical descriptive works on board games. However,
even in 1952 when Murray published A History of Board Games Other than Chess,
research did not suffice to warrant an important shift in attention in the sciences. These
other games had rules, boards, pieces, players and contexts unknown to the academic
world. Sometimes parts were known but never studied, as shown by the first Ph.D.-
thesis on the subject of draughts (or checkers) only in 1997.
Since 1952, some disciplines of research have started to consider games and board
games other than chess. Studies of sculptured game boards in art history (Walker 1990)
and a contextual analysis of board games in anthropology (Townshend 1985) are just
examples from the field of mancala games. This interest from art history, anthropology
and also archaeology (Schädler 1995), psychology (Retschitzki 1990) and linguistics
(van der Stoep 1997) has grown rapidly since the 1980s. International colloquia,
scholarly books, research centres and a growing number of articles and inventories are
being produced for which this annual publication will provide a continuous platform.
Board games are a complex form of games. They consist of boards and various kinds of
pieces (dice, pawns, counters, etc.), a system of rules, and most importantly players.
The context of playing board games includes referees, interfering and non-interfering
spectators, rules of ceremonies or rules of etiquette, club houses and societies, boards
for special occasions, etc. Playing a board game introduces movement, sound,
atmosphere and other elements which are described by poets rather than academics. If
we consider a context with players, boards and pieces, and rules, it appears that these
elements cannot be separated for a complete understanding of a board game. The rules
may influence the board and vice versa. The players may determine the shape and kind
of boards and the specificity of the rules. They form a complex ‘being’ which is a board
Board games in their complexity present the researcher with various questions. For
instance, the (inter)relationship of the aspects of a board game are little understood.
Also, the historical development and distribution of board games has been a point of
discussion which was started in historical works by Murray (1952), Bell (1960), but also
by Falkener (1892) and Hyde (1694) to name a few.
Studies of board games collections (Goodfellow 1997 in BGS) are rare and hardly
ever coincide with fieldwork on context and rules. The results of fieldwork, collection
studies, analyses of rules and the study of players still need to be studied within their
interaction, their dependency and their consequences for the development and
distribution of board games. The methodology for classification appears fundamental for
answering these questions in a systematic way (Eagle 1997 in BGS).
Each article in Board Games Studies makes a rich source of literature available to
scholars. This literature makes it possible to study board games with the necessary
background knowledge. Area studies appear both in need of this literature and are at the
same time instrumental in adding to such literature. This is shown by Depaulis (1997
in BGS) and Verbeeck (1997 in BGS) who contribute considerably to the field of Latin
American studies. However, even interdisciplinary area studies are limited in their
approach. Most board games appear to be distributed across the continents and rare
board games in Asia may only be understood with a thorough understanding of related
games in Africa or their relatives in antiquity (Eagle 1997 & Schädler 1997 in BGS). As
such, board games studies are interrelated studies separate from but dependent on the
known disciplines.
A discipline of research prefers to concentrate on one of the elements of a board
game. Archaeologists and art historians tend to study objects, while computer scientists
are more interested in rules and their consequences. This results in two general problems
for which this journal intends to provide a solution. Firstly, as was stated above,
individual disciplines do not give insight in the complexity of board games. Instead, only
aspects are discussed without the complexity of their interaction. Secondly, research on
board games is presented in many unconnected publications. It is necessary to create a
systematic inventory of board games research in order to get insight in the complexity of
board games as a whole. Colloquia of the past seven years have already made an attempt
in presenting the findings of various disciplines in one publication. This journal is a
direct result of the success of and need for these publications.
In line with the particularities mentioned I sense an ambition for board games research.
It is my belief that, in the study of board games, the individual disciplines need to be
complemented by a perspective which is primarily concerned with the board games
themselves. Since academic disciplines cannot provide us with such a viewpoint, it
should be the role of this journal to develop and show the importance of such a
perspective providing academia with an insight unknown to the practitioners of its
established disciplines.
Alexander J. de Voogt
Articles / Articles / Beiträge
Mancala in Roman Asia Minor? / Ulrich Schädler
Merels games, chess, backgammon and mancala are certainly the most
widespread classical board games in the world. Our knowledge concerning
their origins, both chronologically and geographically, is however remarkably
poor. While entire libraries could be filled with theories about the history of chess, the
opposite is true of backgammon, its evolution having hardly ever excited any interest(1).
The origins of merels are surrounded by the darkness of prehistory and attempts to
lighten it up have rarely been made. The situation concerning mancala is not very much
better, as shown by Philip Townshend’s synopsis of the state of affairs: “The age of
mankala is uncertain. It might be as much as 3,000 years or as little as 1,000”(2). “Some
writers have ascribed to it an Egyptian, Persian, Indian or African origin”(3). It must be
stressed, however, that anthropology and ethnology have rarely tried to advance theories
about the origin and evolution of mancala. Based on observation and literary
descriptions not earlier than the 17th century of the rules adopted in different areas of
Africa (except northern Africa), the Near-East, Asia and the New World, distributional
analyses of variants and different typologies have been applied to gather information
about migrations of peoples or cultural inter-relationships(4). The history of the game in
the long term was no primary concern. The Greaco-Roman world on the other hand was
left to classical archaeology as the traditional field of research. Whether due to the afro-
ethnological domination of mancala-related research or to the lack of a sufficient
archaeological data-base concerning Greek and Roman board games, mancala has not
been regarded as a game played in the Mediterranean during classical antiquity.
This article is a preliminary attempt to contribute to the history of mancala from a
classical archaeologist’s perspective. A number of methodological problems arise from
such a viewpoint. To the archaeologist any board game appears as a tripartite set of data
consisting of a gameboard, the material needed for playing and a set of rules. In contrast
to the nearly complete knowledge of a game collected by anthropology and ethnology
by observation of people playing or explaining the rules adopted, archaeology is more or
less limited to the material remains of games. As far as the ancient Greek and Roman
culture is concerned finds of complete sets of board games are extremely rare(5). Generally
spoken there are strayfinds of gaming stones, dice and other objects on the one hand and
gameboards on the other. In some cases literary sources provide further information as
for example names of games and their rules, but more often they are themselves
problematic, since many of them consist of concise lexicographic entries, poetical or
philosophical allusions to games rather than explanations or are written by late authors,
who gained knowledge not from personal experience but from previous literary sources.
As a substitute for complete games and the observation of people actually playing
representations of board games in progress on wall-paintings or mosaics, in sculpture
and other works of art can be helpful, but follow their own laws concerning style, scale
and perspective, so that often details of the games depicted are not recognizable.
Left alone with the boards the search for cross-cultural analogies can be applied. The
comparative study of board games, however, implies various difficulties. To state that
different games can be played on one and the same board (see for example Alfonsos “Libro
de ajedrez” suggesting fifteen rules for games on the backgammon-board) is to repeat a
banality. Hitherto less considered, however, was the likewise evident fact that on different
gameboards similar games can be played. One should keep in mind that in the first place
a gameboard is a particular disposition of places for the counters. These places may be
shaped as points, holes, circles, squares, intersecting lines, letters or symbols of all kinds.
For the identification of the games played on a given board the particular shape of the
places is less important than their disposition. On Roman XII scripta/alea boards for
example the places exhibit a variety of forms but differ completely from the oblong
triangles of backgammon-boards, and yet the games are very similar. Dara, an African
game where the players try to aline three of their own pieces in order to acquire the right
to take one of the opponent’s and therefore similar to nine men’s morris, is played on a grid
of holes instead of concentric squares(6). In Egypt grids of holes are used for siga, a game
very different to dara. Siga has certain affinities with the Roman game of latrunculi, which
was played on a grid of squares(7). Therefore latrunculi has often been compared to chess,
although it was a completely different game. What follows is that it is important to be
aware of the difference between the structural layout of a gameboard and its formal design.
As far as mancala is concerned this distinction has hitherto not been observed. While
boards with parallel rows of holes have readily been identified as mancala boards, parallel
rows of squares have not been taken into account. From an archaeological and historical
point of view, however, the fact that mancala seems to be played exclusively on rows of
holes during the last centuries is no proof for the assumption that mancala boards must
have had this particular shape from the invention of the game on always and in all cultures.
On the contrary it seems more likely to suggest, that this type of board may be the result
of improvements in design or that cultures importing the game may have played it on
gameboards already existing. That these considerations could of course have consequences
for the study of origin, history and distribution of the game is perfectly clear.
The difference of structure and form of gameboards and the difference of a game
itself and its material remains handed down to us are the subjects treated in this article.
It is dedicated to a class of gameboards consisting primarily of two parallel rows of five
cells. These gameboards are to be found in Roman cities such as those of Asia minor,
where they are frequently found incised in the marble slabs of streets, squares and other
public buildings. Their identification as gameboards is suggested by the proximity of
many of them to other pavement markings definitely identifiable as boards for merels
games or alea(8). I have collected examples at Aphrodisias, Ephesus, Miletus, Cnidus and
ancient Izmir, a similar one at Didyma, but I have not seen boards of this class at
Pergamon, Teos, Claros, Magnesia, Priene, Olympos, Phaselis, Termessos, Perge,
Aspendos or Side. It must be stressed, however, that the list given below is certainly
incomplete, various excavated buildings being re-covered with sand or plants or not
accessible to the public, as for example the theatres at Side and Perge. As for Aphrodisias
I am obliged to Charlotte Roueché for allowing me to make use of her hitherto
unpublished catalogue of the pavement markings in the Sebasteion, the temenos of the
temple of Aphrodite, the Tetrastoon and the southern agora(9).
Gameboards with Two Rows of Five Cells in Asia Minor
Six different types of the gameboard in question can be distinguished (fig.1). These types
can be described as a combination of essentially two elements, that is a frame on one
hand and the filling of the frame on the other. The most elaborate frame (A) is composed
of an oblong rectangle divided into two rows of five cells by one central line running
parallel to the long sides and four intersecting lines parallel to the short sides. A simpler
form of frame (B) omits the short lines and therefore consists of only a rectangle divided
into two oblong halves. The simplest frame consists of a rectangle only (C). Finally the
frame can be completely absent (D). Whereas the cells of the A-group frames can have
no filling at all (1) or can be filled with inscribed circles (2) or cup-like holes (3), these
holes appear to represent the cells in groups B, C and D. The following list of boards
dicussed in this paper contains fifty two examples(10).
Type A1 (BMT R1):
1Ephesus, Hydrekdocheion of Laekanius Bassus,
2Ephesus, Basilike Stoa, stylobate of the eastern inner colonnade opposite the so-cal-
led Rhodian peristyle,
3Ephesus, Arkadiané, eastern colonnade between the 23rd and 24th columns north
of the four-columned monument, 16x33,5cm (fig.2),
4Ephesus, Arkadiané, in the middle of the street close to a circular and a square merels
game, 20x35cm,
5Aphrodisias, theatre, 1st sector [counting the segments of the auditorium from south
(1st) to north (11th)], 3rd step from below,
6Aphrodisias, theatre, 8th sector, 12th step from above, near square merels game and
incised gladiator’s bust,
7Aphrodisias, theatre, northern sector (11th), 3rd step from below,
8Aphrodisias, theatre, northern sector (11th), 6th step from above,
9Izmir, agora, inscribed cup-like depressions in four squares, leaf near the board poin-
ting to the middle square, numerical sign m at one corner, 22x36cm (fig.3),
10 Cnidus, propylon to the precinct of Apollo, 9,5x25cm, holes in the first square of
both rows.
Type A2:
11 Ephesus, Arkadiané, on a threshold in the eastern colonnade. Since in the frames of
the A-group the ten cells are sufficiently defined by the squares, the inscribed circles
should be interpreted as simplified holes.
Type A3 (BMT R6):
12-15 Aphrodisias, theatre, northern sector (11th), 3rd (fig.4), 5th, 6th and 8th steps
from below,
16-17 Aphrodisias, stadium, 5th sector of north-side (counting from east to west), 4th
step from above c. 4m apart,
18-20 Aphrodisias, Tetrastoon, stylobate of the west colonnade, between 2nd and 3rd
column (counting from north), 15x25cm, between 4th and 5th column, 24x39cm,
Fig 1: Types of 2x5-class game boards in Asia Minor
Fig 2: Game board at Ephesus, Arkadiané
and between 5th and 6th column, 18x40cm.
Type B3:
21 Aphrodisias, theatre, 9th sector, 5th step from below,
Type C3:
22 Aphrodisias, theatre, 8th sector, 7th step from below (only one row existing).
Type D3 (BMT H4):
23-25 Aphrodisias, theatre baths, western colonnade, beneath column at northern
26 Aphrodisias, theatre, 9th sector, 5th step from above, close to the stairs between
sectors 8 and 9,
27 Aphrodisias, theatre, northern sector (11th), 8th step from above,
28-29 Aphrodisias, Sebasteion, on the steps at the east end, 9x23cm and 11x24cm,
30 Aphrodisias, temple-temenos, on the steps at the east end of the south colonnade,
31-51 Aphrodisias, southern Agora, northern portico:
31-32 between 27th and 28th column (columns numbered from east to west)
7x15,5cm and 8x24cm,
33 next to 30th column, 17x37cm,
34-35 between 31st and 32nd column, 10x29cm and 15x25cm,
36 next to 33rd column, 7,1x22cm,
37 between 34th and 35th column, 14x29cm,
38 between 36th and 37th column, 10x28cm, next to a xii scripta/alea-board,
39 next to 37th column, 10x33cm and 12x26cm,
40 between 39th and 40th column, 10x19cm,
41-42 next to 40th column, unfinished 9x15cm, and traces of an earlier,
43-44 next to 42nd column, 8x21cm and 6x19cm,
45 next to 45th column, 8,5x16cm, close to a xii scripta/alea-board,
46 next to 46th column, 9x22cm,
47 next to 47th column, 7x20cm,
48 next to 48th column, 9x23cm,
49 next to 49th column, 11x24cm,
50 next to 50th column, 8x23cm,
51 next to 55th column, 9x20cm
52 Miletus, theatre, southernmost sector, 4th step from below, 10x22,5cm.
Thus a survey of the boards of the 2x5-class in the Roman cities of Asia minor shows
that out of fifty two examples ten boards have two rows of five squares, while fourty one
boards – fourty four with the one of type A2 from Ephesus – have two rows of five holes.
Mancala or Five Lines?
Two main questions arise: Were all six types of the 2x5-class boards used for one and the
same game or were different games played on these boards? And what kind(s) of
game(s)? At first sight it might be suggested that at least some of the boards were used
Fig 4: Game board at Aphrodisias, theatre
Fig 3: Game board at Izmir, Roman agora
for some kind of mancala, judging from the striking formal analogy and in particular
from the fact that four, if not five of the six types have holes as cells as is typical for
mancala boards. To support this hypothesis one could add two further arguments. From
a functional point of view cup-shaped troughs are very suitable to grasp a certain number
of pieces at a time with one hand as in mancala, but less suitable to move single pieces
from one place to another. From a practical point of view it would be quite astonishing,
that anyone should take pains over chiselling holes into marble, if he did not believe
them to be appropriate or in fact necessary for the game. But the case is not as easy as it
seems to be. It must be observed that some of the holes of the boards described are
reduced to points rather than holes and that all two-row boards with holes consist of
exactly five cells in each row, a peculiarity they share on a structural level with boards of
a Greek game conventionally named as five lines. The existing literary and archaeological
evidence enables us to create a fairly good picture of that game(11).
Referring to five lines Pollux states (IX 97) that “each of the players had five pieces
on five lines” adding that “on either side there was a middle line called the sacred line.
And moving a piece from it gave rise to the proverb ‘He moves the piece from the sacred
line’”. In another instance (VII 206) Pollux includes five lines in a list of dice games.
Eusthatius in his commentary to Homer’s Odyssey (Od. 1397,28), probably relying on
Suetonius’ lost book about Greek games, says that both players had their own five lines
and that the line between these was the sacred line. Its significance he explains by adding
that “the beaten player goes to it last”, i.e. the player who first manages to place his pieces
on the sacred line is the winner. The earliest reference is a verse by Alcaeus(12), implying
that moving a piece from the sacred line can lead to final victory – in a sense similar to
playing the trump card” nowadays.
Archaeological finds that can convincingly be connected with these references,
totally ignored by Austin(13), add much to their understanding. W. Kendrick Pritchett
catalogued the material hitherto known from mainland Greece, Delos and Cyprus(14). I
do not want to leave the fact unmentioned, that the numerals on some of these boards
induced some scholars to speak of abaci. Although it cannot entirely be ruled out that
the boards may at times have served for calculations, I follow the opinion that they were
primarily designed as gaming boards(15). The earliest example is a painted terracotta
gaming table found together with a cubic die in a grave at Vari in Attika, dating to the
middle of the 7th century BC(16). The board measures 18,3x24,8cm and has on its
surface five incised parallel lines ending in a circular cavity on both sides, thus forming
two rows of five holes along the longer edges of the board. Two distinct groups of five
parallel lines widening at both ends are incised on the surfaces of two stone gaming
tables dedicated possibly during the fourth century BC in the sanctuary of Asclepius at
Epidauros(17). On one of these tables six shallow lines have been added clearly at a later
date in order to create a gaming area with eleven lines next to one with five. Boards with
eleven lines have been found at several places, often with the third, sixth and ninth line
marked by a special sign, pointing to a special significance of these lines. Thus the boards
with eleven lines appear to be boards where two groups of five lines with their sacred
lines in the middle have been joined by adding a line between the two groups. This
5+1+5-layout corresponds to Pollux’ (IX 98) and Eusthatius’ (Il. 633,58) peculiar
expression Lamer(18) came across, that “a line in the middle was called the sacred line”
instead of “the line in the middle”. From the extant gaming boards this expression seems
to correspond to both possibilities, i.e. the sacred line as the middle line of five on the
one hand and of eleven on the other hand.
As far as the modes of playing on these boards are concerned, those points, grooves
or holes at the ends of the lines in some of the boards seem to demonstrate that the
gaming pieces were placed at the ends of the lines. This arrangement is represented on a
terracotta model of a gaming table from Athens, dating about 600 BC, the lifetime of
Alcaeus(19). On its surface measuring 37x12cm are engraved nine parallel lines occupied
by oval knobs at each end, probably representing the gaming stones. Although the
number of nine instead of eleven lines differs from the stone boards, there can be little
doubt that essentially the same game is meant, the reduced number of lines probably due
to the miniature scale and the overall inaccurate design of the model. At both ends of
the board two dice with their upper face showing 6 are preserved and perhaps traces of
a third die in the center. It has been argued that a winning move is represented with all
eighteen points occupied by one player’s counters after the highest possible throw of
three dice(20).
There are several reasons for not accepting this hypothesis. Apart from the fact that
the important role of the ”sacred line“ is not taken into consideration, having seen the
board in Copenhagen I wonder if there really are traces of a third die. Moreover the
corresponding numbers of eighteen points on nine lines and on three cubic dice are
merely coincidental, since the normal number of lines is five or eleven with ten or twenty
two points respectively. Finally the underlying hypothetical rule that the players had to
place a number of pieces on the points according to the result of the throw of dice,
simply does not correspond to the fact that the pieces were moved from one line to the
other. That this was the case is not only clearly indicated by the proverb “moving the
piece from the sacred line” to which the literary sources refer. This way of playing is also
represented on an Etruscan mirror depicting Achilleus and a companion (probably Aiax)
playing on a board with seven parallel lines kept on their knees(21). The circles at the ends
of the lines are generally held to be gaming stones(22). Again the reduced number of lines
can be explained by the general tendency of small scale reproductions to a certain
inaccuracy as to the details. Therefore a Praenestine mirror in the British Museum(23),
dating to the 3rd century BC(24), should also be added to the representations of the game.
Although the gaming table used by the couple shows twelve or thirteen lines, the overall
design of the board differs considerably from boards for XII scripta to which the mirror
has hitherto been attributed(25). Moreover the name of XII scripta, a game similar to alea
with close affinities to backgammon(26), does probably not refer to twelve lines on the
board, but to the use of two dice with twelve spots as the highest result, as already stated
by Nonius (170,22)(27) and confirmed by a board with two dice on it depicted on a
mosaic from Ostia (CIL XIV 607)(28).
We may therefore conclude that five lines was played on a board with normally two
rows of five or eleven points, the opposite points connected by a line and often in the
shape of small troughs. Two players played either on five or eleven lines or on separate
groups of five lines. The use of dice in five lines is attested by both the literary and
archaeological sources. Judging from the find from Vari, one die was used when playing
on five lines, whereas two dice belonged to the larger boards. Not only are two dice
placed on the board from Athens in Copenhagen, but also on the Etruscan mirror two
rectangular objects are depicted between the lines that can be taken as dice. The number
of gaming stones obviously corresponded to the number of lines, each player having as
many counters as lines on the board. Five stones for five lines are mentioned in the
literary sources, while on the clay gaming table in Copenhagen and on Achilleus’ board
all points are occupied by one counter. The pieces moved from line to line according to
the spots on the dice. If the interpretation of the sources is correct, that the aim of the
game was to place all or as many pieces as possible on the “sacred line(s)”, then probably
the pieces had to move around the board several times, for just one turn was surely not
sufficient. Presumably a counter having reached the last line on one side of the board
was shifted along the line to its other end, where it moved in the opposite direction along
the other side back to the first line, where the same manoeuvre was repeated and so
forth. It is this presumed circular movement around a board with two rows of points in
the shape of holes that reminds one of mancala games, hence the somewhat detailed
analysis of the evidence concerning five lines.
There are also other aspects pointing to a possible use for five lines of the two-row
boards in Asia minor. One is the fact that by far the largest number of two-row boards
have exactly 2x5 cells. As exceptions I noticed one board with 2x4 squares on the upper
step of the western curve in the stadium and one of 2x6 squares next to the 50th column
of the northern portico of the agora at Aphrodisias (15x32cm) as well as another next to
the north-eastern anta of the temple of Apollo at Didyma (8,5x30cm). On the other
hand several boards with precisely 2x11 squares exist, one on the stylobate of the temple
of Apollo at Didyma, three boards at Ephesus in the so-called Curetes Street opposite
the monument for Androclus(29), four boards dating to the Byzantine period have been
found at Salamis (Cyprus)(30) and one board has been reported from Petra (Jordan)(31).
This striking correspondence between the number of cells of these boards and the
number of points in five lines boards can hardly be explained as mere coincidence. The
second reason is that neither Pritchett’s catalogue of lined boards contains a single
example from Asia minor, nor did I ever find one in the ruins of the cities mentioned
above. There may be geographical and chronological explanations for this last
observation, as the boards with lines from mainland Greece, Delos and Cyprus as well
as their reproductions on the bronze mirrors all predate the Roman imperial era, whereas
none of the boards with cells in Asia minor and Cyprus is earlier than the 2nd century
AD. Since a board of type A1 is also known from the Roman theatre at Leptis Magna in
Libya(32), the boards of the 2x5-class seem to represent a later type of five lines board,
although we must keep in mind that the data-base at present is relatively sparse.
Perhaps the most important testimony in favour of five lines being played on the
two-row boards discussed in the present article is a board (fig.3) engraved in the
pavement of the State Agora at Izmir, erected in the 2nd century AD(33). As already
mentioned the board preserves the normal design of the A1 type with two rows of five
squares. Next to one corner of the board a sign can be seen, possibly a cursive µ
standing for 1000, while a circle with four spokes near the diagonally opposite corner
seems not to have anything to do with the board. Most interesting is a leaf engraved
close to the board with its base touching the middle square, thus indicating its
importance. Similar leaves appear frequently in connection with late antique represen-
tations of games, races and athletic contests, but also on gravestones, obviously as a
symbol for good luck or victory. As an illuminating example the gladiators holding a
leaf in their raised right hand engraved in the pavement of a public street in Rome
together with different types of gaming boards may be cited(34). It seems therefore not
too far-fetched to suppose that the leaf indicates the “sacred squares” in the middle of
both the two rows of squares, where the counters had to be placed to win the game.
One of the Salaminian boards cited above points to the same direction: It consists of
two parallel groups of eleven cells, the middle cell being larger and marked by an x35.
Mancala in Roman Times?
However, at some time somebody started to grind depressions into the board on the
Roman Agora at Izmir. Compared to the three holes in the squares of the opposite row,
the hole in the square next to the µ is relatively shallow. It seems therefore as if the
original intention was to supply all the cells with holes, but work has been interrupted.
Since five lines, like all other Greek or Roman board games hitherto known, was played
moving single counters from one place to another, there was surely no need whatsoever
for holes like this on a five lines board. Although many five lines boards of the group with
lines have grooves or holes at the ends of the lines, these grooves or holes are relatively
small and could have served only to hold single counters, judging from the
representations of the game described above as well as from the average dimensions of
ancient gaming pieces. The same is true for the small examples of 2x5-boards of types
A3, B3, C3 and D3. On the other hand many of these boards including the board at
Izmir have holes large enough to contain more counters, pebbles, seeds or whatever
might have been used as gaming stones. As already stated above, such depressions make
a game like mancala more comfortable to play.
Can we assume then that a board for five lines or another game was intended to be
transformed into a mancala-board? And may all the boards with large cup-like holes
(types A3, B3, C3 and D3) have served that purpose? Two arguments may support this
suggestion. Not only the formal similarity in design, which would probably induce most
Africans to play mancala on such boards, may be mentioned. Those boards at Didyma
and Aphrodisias with 2x6 and 2x4 squares mentioned above should not be forgotten.
With their equal number of cells these boards cannot have served for five lines, which
required an odd number of lines or cells.
It seems therefore, judging from the evidence presently available, that the two-row
boards presented here, have been used for both a variant of five lines played on squares
instead of lines and a mancala-like game as well, not necessarily played on rows of holes
but also on rows of squares. Concluding from the history of the cities in Asia minor,
where buildings and streets were still repaired in the 6th century, before the Arabian
invasions during the 7th and 8th centuries accelerated their decline, the gameboards
engraved in the pavements should be dated not earlier than the 2nd and hardly later than
the 8th century AD. Thus they are not later in date than the earliest examples from
Africa that have been taken as mancala boards. The boards excavated at Matara and Yeha
in north-western Ethiopia have been dated between the 6th and 8th centuries AD(36).
We have come to touch the question of the origins of mancala. On this subject
modern authors use to refer to Egypt, relying faithfully on Parker’s observations from
the beginning of the century quoted at length by Murray(37). Boards with two rows of
holes have been met with at the pyramid of Menkura (middle of the 3rd millennium
BC), at the temples of Kurna and Luxor (middle of the 2nd millennium BC) and at the
entrance to the Ptolemaic temple at Karnak(38). It has been taken for granted that the
boards date from the time of the erection of the buildings, an assumption not at all
confirmed by modern egyptology(39). On the other hand a board with 3x14 small holes,
associated by Flinders Petrie with senet, has not unreservedly been claimed for
mancala(40). Moreover, if two-row mancala was really known in Ptolemaic Egypt, we
should expect it to spread to other parts of the Hellenistic world more rapidly than in
a period of several hundred years. And if it was not, the time gap of two thousand years
between the boards at Kurna and the earliest examples a little further south in Ethiopia
remains without satisfactory explanation. What is needed is a thorough re-examination
of these boards regarding their chronology, a task that is beyond the scope of the present
article. On the other hand the neolithic board found in Jordan at ‘Ain Ghazal (6th mill.
BC) with its diverging rows of holes(41) as well as the boards from Beidha (7th mill. BC)
with grooves running through the depressions and off the slab at one end are unlikely
to be gameboards(42).
Leaving aside the doubtful old-Egyptian and neolithic evidence, both the
archaeological finds from Asia minor and Ethiopia as well as the silence of the literary
sources until the first mentioning of mancala in Arabian literature in the 10th century
AD correspond surprisingly well in pointing to a possible birth of two-row mancala
between the 6th and 8th centuries AD, or even a little earlier. Judging from the attested
age of five lines and its not having survived classical antiquity, two-row mancala seems to
have replaced the older game. The question therefore arises, whether in fact the context
of the late antique and early Byzantine cultures, in particular some special attitude of
early Christian (or even early Muslim) society towards games might have supported the
dismissal of five lines and the distribution (if not the birth) of mancala in the eastern
mediterranean? In the new light of the evidence and considerations presented here I
would finally like to compare two well attested characteristics of both games. One
peculiarity of mancala is the way of moving by spreading the pieces. In contrast to a dice-
game like five lines, in mancala the reach of a player’s turn is determined by the number
of seeds in a given hole, the spreading of pieces being the easiest way to find out where
the turn ends even without counting. The choice of the hole to be emptied enables the
player to change deliberately the values, i.e. the possible reach of a turn starting from that
point, of the following holes. Thus the function of the die as the oldest principle of
movement in board games has been integrated into the counters.
Peculiar to five lines was the function of the “sacred line” as the predetermined goal
where the pieces had to move to. Thus, in the course of the game the sacred line or
squares accumulated the counters. From Ethiopia, eastern Anatolia and elsewhere
variants of mancala have been reported where holes serve a similar function. The
important difference lying in the fact that these holes have to be captured first(43). They
are not predetermined goals that could be called ”sacred“ as in five lines. Could both this
difference as well as the particular way of moving in mancala by spreading pieces instead
of rolling dice perhaps trace back to the demands of early Christian society, where a line
or square in a game called “sacred” must have been taken as blasphemy and games of
hazard were generally ill-reputed?
However this may be, the possibility that mancala was known to the Graeco-Roman
world in late antiquity changes considerably the perspective towards the origin and
history of the game. This would provide a context for the otherwise isolated evidence for
the game being played in Greece, attested by Galt’s observation of people playing
Mancala on the island of Hydra in 1810(44) and a rock-cut mancala-board in Athens
reported by Townshend but without details as to its location, date or source(45). Moreover,
if the late antique boards in Asia minor predate the Arabian invasions, the existence of
mancala in Syria, the Levant and Anatolia might date back to Late-Roman times instead
of being due to Islamic influence.
Akurgal, Ekrem, 1993. Ancient civilizations and ruins of Turkey. 8th ed. Istanbul.
Austin, Roland G., 1940. Greek Board-Games. In: Antiquity 14: 257-271.
Becq de Fouquières, Louis, 1869. Les Jeux des Anciens. Paris.
Bell, Robert C., 1979. Board and Table Games from many civilizations. Revised ed.
New York.
Bergk, Theodor, 1884. Poetae lyrici Graeci, Vol.3. 4th ed. Leipzig.
Blinkenberg, Christian, 1898. Epidaurische Weihgeschenke. In: Athener Mitteilungen,
23: 1-14.
Breitenstein, Niels, 1941. Catalogue of Terracottas. Copenhagen.
Caputo, Giacomo, 1987. Il teatro Augusteo di Leptis Magna. Scavo e Restauro (1937-
1951). Roma.
Chavane, Marie-José, 1975. Salamine de Chypre VI. Les petits objets. Paris: 192-205.
CIL. Corpus Inscriptionum Latinarum.
Crummy, Philip, 1997a. City of Victory, the story of Colchester - Britain’s first Roman
town. Colchester.
Crummy, Philip, 1997b. Your move, Doctor! The gaming board and other discoveries
from Stanway. In: The Colchester Archaeologist, 10: 1-9.
Culin, Stewart, 1896. Mancala, the National Game of Africa. In: Annual Report of the
US National Museum (Report of the Smithsonian Institution) for the year ending
June 30, 1894. Washington: 595-607.
De Puma, Richard Daniel, 1987. Corpus Speculorum Etruscorum, USA 1, Midwestern
Collections, Ames.
Dunbabin, Thomas J., 1962. Perachora. The Sanctuaries of Hera Akraia and Limenia,
vol.II. Oxford.
Eagle, Vernon, 1995. On Some Newly Described Mancala Games from Yunnan
Province, China, and the Definition of a Genus in the Family of Mancala Games. In:
Alexander J. de Voogt (ed.), New Approaches to Board Games Research. Leiden: 48-
Galt, John, 1813. Letters from the Levant, containing Views of the State of Society,
Manners, Opinions, and Commerce, in Greece and several of the Principal Islands
of the Archipelago. London.
Gatti, Giuseppe, 1904. Roma. Nuove scoperte nella città e nel suburbio. In: Notizie
degli Scavi di Antichità, 1904: 153-157.
Heisel, Joachim P., 1993. Antike Bauzeichnungen. Darmstadt.
Kallipolitis, Vassilios, 1963. Ανασκαφη ταφων Αναγυρουντος. In: Archaeologikon
Deltion, 18, vol. 1: 123-124, 172-173.
Kendall, Timothy, 1991. Le jeu des “vingt cases”. In: Jouer dans l’Antiquité, Catalogue
of the Exhibition at Marseille 1991/92. Marseille: 148-152.
Kirkbride, Diana, 1966. Five Seasons at the Pre-Pottery Neolithic Village of Beidha in
Jordan. In: Palestine Exploration Quarterly, 98: 8-72.
Körte, Gustav, 1897. Etruskische Spiegel, ed. by Eduard Gerhard, vol.5. Berlin
Lamer, Hans, 1927. Lusoria tabula. In: Pauly and Wissowa, Realenzyklopädie des klas-
sischen Altertums, XIII 2. Stuttgart: Vol. 39 col.1970-1973; Vol. 54b col.1996-
Liepmann, Ursula, 1988. Corpus Speculorum Etruscorum, Bundesrepublik
Deutschland 2. München
Lund, John, and Bundgaard Rasmussen, B., 1995. Guides to the National Museum.
The Collection of Near Eastern and Classical Antiquitites. Greeks, Etruscans,
Romans. Copenhagen.
Mansuelli, Guido Achille, 1943. Materiali per un supplemento al “Corpus” degli spec-
chi etruschi figurati. In: Studi Etruschi, 17: 487-521.
Mansuelli, Guido Achille, 1945. Li Specchi figurati etruschi. In: Studi Etruschi, 19: 9-
May, Roland, 1991. Les jeux de table en Grèce et à Rome. In: Jouer dans l’Antiquité,
Catalogue of the Exhibition at Marseille 1991/92. Marseille: 166-189.
Murray, Harold J.R., 1952. A History of Board Games other than Chess. Oxford.
Murray, Margaret A., and Ellis, J.C., 1940. A Street in Petra. London.
Pankhurst, Richard, 1971. Gabata and Related Board-Games of Ethiopia and the Horn
of Africa. In: Ethiopia Observer, 14, no.3: 154-206.
Pankhurst, Richard, 1982. Gabata and Other Board-Games of Ethiopia and the Horn
of Africa. In: Azania, 17: 27-42.
Parker, Henry, 1909. Ancient Ceylon: An Account of the Aborigines and of Part of the
Early Civilization. London.
Petrie, W.M. Flinders, 1927. Objects of Daily Use, with over 1800 figures from
University College, London. Reprint London, 1974.
Pritchett, W. Kendrick, 1968. “Five Lines” and IG I2324. In: California Studies in
Classical Antiquity, vol.1: 187-215.
Pusch, Edgar B., 1979. Das Senet-Brettspiel im Alten Ägypten, Teil 1.1: Das inschrift-
liche und archäologische Material. München.
Rollefson, Gary O., 1992. A Neolithic Game Board from ‘Ain Ghazal, Jordan. In:
BASOR 286: 1-5.
Roueché, Charlotte, 1993, Performers and Partisans at Aphrodisias. Journal of Roman
Studies Monograph 6. London.
Roueché, Charlotte, and Bell, Robert C., in press. Pavement signs and gameboards in
the Greek and Roman periods: a British Museum typology. In: Finkel, Irving (ed.),
British Museum Colloquium on Ancient Games, London 1990. London (forthco-
Schädler, Ulrich, 1994. Latrunculi - ein verlorenes strategisches Brettspiel der Römer. In:
Homo Ludens. Der spielende Mensch IV. Salzburg: 47-67.
Schädler, Ulrich, 1995. XII Scripta, Alea, Tabula - New Evidence for the Roman History
of “Backgammon”. In: Alexander J. de Voogt (ed.), New Approaches to Board
Games Research. Leiden: 73-98.
Townshend, Philip, 1977. Mankala Games. In: Bulletin of the International Committee
on Urgent Anthropological and Ethnological Research, 19: 47-54.
Townshend, Philip, 1979. African Mankala in Anthropological Perspective. In: Current
Anthropology, 20 no.4: 794-796.
Townshend, Philip, 1979b. Mankala in Eastern and Southern Africa: a Distributional
Analysis. In: Azania, 14: 109-138.
Townshend, Philip, 1980. Games of Strategy: A New Look at Correlates and Cross-
Cultural Methods. In: Helen B. Schwartzman (ed.), Play and Culture. West Point:
Voigt, Eva-Maria, 1971. Sappho and Alcaeus, Fragmenta. Amsterdam.
Walker, Roslyn A., 1990: Sculptured Mankala Gameboards of Sub-Saharan Africa.
PhD-thesis Indiana University.
Walters, Henry B., 1899: Catalogue of the Bronzes, Greek, Roman, and Etruscan in the
Department of Greek and Roman Antiquities, British Museum, London.
1. See Schädler 1995.
2. Townshend 1979: 794.
3. Townshend 1977: 51.
4. See a.e. Pankhurst 1971 and 1982, Townshend 1979b; critical Eagle 1995: 51-52.
5. I know of only two exceptions. One is a complete set of XII scripta/alea found in a grave Qustul,
Nubia (Schädler 1995: 80). Recently in a British grave at Stanway, Colchester, the remains of
a wooden board of presumably 12x8 squares with thirteen white counters, one of which was
smaller in size, and thirteen black counters, one of which was used upside down, still in place
have been found (Crummy 1997a: 68-70, Crummy 1997b: 1-9). The game played on the
board is probably not the Roman Latrunculi game, mostly because of the following reasons
(compare Schädler 1994): 1) in the written sources about Latrunculi no extra piece is men-
tioned; 2) the ratio of Latrunculi boards is more or less square, the boards having normally
nearly the same number of squares horizontally and vertically; 3) 12 pieces for each player
seems too small a number in a game where an enemy piece must be captured by enclosing it
from two sides; 4) in Latrunculi probably there was no initial position, if the position of the
pieces on the Stanway board indeed reflects the beginning of the game.
6. See Murray 1952: 48-50; Townshend 1980: 218.
7. For siga see Petrie 1927: 56-57, Murray 1952: 54-55, 82; for latrunculi see Schädler 1994.
8. See Schädler 1995.
9. Roueché in press.
10. In the list below the type denotations of the British Museum Typology of gameboards (BMT),
developped by Charlotte Roueché and Robert C. Bell, are given in brackets. As the BMT does
not distinguish my types B3 and C3 from type D3, boards of these types may be included in
the boards nos. 28-51. For the BMT see Roueché/Bell 1993: 249-251.
11. Completely misleading Becq 1869: 397-405 and Murray 1952: 28; disappointing May 1991:
12. Bergk 1884: 177 no.82, Voigt 1971: 320 no.351.
13. Austin 1940: 267-271.
14. Pritchett 1968: 189-198.
15. For the discussion see Pritchett 1968: 200-201.
16. Kallipolitis 1963: 123-124 pl. 53-55.
17. Blinkenberg 1898: 2-5 fig.1-8; Pritchett 1968: 189-191 nos.1-3.
18. Lamer 1927: 1971.
19. Lund/Rasmussen 1995: 67; Pritchett 1968: 197 pl.7,1; Breitenstein 1941: 19 no.171 pl.19.
20. Blinkenberg 1898: 9.
21. 3rd century BC: Mansuelli 1945: 58.
22. Körte 1897: 144-146 pl.109.
23. Körte 1897: 191-193 pl.146; Walters 1899: 377 no.3213.
24. Compare the mirrors Mansuelli 1943: 517-518 pl.40 no.13 (2nd half of the 3rd cent. BC),
Liepmann 1988: 43-45 no.17 (early 3rd cent. BC), and de Puma 1987: 38-39 no.21 (early
3rd cent. BC).
25. Walters 1899: 377; Bell 1979: 30 fig.25; May 1991: 179 fig.174, who wrongly dates it to
Roman times.
26. Schädler 1995.
27. See Schädler 1995: 84.
28. Wrongly identified as a ground-plan (!) by Heisel 1993: 193.
29. See Lamer 1927: 1999.
30. Chavane 1975: 195 pl.53 and 72 no.575, 197 pl.53 no.576, pl.54 and 73 no.577, 204 fig.12
and pl.54 no.578. Chavane fell into the same trap described above identifying the boards as
“jeux des douze lignes”, i.e. XII scripta/alea boards. But the boards are clearly 2x11-squares
31. Murray 1940: 35 fig.10, incised in the rock together with two boards of 2x10 squares.
32. Caputo 1987: 121 pl.94,3.
33. Akurgal 1993: 122-123.
34. Gatti 1904: 153-155 fig.2 and 4; for the other gaming boards see Schädler 1995: 89 fig.11a,
94s. fig.12a. For gaming boards with leaves see Schädler 1995: 87 fig.6a and 6h, 88 fig.7.
35. Chavane 1975: 197 pl.54 and 204 fig.12 no.578.
36. Pankhurst 1971: 154; Walker 1990: 37.
37. Parker 1909: 579-644; Murray 1952: 160-161.
38. See Walker 1990: 34-35.
39. I quote from a letter dating 14th of July 1996 by Edgar B. Pusch, Hildesheim, to the present
writer: “Das Mankala-Spiel ist mir – trotz Murray – aus Alt-Ägypten nicht bekannt. Zwar
gibt es Ritzungen auf Tempeldächern und in Höfen, welche an eine entsprechende Aufteilung
erinnern, jedoch ist die Zeitstellung völlig offen und vermutlich sehr spät (Islam).” On the
roof of the temple at Kurna there are also siga or dara boards, certainly of rather late date.
40. Petrie 1927: 55 no.175 pl.74; Walker 1990: 35. The unique small scale clay models of gaming
tables from Perachora in Greece showing three rows of at least six cup-shaped depressions and
dating to the 7th and 6th centuries BC (Dunbabin 1962: 131-132 pl.39 and 132 nos. 1325-
1328) have never been regarded as mancala-boards and are likely to be associated with
contemporary senet boards showing three rows of circles (see a.e. Petrie 1927: 53-55 pl.48
no.3, no.4 = Pusch 1979: 374 pl.94 no.80, 376 pl.95 no.81; Kendall 1991: 151 fig.145 wron-
gly identified as “game of twenty squares”), certainly not with XII scripta/alea as Dunbabin
1962: 132 took it.
41. Rollefson 1992: 1-4 fig.1.
42. Kirkbride 1966: 34 fig.8.
43. Pankhurst 1982: 35; Townshend 1979b: 119-122.
44. Galt 1813: 242; Murray 1952: 158.
45. Townshend 1977: 47.
Inca Dice and Board Games / Thierry Depaulis
Very little is known of the games played in Precolumbian Andean cultures.
Significantly none is mentioned in H. Murrays History of board games other than
chess (Murray 1952) or in R.C. Bell’s Board and table-games from many
civilizations (Bell 1979) where we find a wide scope of games, ancient and modern, from
all over the world. The Aztecs are known to have had ball and dice games, notably the
patolli, a race game. Did the Incas really play? Can we use for their games the same
sources as we have for other extinct civilisations? It is the object of this paper to present
a sketch of the games that were played by the Incas and some other Andean peoples.
It is through folk funerary rites that Americanists like Nordenskiöld, Karsten, or
Rivet have encountered games that were still played by the Indians in some remote
villages of Peru or Ecuador. There they observed the use of a very typical pyramidal die
and the practice of mock gambling for distributing the defunct’s possessions. They
realised that these games were probably also played in Inca times and they tried to search
for more details. Although they have found some interested sources, their enquiries both
ethnological and textual are not satisfying.
The first scholar who tried to survey the games of the Incas was Emilia Romero, who
published an article in 1941 then a booklet printed in Mexico in 1943 (Romero 1943).
She mainly used the so-called Spanish ‘chroniclers’ to present what she had collected.
But she could not explain what the games were. A games expert’s eye was necessary. By
comparison with other dice and board games from other civilisations, it is nevertheless
possible to trace what the games the Incas played looked like. Archaeological finds can
be taken into consideration too, although one of them is much discussed and is more
probably an abacus. Ethnological surveys can also help understanding how a game was
I. Spanish Chroniclers and Others
The Andean civilisations had no writing. So no written record can give us the
description of the games that were played before Columbus in South America. And,
contrary to Middle America, pictures are missing here too: the Inca culture did not
favour representations of its daily life. Therefore we have to rely first on the records the
Spanish made of what they had noted. We must keep in mind that their accounts can
only deal with the Incas, late-comers who ruled over what is now Peru and Ecuador, and
parts of Bolivia and Chile during the 15th century and the beginning of the 16th. They
had been preceded by other civilisations like the kingdom of Chimú (10th to 14th
century), and previously by the civilisations of Wari and Tiahuanaco (7th to 9th
century), Moche and Nasca (100 BC - 600 AD), to name but a few. We know almost
nothing of the games these peoples played, except for the Mochicas who have left us
many representations of their everyday life on their delicately painted vases.
At the beginning of this century very few Spanish texts related to ancient Peru were
available in print. The conquistadors were not interested in Inca culture, and their
accounts deal only with what they did. The only source which gave a comprehensive
description of the Incas’ daily life was Father Bernabé Cobo’s Historia del Nuevo Mundo,
written in 1653 but printed only in 1890-3, ‘the best and most complete description of
Inca culture in existence’ according to J.H. Rowe. Here we find a chapter (Bk. XIV, ch.
17) entitled ‘De los juegos que tenían para entretenerse’ (‘Of the games they have for
entertaining themselves’). Martín de Murúa’s Historia del origen y genealogía real de los
reyes incas del Pirú, our second most important source, was published only in 1922-5 in
a poor edition printed at Lima (even the author’s name was wrongly given as Morúa!).
Waman Puma’s (or Guamán Pomas…) extraordinary illustrated manuscript was still
lying in the dust of the Danish Royal Library.
Things changed just before World War II. In 1936, Felipe Waman Puma de Ayala's
Nueva crónica y buen gobierno (written about 1615) was printed at Paris in a facsimile
edition, although without notes or index, and, in 1946, C. Bayle gave a correct edition
of Murúa's Historia from the original manuscript held in the Jesuit Archives. Even
Father Cobo's works were published in a better edition in 1956. Soon another, longer
version of Martín de Murúa's Historia was discovered in the Wellington Papers and
published in 1962-4. Unfortunately games are no longer present. At last an annotated
(and indexed) edition of Waman Puma's Nueva crónica was published at Mexico-City in
These sources are not very informative. The celebrated half-Inca half-Spanish writer
Garcilaso de La Vega has a brief allusion to gaming among the Incas in his Comentarios
reales de las Indias (1609). Of the already mentioned authors, Waman Puma offers two
puzzling lists and an interesting drawing (which we shall examine further). Only Cobo
and Murúa (in his shorter earlier version) give broad descriptions of actual games,
though their accounts do not tally.
To these texts Emilia Romero had the excellent idea to add the numerous dictionar-
ies of the local languages, Quechua and Aymara, which were compiled in the late 16th
and early 17th centuries. Spanish missionaries quickly understood that, if they wanted
to convert the Indians, they had to speak their languages. Fortunately the Incas had
already imposed one language known as Quechua, that the Spaniards named ‘la lengua
general del Inga’. (Quechua is still living and is today spoken by millions of speakers in
Peru and Bolivia.) They rapidly drew up grammars and dictionaries – as early as the mid-
16th century, whereas the European languages were not so well treated.
All these dictionaries offer entries to games with short descriptions in Castilian. The
earliest of these is Domingo Santo Tomás’s Quechua dictionary, entitled Lexicon o
vocabulario de la lengua general del Perú (Valladolid, 1560). Then came Arte y vocabulario
en la lengua general de Peru llamada quichua, printed in Lima in 1586, and assigned to
Juan Martínez. The most important Quechua dictionary is Diego González Holguín’s
Vocabulario de la lengua general de todo el Perú, llamado lengua qqichua o del Inca (Lima,
1608). It was followed by Diego de Torres Rubio’s Arte de la lengua quichua (Lima,
1619). Although Quechua was the most widespread language of the former Inca empire,
Aymara, a close relative spoken farther south, was not forgotten. The Italian Father
Ludovico Bertonio published his Vocabulario de la lengua aymara at Lima in 1612. It is
by far the most informative source for some of the games we are looking for.
All these sources offer words only. And their descriptions, if any, are very confusing.
Strangely we get the impression that the Incas had a wide variety of dice and board
games: about a dozen of names can be listed. This may be deceptive since, contrary to a
general belief, Quechua was in fact – and still is – split into several dialects of which ‘la
lengua general del Inga’ was the dominant one. Obviously there are synonyms owed to
the different dialects the Spaniards came in contact with. Enough is said, however, to
broadly categorise the games I have collected: clearly there are pure ‘dice’ games, race
games and hunt/war games. I will describe them in this order, giving the different names
and different spellings that are to be found. (Unfortunately there is no standardised
spelling for Quechua, and every author from the 16th century on has his own spelling!)
As far as the Incas are concerned, no iconography can help us (save in one case), and the
archaeological finds are so rare that they can hardly be used – but I will further discuss
some of them. Ethnology, which was so helpful in understanding the Aztec games, is
here very disappointing. However, as we shall see, comparisons with neighbouring
cultures, like that of the Mapuches, or Araucanians, offer some clues to understanding
how some of the Inca games were played.
II. Inventory of the Inca Games
A.The Inca Die
pichqa (pichka, pisca, pichca, pichiqa)
wayru (huairu, guayro)
The Incas had a very special six-sided pyramidal die which they used for pure gambling
as well as for race games. It is the only game to be mentioned and described by all
chroniclers and in all dictionaries, though under two different names: pichqa (Quechua
five’) and wayru. Inca dice have also been found in archaeological diggings, and
ethnologists have collected modern examples. (Fig. 1 and 2)
After López de Gómara (1552), Father Cobo (Cobo 1653) says that ‘el llamado
Pichca era como de dados: jugábanlo con un solo dado de 5 puntos’. Murúa (Murúa
1590, II, 13) describes it as a teetotum (‘como una perinola’), adding further ‘the Indians
play with one die, called pichca, with 5 points on one side, 1 on the other, 2 on the other
and 3 on the other, plus side 4; the crossed top is 5, and the bottom of the die is 20’ (III,
25). For González Holguín (1608), pichkana is a ‘six-sided piece of wood’ (‘Ppichca. Un
juego como de dados. — Ppichcana. Un palo seizavado con que juegan.’). This is exactly
what the ethnologists (Rivet 1925; Karsten 1930; Hartmann & Oberem 1968;
Hocquenghem 1979) and the archaeologists (e.g. at Machu Picchu: see Bingham 1915
b, p. 176; 1930, fig. 172, b-h, quoted in Rowe 1946) have found.
The pichqa die is played either as a simple dicing game, where the winner must make
the maximum score, or as part of a race game (Cobo 1653, XII, 15; Murúa 1590, II,
Another name for the Inca die is wayru, but no satisfactory explanation has been
given for this differentiation. González Holguín (1608) presents both names as
synonyms: ‘Huayru, Ppichca. Juego de los naturales.’ adding: ‘Huayru. El mayor punto,
o el mejor que gana’. Santo Tomás (1560) has guayroni meaning ‘jugar juego de fortuna
or ‘jugar juegos de dados’. The Vocabulario of 1586 defines huayru as ‘un tanto, o azar al
juego de los Yndios’ and pichca, pichcana as ‘un cierto genero de juego de Indios’. In
Aymara too the words are roughly the same: Bertonio (Bertonio 1612) offers huayrusitha
and piscasitha for two very similar board games (Romero 1943: 19).
Wayru/pichqa was obviously played in relationship with funerals. Already in the early
17th century Father Arriaga (Arriaga 1621: VI, 60) had noted that piscawas played to
keep watch over the body of a dead (with small scoring sticks, palillos con diversas rayas).
Juan de Velasco (Velasco 1789: II, 152) states that the five-sided huayru (‘gran dado de
hueso con cinco puntos’) was preferably played in November. This ritual use of the
wayru die was noted by many ethnologists (Rivet 1925; Nordenskiöld 1910; Karsten
1930; Roca Wallparimachi 1955; Cavero 1955; Hartmann & Oberem 1968). Rivet
thought that pichqa was kept alive in Peru whereas wayru prevailed in Ecuador.
The Inca die is not necessarily very old: the Mochicas, whose games are represented on
painted vases, seem to have nothing like this. Instead, they used two-colour beans. It is
interesting to note that wayru was originally the name of a red-and-black bean
(Erythrina americana).
Amazingly the dictionaries offer alternative names for the Inca die. According to
Bertonio 1612, chunka (Quechua ‘ten’) is its Aymara counterpart: ‘Chunca: Tagua de
madera para jugar.’ The same dictionary has ccanccallu (kankallu, kancalla) for the
wooden die; in modern Aymara kancalla means ‘knucklebone’ (De Lucca 1983: ‘taba,
astrágalo, hueso del pie’). Paul Rivet has pointed out that the die is also called tawa,
tahua, tagua (Quechua ‘four’) in the Cuzco dialect; he saw in this word the origin of the
Spanish word taba (Rivet 1925). But this is much discussed, since for others the term
derives from the Arabic tâb, also a game term (see J. Corominas and J.A. Pascual,
Diccionario crítico etimológico castellano e hispánico, Vol. V, Madrid, 1983).
B. Board Games
From literary evidence, we know the Incas had several board games, most of them of the
race game type. Let us first examine the race games.
B.1. Race Games
Our ‘chroniclers’ offer about six different names for what clearly appear to be race games,
i.e. board games played with a die. Some of them must have been synonyms. If chunkana
(from Quechua chunka ten’) is often mentioned, aukay, takanako, halankola / hunkuña
appear more rarely. Pichqa and wayru also seem to have designated actual board games.
All these games make use of one ‘Inca die’.
aukay (awkai, aucai) (Murúa 1590) = takanako?
(Romero 1943: 23)
Murúa 1590 (II, 13) presents this game: ‘llaman también aucai, que en una tabla con
fríjoles de diversos colores y dificultoso en jugar, también echando los puntos con la
Fig. 1: Pichqa die
found at Macchu
Pichu by H. Bingham
Fig. 2: Drawing of a modern
wayru die used at Sigsig
(Ecuador) (from Hartmann
& Oberem 1968)
pisca como queda ya dicho, el cual es un juego muy gustoso’ (‘aucai is a board with
multicoloured beans; it is difficult to play, points are counted with the pisca die…; it is
a very nice game’). In classical Quechua auccay means ‘war’ (González Holguín 1608).
Although Murúa is the only author who mentions this game, his account is very close
to that given by Cobo (Cobo 1653) for takanako. In Aymara, according to Bertonio
1612 (s.v. ‘Juego’), the word aucattana designates the halankola board.
chungani, chuncani (Santo Tomás 1560; Torres Rubio 1619), chuncana (Vocabulario
1586), chuncaycuna, ccullu chuncana (González Holguín 1608), chuncara (Cobo
(Romero 1943: 22)
A race game played on five squares with multicoloured beans and the pichqa die. Scoring
is by tens, from 10 to 50. Although a complete description is lacking, it is the best
documented race game. In Santo Tomás 1560 we find chungani meaning ‘juego de for-
tuna’ and ‘jugar a… los naypes’ (to play at cards). The Vocabulario of 1586 gives
‘Chuncana. qualquier juego de fortuna’, and ‘Cullu chuncana. ajedrez, o tablas, &c.’ (i.e.
both chess and backgammon – from ccullu = ‘wood’). González Holguín (1608) repeats
all this (s.v. chuncaycuna and ccullu chuncana) and adds chuncana cuna gaming in-
strument’ (Lara 1978: chunkanakuna ‘instrumentos u objetos que se emplean en un
It is once more Bernabé Cobo (Cobo 1653) who gives the most detailed account of
this game: chuncara [wrong spelling for chuncana?] was another game with five little
holes dug in a flat stone or table: they played with multicoloured beans; when the die
was cast, and according to lots, they move them from one square to another until the
end; the first square was worth 10 and the others went increasing by tens up to the fifth
which was worth 50.’ (‘Chuncara era otro juego de cinco hoyos pequeños cavados en
alguna piedra llana o en tabla: jugábanlo con frisoles de varios colores, echando el dado,
y como caía la suerte, los mudaban por sus casas hasta llegar al término: la primera casa
valía 10, y las otras iban creciendo un denario hasta la quinta, que valía 50.’) Note that
Murúa is silent about this game.
Chunka seemed to have had a special meaning for the gamesters. According to
Garcilaso de La Vega (1609: II, 62), ‘llaman chunca a cualquier juego… (…) Tomaron
el número diez por el juego…’, they held the numeral ‘ten’ for gaming. It was also one
of the names of the die in Aymara (Bertonio 1612).
halankola, jalankola or hunkuña, juncuña
halancola, halancolatha, halancolasitha (verb) = huncusitha (verb), huncosiña
(Bertonio 1612), hilancula (Waman Puma 1615)
(Romero 1943: 22)
According to Bertonio 1612 halankola was a race game somewhat resembling
backgammon, played with the pichqa die on a board called aucattana (‘se parece algo al
de las tablas, y van adelantando las casas con estas palabras: Halancola; y a su traza
llaman, aucattana, y al dado de madera que usan, pisca; y a los agujeros o hoitos del
juego les dicen Halancola.’, see under ‘Juego’). There are other entries for this game in
Bertonio’s dictionary:
‘Halancola. Los agujeros, o hoytos, de un juego assi llamado que algo se parece al de las
tablas.’ (s.v. Halancola) [Halancola = the little holes (i.e. squares) of a game so-called
which resembles backgammon]
‘Huncusitha. Jugar como a la tagua con un dado grande de madera, adelantando unas
piedrecitas en sus casas u hoyos; lo mismo que el halancolatha(s.v. Huncusitha)
[Huncusitha (verb): To play as with knucklebones with a large wooden die, moving
some small stones in the squares or holes]
At the same time Felipe Waman Puma de Ayala offers a list of games that were played
by the ‘noble lords’ in April (Waman Puma 1615: 243). Among these was hilancula
(which here too seems to need a pichqa die). There is another mention further (Waman
Puma 1615: 780), in another list of games with cards, dice, chess, then ‘hilancula, chalco
chima’, etc. Again in association with pichqa, hilancula is quoted a third time (Waman
Puma 1615: 844: ‘con la hilancula, pichica’).
pasa (from Quechua pachak ‘100’)
(Juan & Ulloa 1748: VI, 6, no. 941 = II, 549; see also Culin 1898: 805)
A race game observed c. 1735-44 among the Aymara Indians by Jorge Juan and Antonio
de Ulloa, played on a wooden board (or a counter?) in the shape of a double-headed
spread eagle of wood with 10 holes on each side, with pegs and a seven-face die, one of
the faces being called guayro (= wayru). The game was won by the first player who
reached 100. According to Juan & Ulloa, it was ‘el único que los Indios del Perú suelen
jugar’ (Index). Maybe a form of wayrusitha.
pisca (Murúa 1590), piscasitha (Bertonio 1612)
wayrusitha (Bertonio 1612)
(Romero 1943: 19)
Obviously pichqa and wayru also meant a race game. Murúa (1590) mentions ‘another
game’ played with the pisca die and with ‘its table and its holes or marks, where they
move their men’ (‘que es muy ordinario questos Indios llaman la pisca con su tabla y
agujeros o señal, donde iban pasando los tantos.’). Bertonio 1612 gives two verbs
huayrusitha and piscasitha with the meaning ‘to play with little stones, by moving them
in their holes [squares] according to the score of a kind of large die [huayru or pisca]; in
one of these games, they move the stones all around or in circle; in the other [game],
they go winding about like a river.
(‘Huayrusitha, Piscasitha. jugar con unas piedrecillas adelantandolas en sus hoytos, segun
los puntos de una manera de dado grande; en unos destos juegos van adelantando las
piedras alderredor o en círculo; en otros dando buelta como río, &c.’)
Such a game may well fit what Roswith Hartmann and Udo Oberem watched in 1965
at Sigsig (Ecuador) under the name of huairu. It is a race game played with the
pichqa/wayru die, on a board with 29 holes arranged in a somewhat triangular circuit
divided by a central line (see Fig. 3). An unstated number of men (maize ears, beans,
etc.) are used. The tracks of the two players are not identical: the shorter track (1) has
10 cells, the longer track (2) has 20 cells. The first player who gets the crossed side of the
wayru die starts. The first who reaches the centre wins. Unfortunately Hartmann and
Oberem have forgotten to complete their information so we have no other detail nor do
we know what happens when two opponents meet.
We will find a somewhat similar arrangement of the holes in a race game played by
the Mapuches (Araucanians), called kechukawe (see further). As we shall see, the
Mapuche game, which uses the same pyramidal die, may well be a close relative of the
Inca game.
Fig. 3: ‘Huairu
board as drawn by
Hartmann &
Oberem (1968).
The two players
have different
takanako (Cobo 1653) = aukay?
(Romero 1943: 22)
A race game ‘like backgammon’, with multicoloured beans and the pichqa die.
Cobo (1653) writes: ‘another kind of game was called tacanaco; it was played with the
same die and with multicoloured beans, like backgammon (como el juego de tablas)’.
Cobo is the only one who mentions this game but his description looks very much like
that of aukay mentioned by Murúa. Romero 1943 identified takanako with
B.2. Strategy Games
Besides race games, some sources account for the existence of a ‘strategy’ game (without
dice) under the names of taptana or komina. These ‘Inca chess’ are in fact a ‘hunt game’,
played on an ‘alquerque’ board with a triangle added to one side.
komina, cumi
comina (Santo Tomás 1560), cumisiña (Bertonio 1612)
(Romero 1943: 24)
A hunt game. Comina appears in Santo Tomás's dictionary of 1560 as a synonym of
taptana: Taptana, o comina: axedrez, tablas, o alquerque’ (‘chess, backgammon or
alquerque’!) and ‘Comina, o taptana: alquerque’. Bertonio 1612 has an entry under
kumisiña or kumisitha: ‘Cumisitha. Jugar a un juego como al que llamamos oca, aunque
en muchas cosas diffiere. — Cumisiña. Juego assi.’ (‘Cumisitha. To play a game like the
one we call the game of the goose, although it is different in many things’); he also gives
kumisiña as an equivalent of Spanish alquerque and ajedrez: Alquerque. Cumisiña, y lo
mismo significa Axedrez, porque los Indios no distinguen los juegos, si no miran al
modo.’ (‘and the same means chess, because the Indians make no distinction between
games if they do not watch how they are played’). In a modern Quechua dictionary (De
Lucca 1983) cumi is defined as ‘juego que en castellano se llama el león y las ovejas’,
which is indeed a hunt game equivalent to the English fox and geese. El león y las ovejas
(literally ‘the lion and the sheep’) is a popular game in South America where león stands
for puma (there are no native lions in America!).
puma (González Holguín 1608, Cobo 1653, Torres Rubio 1700)
(Romero 1943: 28)
There is also a board game called puma.
Unfortunately the rare accounts we have are quite uninformative. González Holguín
(1608) writes: ‘Puma. Un juego de Indios. — Pumani. Jugar a este juego.’, and Torres
Rubio (1700) has: ‘Puma. Cierto juego de Indios’. Cobo (1653) mentions it, together
with apaytalla, at the end of the ingeniosos games; for him, apaytalla and puma are ‘less
prestigious’ (menos principales)
We know puma also means the Andean ‘lion’, and it is tempting to connect the game of
puma with el león y las ovejas, in other words with komina. We will see that the Mapuches
call their game komikan, el leoncito.
Fig. 4: The Inca emperor
Atahualpa playing
taptana in jail with his
guard (from Waman
Puma 1615).
taptana (Santo Tomás 1560; González Holguín 1608; Waman Puma 1615), tapta
(Vocabulario 1586), probably also atapta (Murúa 1613: II, 89°, p. 323)
The game of taptana must be a war or a hunt game, if we trust the equivalents offered
by the Spanish lexicographers. In modern Quechua taptana means ‘chess’.
Santo Tomás 1560, the Vocabulario of 1586 and González Holguín 1608 all have an
entry for a game called taptana that they render as ‘alquerque’ or ‘ajedrez’. The
Vocabulario of 1586 adds that taptana means ‘chess-board’, using tapta for the game
itself. More interestingly Santo Tomás 1560 gives to taptana the synonym comina. In his
‘long’ and later version, Martín de Murúa mentions only one game that he calls atapta
que es como a las tablas reales’ (‘like backgammon’); it was played by the Inca Tupac
Olaf Holm (Holm 1958) has rightly remarked that Waman Puma alludes to the
game played by the last Inca emperor Atahualpa in the Cajamarca jail, before his death
in 1533. According to the Nueva crónica, Atahualpa ‘played chess [ajedrez] that they [the
Indians] called taptana’. A drawing shows the scene (Waman Puma 1615: 388 [390]),
with Atahualpa in chains in front of his guard. What is at first sight an alquerque board
lies between them (Fig. 4).
We now know that taptana is also called komina (and perhaps puma too…), and that
it is a hunt game known today as el león y las ovejas. As we shall see, the same game is
played by the Mapuches under the name of komikan, probably a cognate to Quechua
Figs. 5a-b: Taptana
boards scratched on
Pre-Hispanic walls at
Chinchero (near
Cuzco) (from Alcina
Franch 1980).
komina and Aymara kumisiña. The Mapuche komikan has the same latticed board as
alquerque with a triangle added on one side. It is possible to imagine that it was the game
intended by Waman Puma in his drawing.
Archaeologists have been fortunate enough to find the same design scratched on the
Pre-Hispanic walls of the square of the church at Chinchero (near Cuzco). These graffiti
were discovered and analysed by J. Alcina Franch (Alcina Franch 1980). One of them
(no. 32 – Fig. 5a & Fig. 6) is without doubt a taptana/komina board; another one (no.
37 – Fig. 5b) looks like a ‘spoiled’ board. Although the author dates these graffiti in the
17th century, he demonstrates that they are related to Precolumbian traditions. Stewart
Culin too reports a Peruvian game called solitario (Culin 1898: 876, fig. 183; also
Murray 1952: 100, no. 5.2.1) which shows the same triangular appendix (Fig. 7).
Is taptana/komina an indigenous game? The board illustrated by Waman Puma de
Ayala is nothing but an ordinary Spanish alquerque board, either used for a war game
(alquerque de doze) or for hunt games (cercar la liebre, castro). Moreover all our sources
are later than Columbus. Other such war games are known in Europe and in Asia, as
well as in Mexico and in the South-West culture of the United-States (Keres, Zuñi,
Pima, Papago, and Hopi Indians: see Culin 1907: 794-5; Murray 1952: 67). However,
triangular appendices only appear in Southern Asia. Hunt games on the same board also
exist, but added triangles are known only in China and Japan (Murray 1952: 100-101).
These triangles, always in pairs, one on each opposite side, are definitely different: the
Fig. 6 (left): Schematic
drawing of a taptana
board (from Alcina
Franch 1980).
Fig. 7 (right):
Board game for
solitario, Peru, late
19th century
(after Culin 1898).
South-American game cannot derive from these Asiatic forms. Since they cannot have
borrowed their appendix from Europe, it is reasonable to think this game is
C.Apaytalla:The Game with Beans
In his celebrated Comentarios reales Garcilaso de La Vega (La Vega 1609) says that besides
edible beans, the Incas knew other kinds of beans called chuy, que no son de comer’ (‘not
edible’), round, of different coulours and of small size (‘del tamaño de los garbanzos’, i.e.
like chickpeas), to which they gave ridiculous or ‘well suited’ names and which they used
for many children and adult games. Garcilaso remembered that he himself used to play
with these beans (Romero 1943: 14).
apaytalla (Murúa 1590; Cobo 1653)
(Romero 1943: 23)
Murúa 1590 mentions a game called apaitalla, using beans ‘of different kinds and
appellations’, cast on the ground from the top (‘con la cabacera alta’), with lines and
arches like furrows (‘rayas y arcos a manera de surcos’); the winner was the player who
‘went ahead and was the noisiest [!]’… A legend attributed the invention of the game to
Fig. 8: Cutaway
drawing of a Mochica
vase (first centuries AD)
showing gods
(or heroes) playing with
Queen Anahuarque.
(‘es un género de fríjoles redondos de diversos géneros y nombres, e hizo en el suelo con
la cabacera alta de donde sueltan los tales fríjoles, y el que de ellos pasa adelante y hace
ruido más, gana a los otros, está con sus rayas y arcos a manera de surcos’)
For Cobo (1653), it was just a ‘less prestigious game’. The lexicographers (Santo
Tomás, González Holguín, Bertonio) have no entry for that word or for any similar
game. González Holguín (1608) only has ‘Chuui o chuy. Unos frisoles muy pintados
como garvanzos y otros menores larguillos.’ (‘multicoloured beans, like chickpeas and
other smaller beans’).
It is very tempting to identify this Inca game of apaytalla with the game shown on
many Mochica potteries (c. 100 BC-c. 600 AD). On these gaming scenes players are
shown handling multicoloured beans and waving sticks which were more probably used
for keeping the score than as dice, since each player has his own set (Fig. 8). The
undulating ground seems to be made of sand; beans are put in hollows as well as on
ridges. The Mapuches have such a game, called lligues (llüqn) or awarkuden (see further).
Although these Mochica scenes have been interpreted as ‘writing workshops’ by R. Larco
Hoyle (who tried to demonstrate that writing was known in Precolumbian South-
America), they are now understood as gaming scenes (Vivante 1942; Romero 1943: fig.
1; lastly Hocquenghem 1979).
This implies that prior to the using of the rather sophisticated Inca die, the peoples
who lived in the Andean area used beans as dice, like the Aztecs.
D. Unidentified Games
Waman Puma 1615 offers two lists of games played by the ‘noble lords’. Some of these
names are unknown elsewhere and cannot be explained. I have nevertheless decided to
publish them here.
p. 243 [245]: ‘todo el mes [abril] juegan los señores principales al juego de riui, choca,
al uayro de ynaca, pichica de hilancula y de challco chima’ (1987 edition).
p. 766 [780]: ‘y se enseñan a jugar con naypes y dados como españoles, al axedres,
hilancula, chalco chima, uayro, ynaca, riui, pampayruna, yspital, uayro ynaca [sic]’
(1987 ed.).
riuichoca (riwichuqa) (Waman Puma 1615: 243 [245])
A throwing game, today known as riwi or lliwi (Lara 1978) and, in Spanish, as bolea-
doras. It is an old hunting weapon made of three stone or lead balls (bolas), tied with a
cord. The balls are thrown as far as possible.
ynaca (iñaka?) (Waman Puma 1615: 243 [245]; 766 [780])
The meaning is unclear. In Quechua iñaqa (or iñaka) means ‘mantilla’ (González
Holguín 1608); but in Aymara it means ‘noble woman from the Inca caste’ (Bertonio
1612). It is possible to understand ‘al uayro de ynaca’ (= wayru [de] iñaqa) as ‘to the
wayru game of the noble Inca women’… In the second list, Waman Puma repeats uayro
cha[l]lco chima, challkuchima (Waman Puma 1615: 243 [245]; 766 [780])
‘Challcochima’ is the name of an Inca war lord (Challcochima, Challicuchima,
Challkuchimaq, Challkuchima…), supporter of Atahuallpa, victorious of Huascar, and
finally killed by the Spaniards. In Aymara kallko (now extinct) meant ‘five’ (De Lucca
pampayruna (Waman Puma 1615: 766 [780])
There is only one meaning for pampayruna: ‘prostitute’! González Holguín 1608 has:
‘Pampayruna. Muger pública comun a todos’; and Torres Rubio 1619: ‘Panpayruna.
Ramera’. In modern Quechua the meaning has not changed: ‘Panparuna. Prostituta
(Lara 1978)! Is that also the name of a game?
yspital (Waman Puma 1615: 766 [780])
Another puzzling word whose exact meaning is unknown (hospital?).
E. Game Board or Abacus?
Another artifact has sometimes been presented as ‘Inca chess’: this puzzling object,
which sometimes looks very much like a miniature castle, has alternately been
interpreted as an abacus, a model fortress, or a game board (Fig. 9). It is Nordenskiöld
who made the hypothesis it was a board game. His demonstration was attractive: the
Chaco Indians have a very simple race game called tsuka, chukanta, or shuke (from
Quechua chunka, ten’) which they play with throwing sticks as dice. (Actually this was
reported in the early 20th century.) Nordenskiöld inferred from this that this game was
borrowed from the Incas. Cobo’s reference to chunkara appeared as a good justification
for Nordenskiöld’s theory. Moreover when transposed on the Inca artifact the Chacoan
rules work!
However, Cobo’s description does not fit the Chacoan game at all, and it is hard to
believe that such a complicated multi-level object was used for a board game. More
recent investigations, undertaken by Carlos Radicati di Primeglio (Radicati di Primeglio
1979) have shown that it is in fact a yupana, the Inca abacus. Other objects, which have
been sometimes presented as board games, are clear relatives of this abacus (e.g. Holm
1958; Figge 1987). Although some scholars have tried to support Nordenskiöld’s theory
(Smith 1977; Pratesi 1994), there are good reasons to accept Radicati di Primeglio’s well-
documented demonstration. (facing page) Fig. 9:
Abacus or game
Plate from P. Rivet &
R. Verneau,
ancienne de
Paris, 1912.
III. The games of the Mapuches (Araucanians)
Because the study of neighbouring cultures can throw some light on the Inca games, I
have studied too the games of the Mapuches. The former Araucanians lived to the south
of the Inca empire and were partially conquered by them. Today they call themselves
Mapuches and live in the north of Chile. The Mapuches were influenced by the Inca
culture: it is no surprise if their games show strong similarities with the Inca games. By
chance, early descriptions of the Araucanians are very informative (e.g. De Ovalle 1646)
and all give detailed accounts on their games. Alonso De Ovalle has even illustrations
showing two games in action (Fig. 10).
kechukawe (quechucayu, quechucague)
According to De Ovalle’s description (De Ovalle 1646), quechucague is a race game
played on a semi-circular board, with segments of five squares each (see picture ‘Ludus
quechucague’ in Fig. 10 and my Fig. 11); the sole die looks like an elongated pyramid;
men are little stones. Rosales (post 1674, in: Pereira Salas 1947) says it is a gambling
game, but does not mention any circuit; he nevertheless describes a perched ring through
which this ‘triangular’ die was cast. Comparison with De Ovalle’s engraved plate shows
strong similarities between quechucague and the race game which Hartmann and
Oberem observed in the 1960’s in Southern Ecuador (Hartmann & Oberem 1968).
There the game was called huayru (wayru), which, as we know, is a synonym of pichqa.
Like pichqa, kechukawe is derived from the Mapuche word kechu meaning ‘five’!
There was in Francisco Fonck’s collection a ‘gaming stone’ with five little holes on
each side, which was found in the Group IV of El Retiro (Fonck 1912: 5). Unfortunately
the drawings that Fonck had prepared were not printed.
M. de Olivares’s Historia militar, civil y sagrada … del reino de Chile (written c. 1758,
quoted in Medina 1952) gives more detail about the Araucanian die: it is an ‘isoceles
triangle’ with faces bearing 1, 2, 3, 5 (?) and 0; the game of quechuncague or quechucan
is a race game ‘al modo de la oca’ (like the game of the goose) were pieces (tantos) are
moved according to the throw of the die. More interestingly we learn that every man
which encounters another man ‘eats’ it ‘al modo del ajedrez’. So kechukawe, as a board
game, seems to have been a race game with capture, a class of games not unknown in
other civilisations (e.g. the Arabo-Muslim tâb wa-dukk and its many relatives). Whether
the perched ring, known as chúgudhue according to Fébres’s Araucanian dictionary of
1765 (in Medina 1952), was used or not is unclear.
J.I. Molina (Molina 1787: II, x) explains that quechu, que [los Indios] aprecian
infinito, tiene una grande analogía con el juego de tablas, pero en lugar de dados se
sirven de triángulo de hueso señalado con puntos que echan por un arillo sostenido de
dos palillos, como era quizá el fritillo de los antiguos romanos.’ (see also Murray 1952:
147-8, no. 6.7.3.).
(‘quechu, which the Indians liked very much, is very similar to backgammon, but instead
of dice they use a wooden triangle marked with dots that they cast through a circle
Fig. 10: (facing page)
Two Araucanian games
in action (plate from
De Ovalle 1646).
perched on two sticks, as was perhaps the fritillus [dice cup] of the ancient Romans’)
Modern scholars have described a game called kechukawe, but it is a simple dice game.
Manquilef 1914 (§ 5. ‘El kechukawe’) reports that the die is a five-sided prism cast
through a ‘funnel’ in a circle on the ground. The score is kept with sticks (palitos). This
is roughly what De Ovalle’s picture ‘Modi ludendi Indorum’ (De Ovalle 1646) shows
(Fig. 10). In the early 19th century Luis de La Cruz (La Cruz 1835: 66) observed the
same game among the Peguenches under the name guaro [wayru!] played with a quechu
die, palitos and a perched ring.
There is a strong parallel between the Inca pichqa and the Mapuche kechu: not only
have they roughly the same shape (see Mátus Z. 1918-19: fig. 49 et 50 for two dice from
the Museo nacional de Chile [Fig. 12]; Cooper 1949 states that the ‘pyramidal’ die –
either with 5 or 7 sides – is common to all Andean cultures), not only both words mean
five’, but they were both used for two related games, a simple dice game and a race
game. So kechukawe is the exact equivalent of the Inca pichqa, and it is reasonable to
think the race game kechukawe is a likely cousin of the Inca huayrusitha/piscasitha.
komikan, comican
According to J.I. Molina (Molina 1787: II, x) the Araucanians knew ‘el artificioso juego
del ajedrez, al cual dan el nombre de comican(‘the ingenious game of chess, to which
they give the name comican). Komikan was still played in the early 20th century and was
described by Manquilef (Manquilef 1914: § 3, ‘Komikan’) and by Mátus (Mátus Z.
1918-19): it is a hunt game played on a latticed board with a triangle added to one side
(see Fig. 12). It is also called leoncito. There are 38 points (25 on the main board + 13
on the added triangle). One side has 12 men (Spanish perros dogs’ or perritos ‘little dogs’)
and the other has one bigger and more powerful piece called komikelu or leon. The
perritos move one step ahead; they try to hem the lion in. The leon alone has the power
to capture by leaping over a perrito. Multiple short leaps are possible. What exactly
happens on the triangle is not revealed by our sources. Mátus reports he had seen the
game played ‘among the Indians of inner Valdivia; but I could not clarify this subject
Fig. 11: Schematic
drawing of ‘Ludus
quechucague’ (from
De Ovalle 1646).
[the rules of the game] with them because they refused to give me details’ (Mátus Z.
1918-19: 169).
Because he had too little information, Murray 1952 classified comican (‘said to
resemble chess’, after Molina) with his ‘War-games of which we have no certain
knowledge’ (Murray 1952: 97). Instead the Mapuche game would rather belong to what
Murray called ‘tiger games’ (Murray 1952: 107-12). However, although komikan
somewhat resembles any hunt game played on the alquerque board, it has its own
features: no European game has any added triangle; and it is dubious that the Mapuche
game came from Malaysia or Indonesia! Even in these countries the rules and initial
position of the men are different.
The Mapuche game of komikan is no doubt the equivalent of the Inca komina also
known as taptana (Santo Tomás 1560: ‘alquerque’). Komikan must be a cognate to
Quechua komina and to Aymara kumi. The similarity between the games have already
been noted. Although komikan is not described before the late 18th century (it is not
mentioned by De Ovalle, Rosales, or Olivares), it is hard to suppose that it would be just
a slightly modified European import.
Fig. 12 (left): Two
Mapuche dice from
the Museo nacional
de Chile (from
Mátus Z. 1918-19:
fig. 49 et 50).
Fig. 13 (right):
Komikan board
(from Mátus Z.
opening positions.
llügün, lüqn, lünq, lüq, llique, lligues (Mapuche lüq, liq, white’ according to Vivante
1946: 33); modern Mapuche awarkuden ‘beans game’ (Vivante 1942; Vivante 1946)
Andean people also played ‘beans’ games, where half-blackened beans were used instead
of dice. This tradition can be traced back to the Mochicas (c. 100 BC-c. 600 AD); the
Mapuches used to call it llügün (or lligues), and call it now awarkuden.
The game needs 8 beans, peeled and blackened on one side, spotted with dots; the
beans are cast on a mat (pontro), and 40 ‘counters’ (kob, kou) – sticks, seeds or beans –
for keeping the score (20 for each player). The games are supported with incantations.
The score is won or lost according to the number of faces up: 4 black and 4 white = 1
point; all black or all white = 2 pts. (Manquilef 1918-19: § 6, ‘Awarkuden’).
De Ovalle 1646 describes, without naming it, a game of porotos o habas(beans):
they choose for that the white and they paint them black on one side (…); they drop
them on the ground through a suspended circle or a large ring; the player whose beans
fall with painted faces up wins the highest score.’ Moreover, the players blow themselves
on their breasts! (see Fig. 10 De Ovalle's picture ‘Modi ludendi indorum’ with the
perched ring). In the late 17th century Rosales (post 1674) described a game called uies
said to resemble dice. The player shout at the beans (in Pereira Salas 1947: 219).
According to Carvallo's Descripción … del reino de Chile, c. 1796 (in Medina 1952),
lligues are ‘12 halves of beans, the ones black, the others white’.
Armando Vivante suggested that the Auracanian awarkuden was the same game as
the Inca apaytalla and as the game represented on Mochica potteries (Vivante 1942,
1946; cf. Hocquenghem 1979 and Hocquenghem 1987). However, there are some
differences: palitosare never mentioned by the early sources. On the Mochica vases there
is no trace of any throwing ring, but an undulating sandy ground is depicted with beans
placed in hollows and ridges.
It is not easy to get a precise picture of the games the Incas played. However, from the scarce
and confusing sources I have presented it is possible to go further than the scholars who had
studied these games previously. One wonders why the Spanish chronicles and dictionaries
are so poor, compared to the good accounts we have about the Aztec patolli. Expected
sources are silent, and there is no useful iconography. After all, it seems that games and
gaming had little importance in the Inca world, at least less than in Mesoamerica where all
the chroniclers were impressed by the Indians’ addiction to gambling (and this was true too
for European travellers in North America, who observed the same phenomenon).
Contrary to this, the Incas seem to have had a large variety of games, but no specific
enthousiasm for one of them. Did they prefer ‘thinking’ games? This is what Father
Cobo inferred when remarking: ‘Although they were barbaric, these Indians invented
some ingenious games that correspond to our dice and to other games of ours; but they
used them more for entertainment than for the lure of gain.’
‘Aunque bárbaros, inventaron estos Indios algunos juegos ingeniosos que correspon-
den á el de los dados y á otros de los nuestros; pero usábanlos más por entretenimiento
que por codicia de la ganancia’ (Cobo (1653: XIV, ch. 17).
A. Early sources
Arriaga, Pablo Joseph de 1621. La extirpación de la idolatría en el Perú (1621). Lima,
Bertonio, Ludovico 1612. Vocabulario de la lengua aymara. Lima; reprint 1984.
Cobo, Bernabé 1653. Historia del Nuevo Mundo (1653). In: Obras del P. Bernabé
Cobo, F. Mateos, ed.,. Madrid, 1956 (Biblioteca de Autores españoles, 91-92, 2
Vol.). Vol. II, 270.
De Ovalle, Alonso 1646. Histórica relación del reyno de Chile. Roma; Italian ed.:
Historica relatione del regno de Cile. Roma, 1646.
González Holguín, Diego 1608. Vocabulario de la lengua general de todo el Perú, lla-
mado lengua qqichua o del Inca. Lima; reprint Lima, 1952.
Juan, Jorge & Ulloa, Antonio de 1748. Relación histórica del viage a la América meri-
dional. Madrid (4 Vol.).
La Vega, Garcilaso de 1609. Comentarios reales de las Indias (1609). Madrid, 1960-65
(4 Vol.).
Molina, J.I. 1787. Saggio sulla historia civile de Chili. Bologna; Spanish transl.:
Compendio de la historia geográfica, natural y civil del reino de Chile, 1788-1795.
Murúa, Martín de 1590. Historia del origen y genealogía real de los reyes incas del Perú
(‘short’ version, 1590), C. Bayle, ed. Madrid, 1946.
Murúa, Martín de 1613. Historia general del Perú: origen y descendencia de los Yncas
(‘long’ version, c. 1613), M. Ballesteros, ed. Madrid, 1962-64 (4 Vol.); later Madrid,
Santo Tomás, Domingo 1560. Lexicon o vocabulario de la lengua general del Perú.
Valladolid; reprint 1951.
Torres Rubio, Diego de 1619. Arte de la lengua quichua. Lima; revised ed.: Arte y voca-
bulario de la lengua quichua general de los Indios de el Peru. Lima, 1700.
Velasco, Juan de 1789. Historia del reino de Quito… (1789). Quito, 1978 (2 Vol.).
Vocabulario 1586. [Martínez, Juan?], Arte y vocabulario en la lengua general de Peru
llamada quichua. Lima; reprint Lima, 1951 as Vocabulario y phrasis de la lengua
general de los Indios del Peru llamada quichua.
Waman Puma (Guamán Poma) de Ayala, Felipe 1615. Nueva crónica y buen gobierno
(1615), J.V. Murra, R. Adorno, J.L. Urioste, eds. Mexico, 1980; later Madrid, 1987.
B. Modern sources
Alcina Franch, José 1980. Juegos y ritual funerario en Chinchero (Cuzco). In: III
Congreso peruano: El hombre y la cultura andina (31 de Enero-5 de Febrero 1977).
Actas y trabajos. Segunda serie, IV. Lima: 441-456.
Arellano-Hoffmann, Carmen 1994. Chuncana-Spiel bei den Chimu?. In: Tribus, 43:
Bell, Robert Charles 1979. Board and table-games from many civilizations. Revised ed.
New York.
Cavero, Luis E. 1955. Rito funerario: el pichqa. In: Archivos peruanos de Folklore, I,
no. 1: 154-156.
Cooper, John M. 1949. The Araucanians. In: Handbook of South-American Indians,
Vol. 5. Washington, D.C.: 503-524.
Culin, Stewart 1898. Chess and playing cards. Washington, D.C., 1898 (= U.S.
National Museum Annual Report for 1896, Vol. II: 665-942); reprint New York,
Culin, Stewart 1907. Games of the North American Indians. Washington, D.C., 1907
(24th Annual Report of the Bureau of American Ethnology, 1902-3); reprint New
York, 1975, 1986.
De Lucca D., Manuel 1983. Diccionario aymara-castellano, castellano-aymara. La Paz.
Figge, Horst 1987. Rechnen mit dem peruanischen Abakus – spielen mit dem Objekt
DB-02 MRI: Versuch einer Rekonstruktion von Regeln. In: Indiana, 11: 143-165.
Fonck, Francisco 1912. Formas especiales de los utensilios caseros de los aborígenes :
¿folklore ó no?. Santiago de Chile.
Hartmann, Roswith 1980. Juegos de velorio en la Sierra ecuatoriana. In: Indiana, 6:
Hartmann, Roswith, & Oberem, Udo 1968. Beiträge zum ‘Huairu-Spiel’. In: Zeitschrift
für Ethnologie, 93: 240-259.
Hocquenghem, Anne-Marie 1979. Le jeu et l’iconographie mochica. In: Baessler-Archiv,
n.s., XXVII: 325-346 (see also her Iconografía mochica, 3rd ed. Lima, 1989: 8b. Las
escenas de ‘juego’).
Holm, Olaf 1958. Taptana, o el ajedrez de Atahualpa: a los 425 años de Cajamarca. In:
Cuadernos de Historia y Arqueología, VIII, nos. 22-23-24: 91-109.
Karsten, Rafael 1930. Ceremonial games of the South-American Indians. In: Societas
Scientarum Fennica. Commentationes Humanarum Litterarum, III, 2: 1-38.
La Cruz, Luis de 1835. Descripción de la naturaleza de los terrenos que se comprenden
en los Andes poseidos por los Peguenches (…). Buenos Aires.
Lara, Jesús 1978. Diccionario qhëshwa-castellano, castellano-qhëshwa, 2nd ed. La Paz.
Manquilef, Manuel 1914. Comentarios del pueblo araucano. II. La jimnasia nacional
(juegos, ejercicios y bailes). In: Revista de Folklore chileno, IV, 3-5: 75-219.
Mátus Z., Leotardo 1918-19. Juegos i ejercicios de los antiguos araucanos. In: Boletín
del Museo nacional de Chile, XI: 162-197.
Medina, José Toribio 1952. Los aborígenes de Chile. Santiago: 311-312.
Murray, Harold R.J. 1952. A history of board games other than chess. Oxford; reprint
New York, 1978.
Nordenskiöld, Erland 1910. Spiele und Spielsachen im Gran Chaco und in
Nordamerika. In: Zeitschrift für Ethnologie, 42: 427-433.
Nordenskiöld, Erland 1918. Spieltische aus Peru und Ecuador. In: Zeitschrift für
Ethnologie, 50: 166-171.
Pereira Salas, Eugenio 1947. Juegos y alegrías coloniales en Chile. Santiago: 218-219.
Pratesi, Franco 1994. Giochi andini e protoscacchi. In: Scacchi e Scienze Applicate, Vol.
12 (fasc. 14): 15-19.
Radicati di Primeglio, Carlos 1979. El sistema contable de los Incas: yupana y quipu.
Rivet, Paul 1925. Coutumes funéraires des Indiens de l’Equateur. In: Actes du Congrès
international d'histoire des religions tenu à Paris en octobre 1923, Vol. I. Paris: 376-
412; Spanish transl. In: Boletín de la Biblioteca Nacional de Quito, n.s., II/8, 1927.
Roca Wallparimachi, Demetrío 1955. Ceremonias de velorios funebres. In: Archivos
peruanos de Folklore, I, no. 1: 138-153.
Romero de Valle, Emilia 1943. Juegos del antiguo Perú: contribución a una historia del
juego en el Perú. Mexico.
Rowe, John H. 1946. Inca culture at the time of the Spanish conquest. In: Handbook
of South-American Indians, Vol. 2. Washington, D.C.: 288-289.
Smith, John 1977. Recuay gaming boards: a preliminary study. In: Indiana, 4: 111-138.
Vivante, Armando 1942. El juego mochica con pallares. In: Revista Geográfica
Americana, IX, Vol. xviii, no. 110: 275-282.
Vivante, Armando 1946. Un antiguo juego peruano. In: Revista Geográfica Americana,
XIII, Vol. xxvi, no. 154: 27-33.
On a Phylogenetic Classification of Mancala Games,
with some Newly Recorded Games from the “Southern
Silk Road”, Yunnan Province, China / Vernon A. Eagle
Mancala is a family of games of calculation, played widely in Africa and Asia,
whose details differ profoundly from one venue to another, but whose
distinctive features point persuasively to a common origin. Their antiquity
appears to be on the order of several millenia. Their present diversity, we may conclude,
is the product of a complex evolution whose reconstruction, interesting in its own
right, would also help illuminate the (largely unknown) history of cultural contact and
human displacement which has taken place in the Asian and African continents outside
the boundaries of the written record.
To r econstruct that history in the absence of historical evidence older than a few
centuries would appear at first blush to be hopeless, but in fact it is in the richness of
the present material that hope may be found. For the hundreds of mancala games
described to date, and the, perhaps, thousands of games in existance are not simply
diverse. They are diverse in a certain way: their diversity is the product of their actual
evolution, and in the organization of that diversity we may seek the reconstruction of
their history.
Mancala games are played on boards, which may be carved of wood or scooped out
of the ground, and which consist of a number of holes, usually arranged in rows, most
often two or four. The playing pieces are simple counters, commonly pebbles or seeds,
which are usually completely undifferentiated. Play consists of distributing and
redistributing the counters in the holes, typically by lifting the contents of a hole and,
beginning with a neighboring hole, dropping these counters one by one in successive
holes along a row, then back down the holes of the neighboring row in the other
direction, and so cyclically around the board. This operation is known as “sowing”.
Depending on where the final counter drops, and the configuration this produces, the
player may sow again, or remove counters from the board, or his turn may end.(1)
Why Classify Mancala Games?
Classification arises in human thought in several ways. Often, a classification is
imposed upon a set of entities for convenience only. For example, a library may be
arranged by the size of its books, for economical use of shelf-space. The same books
might be arranged alphabetically by author, or grouped by the languages in which they
are written, or by subject-matter. Or, a classification may be proposed, or imposed, to
reflect a certain point of view as to what differences are most important to the analyst,
or to the task at hand. For example, a field guide to flowering plants might be arranged
by the color of the blooms. The task may be mnemonic: thus, the night sky, organized
into constellations, becomes easier to keep in mind. Such classifications, while not
necessarily arbitrary, are subjective, and cannot be shown to be “right” or “wrong”.
They are useful, or not; appealing, or not; illuminating or confusing as the case may be.
A special situation arises, however, when the entities to be classified have come into
being by a process of evolution from a common stock. Then, the historical task of
attempting to describe the actual evolution in question, the sum total of all the
speciation” events by which new entities are generated out of old ones, gives rise to the
attempt to construct a hierarchical classification faithfully reflecting that evolution.
Such a classification is said to be “phylogenetic,” and it has the property that its cate-
gories, its “taxa” at every level, consist of all the descendents of a single ancestor.(2) Such
taxa are said to be “monophyletic.” Begging the question of whether the entities them-
selves are well-defined, a phylogenetic classification is objective, not subjective. It
carries with it the implicit prediction that characters yet to be examined will be found
to be distributed in accordance with its groupings. It can be refuted by evidence, and
is hence, in principle, scientific.
The paradigm for phylogenetic classification is, of course, the evolution of biological
species. Especially in the 1960’s and later, rigorous attention to principles of phylogenetic
classification has greatly changed, and strengthened, the discipline of biological
taxonomy. Human languages, analogous to biological species, present a similar task. We
take the view here that board games, and mancala games in particular, also present a case
of present diversity resulting from a process of evolution from a common ancestor, and
hold out the hope of constructing a phylogenetic taxonomy reflecting the actual course
of history. This taxonomy should complement, not mirror, a similar classification of
languages. On the one hand, a game can spread by diffusion from one culture to another,
crossing a language boundary. On the other, the subjugation of one people by another
may extinguish a language, leaving a game to survive.
Mancala Games Have a Common Origin
The point here is that similar and often quite complicated modes of play exist in far-
distant parts of the world ... [which] cannot conceivably be of independent invention
and parallel development.” (Townshend 1977b).
The hundreds, perhaps thousands, of mancala games played in the African and Asian
continents, although differing widely in their rules of play: capture methods, initial con-
figuration, method of relay, have also striking fundamental similarities which argue per-
suasively for a common origin for the entire group in time and space. That is, it is reason-
able to suppose that they are all the descendants of a single ancestral game. First, game
equipment consists of a board of two rows of holes: this is the generic shape throughout
the area of mancala play (other configurations can all be regarded as derivative), and
across completely different capture-method types; together with a set of identical playing
pieces. Secondly, play consists of sowing, that is, lifting the contents of a hole and distrib-
uting them one at a time, starting with a neighboring hole and proceeding consecutive-
ly and cyclically around the board. Moreover, typically sowing is compound, that is, a
typical move consists of a sequence of sowings, the placing of the final piece of each
sowing determining the following one. Moreover, it is typical of all these games (again
with exceptional cases) that the player’s free choice is exercised only at the start of a move,
and consists in choosing from which hole, and perhaps in which direction, to play: the
result is then determined by the rules of sowing and capture of the particular game.
Mancala games are quite singular among board games in that the playing-pieces used
by the opponents are undifferenciated. Mancala play, excepting relatively recent spread
from the Old World to the New, occupies a vast, but essentially contiguous zone on the
Afro-Asian land mass. The simplest explanation for the existence of the multiplicity of
mancala games all exhibiting the same (actual or derivative) complex nature is a
common origin, and in the absence of strong contrary evidence, the simplest explanation
is to be preferred.
The Relationship between Mancala and Other Board games
The nature of a board game as an intellectual contest between two opposing players is
reflected in the usual condition of the games not of mancala type, that is, that the
playing pieces are divided into two camps, one for each player. This is noted by Deledicq
and Popova (Deledicq & Popova 1977, p. 21), whose apparantly inelegant term “Anti-
Mancala” for the collection consisting of almost all other board games (race-games, war-
games, position-games, hunt-games and so forth) is actually quite apt. Mancala games,
I suspect, are absolutely unrelated to these other “two-camps-of-playing-pieces” games.
It is misleading even to refer to mancala counters as “pieces,” in that this suggests
they are homologous with, eg, pieces in games like draughts or backgammon or weiqi,
which are placed or moved, on game-boards. Mancala counters have a dual role.
Consider, for illustration, a board game like “Monopoly.” The role of the counters in a
mancala game is not only like that of the moving pieces in Monopoly, which the players
move around the board. In fact, mancala counters are also like the Monopoly money,
which the players compete to obtain. In this sense, mancala games are comparable to
card games, in which the equipment, the deck of cards, is neutral, to be used by both,
more generally by all, the players. The deck of cards, indeed, is not a priori dedicated to
a contest. It may be used for a solitaire, and also, indeed, for divination.
Mancala games may have arisen, not by evolution from earlier games, but from a
divinatory method.
Divination generally requires an element of randomness, or unpredictability: which
card will be dealt from the shuffled deck? which way will the crack develop in the heated
carapace of the turtle? what pattern will the tea leaves form? (cf Townshend 1977c, p.
95). The mancala board and counters provide just such unpredictability when used to
perform the act which is the quintessence of mancala play: compound sowing. An
extended sowing in a mancala game is reminiscent of a spinning roulette wheel or wheel-
of-fortune: “round and round she goes, where she stops nobody knows”. Moreover, the
result of an extended sowing is not only the identity of the last hole sown in, but also
the configuration of the entire board as the pieces have been redistributed in the holes.
I suggest that the closest relatives of the ancestral mancala game were, not other
games, but other activities, perhaps divination, performed with the same equipment. My
hypothesis is that not only the board, but the activity of compound sowing, existed
before the game.
On Phylogenetic Classification of Mancala
Whether or not mancala games are genetically related to any other games and whether
or not they arose from divination, if in fact they do have a common origin, then the
possibility of a phylogenetic classification arises. Games, unlike living species, have no
physical genomes, and are far less complex than living organisms. Too, while hybrids
exist in biology, they may be far more common in the evolution of games, so that the
resulting branching diagrams illustrating the history of their speciation may be rich in
cycles. Recognizing that the analogy between biological species and board games is not
perfect, still the core insights of the “cladistic revolution” in biological taxonomy are fully
relevent to the task of classifying them, as these insights are based, not on the physical
mechanism of evolution, but on the fact (or hypothesis, or assumption) of evolution
having actually taken place.
The properties, or attributes, of the entities to be classified used in their classification
are known as “characters” – for example, in classifying flowering plants, the position of
the ovary, or the number of stamens; or, in languages, the presence or absense of tones;
or, in mancala games, the number of rows of holes constituting the board. As evolution
takes place, characters change. An ancestral fin becomes a leg, then a wing. When a
character changes its state, the earlier condition is called “ancestral”, the later condition
derived.” A central insight of the cladistic method is to note that in adducing evidence
for relative closeness of genetic relationship, one must seek for shared derived character
states, and completely disregard shared ancestral character states. But this must be done
with care. Birds and bats both have wings, but together they do not constitute a
monophyletic group: wings arose more than once in the history of vertebrate anatomy.
Moreover, the fin that became a leg became a fin again among the cetacians (whales and
their relatives). So the piscine fin is ancestral to the mammalian leg which in turn is
ancestral to the cetacian fin. Whether a given character state is ancestral or derived
depends upon the context, that is, upon what taxonomic level is being considered.
The model is this: a monophyletic group consists of all the descendents of a putative
common ancestor. This ancestor is described by character-states, all, in this context,
ancestral. Over time, characters change state: new forms, descended from the ancestor,
come into being. Assuming a given derived character-state arose only once in the group,
and at the level, under consideration (“uniquely-derived”), then sharing it implies
common descent not only from the ancestor of the whole group, but from the earliest
of its descendants in which the new character-state is to be found. Sharing an ancestral
character-state, on the other hand, is evidence only for descent from the ancestor of the
whole group, a tautology.
Character states, then, must be assigned a polarity, an orientation in time, if they are
to be useful in classification. The principal method for accomplishing this, in the
absence of fossil evidence, is “outgroup comparison”. The idea is: to help in deciding, for
entities in a given group, which state of a given character is ancestral, consider entities
outside the group, but as closely related to it as possible. If these possess the character in
a consistent state, then that state is likely to be ancestral for the group under study. (Here
“likely” means more precisely that this explanation is preferred over others because it
requires fewer additional hypotheses about unknown ancestors.)
We accept as taxa, then, only monophyletic groups, defined on the basis of shared
uniquely-derived characters. Groups unacceptable as taxa, like birds + bats, which are the
result of convergent evolution, are called “polyphyletic.” This means they are composed
of two or more monophyletic groups lumped together. The use of shared ancestral
characters in classification generates a second type of unacceptable group, known as
paraphyletic”. This, in effect, is a group created by subtracting one monophyletic group
from another. For example “the great apes,” excluding man, or “monkeys,” excluding
apes, or even “reptiles,” excluding birds, are all to be rejected as paraphyletic. To take the
second example, the possession of an external, visible tail is ancestral among primates;
its loss among apes, including man, is derived.
We will find that both polyphyletic and paraphyletic groups have been proposed,
and must be rejected, in classifying mancala games.
Classifications of Mancala in the Literature
1. Murray (1952) divided mancala games into three groups, distinguished on the
number of rows making up the board. Thus, he had Mancala II, Mancala III, and
Mancala IV. Next, he sorted the Mancala II games by geographic region, and in one of
these regions (“West Africa: Guinea from the Senegal to the Gabon and the Sudan”)
classified the games into nine groups (including a “none of the above” miscellaneous
group), organized by a key (Murray 1952, pp. 178-179). He uses, first, sowing rules,
and then, capture methods, for organizing these groups. He didn’t really know the
Mancala III games. Mancala IV he divided into two types: (a) and (b) (Murray 1952,
p. 207). The type-(a) games are characterized by captured counters being taken out of
play. The type-(b) games are characterized by captured counters being sown back into
the game by the player who has captured them. Murray then further divided the
Mancala IV-(a) games into five groups, based on differences in the rules for capture. He
organized the IV-(b) games according to the number of “reverse holes” – holes from
which a player may reverse the usual sense of play and sow clockwise in order to
2. Deledicq and Popova (1977) divided mancala games into two groups. The first
group, “wari,” consisting of most 2-row games and all 3-row games, is characterized by
the players’ sowing in the holes of the entire board (exceptional holes allowed), while the
second, “solo,” consisting of all 4-row games known at that time and a few exceptional
2-row games, is characterized by the division of the board into two halves, each player
sowing in his own, and capturing from his opponent’s half. They appear to be unfamiliar
with the diversity of solo games, but give a typology for wari, (Deledicq & Popova 1977,
pp. 102-105), according to the states of four characters: (s, a, p, m)
i. sowing is simple or compound: s = 0, 1 respectively.
ii. accumulation holes (sinks) are absent entirely, appear during the course of play,
or exist a priori from the start of the game: a = 0, 1, 2 respectively.
iii. captures are from the final hole of a sowing, or from another hole or holes deter-
mined by it, or take place during the sowing: p = 1, 2, 3 respectively.
iv. play is in a single round, or in multiple rounds: m = 1, 2 respectively.
There result from this 2 x 3 x 3 x 2 = 36 possibilities, of which 15 are realized by games
known to the authors. The authors hazard various hypotheses deduced from the
purported nonexistence of the remaining 17.
3. Townshend (1977a, 1977c, 1979, 1986) agrees with Deledicq and Popova in
dividing mancala into wari and solo, and follows Murray in dividing solo into two types,
inexplicably reversing Murrays nomenclature, so that Townshend’s type A is Murray’s
type (b) and Townshend’s B is Murray’s (a). As Murray clearly has priority, when letters
“a” and “b” are used here, it will be in Murray’s sense. Townshend proceeds to provide a
far richer and more useful typology than any previous writer. He divides (Townshend
1979) wari games into five types, (designated a, b, c, d, e) distinguished by the method
of capture. He divides Solo-(a) games into four groups, again on the basis of capture
method only, and describes two “intermediate types” of Solo games with mixed
characteristics of (a) and (b) type. He first (1977a) divides solo-(b) games into five types:
sombi, mangola, cisolo, kibuguza, and Swahili bao. Later (Townshend 1986), he describes
bao as belonging to the sombi group.
4. Russ (1984) provides a survey of mancala games, and the organization of his book
is not perhaps intended as a formal classification. He retains Murray’s categories of two-
row, three-row and four-row games. Some of his chapters correspond roughly with
Townshend’s typology of two-row wari games, and he groups together two-row games
lacking compound sowing.
5. Santos Silva (1995) gives keys to typologies for solo-(a), solo-(b), and “wari”,
(which he calls Mancala IV-B, Mancala IV-A, and Mancala II, respectively). He
apparantly is not familiar with the work of Pankhurst or Townshend. He gives a key to
7 types classifying 37 solo-(a) games (Santos Silva 1995, pp. 125-131); a second key to
17 types classifying 28 solo-(b) games (pp. 145, 150-153), and a third key to 44 types
classifying 144 “wari” games (pp. 194-217).
Let us now examine some of the categories suggested by the referenced authors.
1. Wari. Unless one is prepared to argue that the original mancala game was a 4-row
game played on a double-board, and that 2-row games arose by a simplification of the
board, it is clear that all that is meant by “wari” is “mancala which is not solo.” In other
words, “wari” is a paraphyletic group, and we reject it as a taxon. When we write “wari,”
we mean “mancala games in which both, or all, players, play on the whole board
(exceptional holes allowed). It is a useful word, but not acceptable as a taxon.
2. Mancala III. As we have noted above, Deledicq and Popova reject this grouping
on the grounds that it is essencially just a variant of 2-row mancala, and that it is closer
to the 2-row than to the 4-row games because it is, like the former, composed of “games
of one cycle” in which the players all play over the entire board. But in fact, if we accept
the proposition that 4-row games are also derived from 2-row games (by doubling the
board), there is no reason a priori to hold that the dichotomy between 4-row games and
all others is any more profound (or more ancient) than that between 3-row games and
all others. The real difference in the situation of 3-row as opposed to 4-row games
becomes clear when one examines them against Townshend’s typology of “wari.” For his
capture-method analysis applies equally to 4-row games, and it is seen that all 4-row
games, of both types-(a) and -(b), employ type-(d) capture, that is from holes opposite
the final hole of a sowing. On the other hand, 3-row games can be found employing
several capture methods also found in 2-row games. The conclusion is that while this is
consistent with the monophyly of 4-row African games played on double-boards
(together with the few African 2-row games also played on double-boards, which also
employ type-d capture), it implies that Mancala III is polyphyletic.
Thus we accept the category solo defined by Deledicq and Popova, and reject
Murray’s Mancala III. However, it is clear that there must have been a first 3-row game,
and, furthermore, the geographical location of all extant 3-row games in the horn of
Africa suggests that all of these games do have a common origin. What complicates the
taxonomy is that the likeliest explanation for the multiplicity of capture-method is that
some of these games are hybrids. Further analysis may resolve this, perhaps enabling us
to construct a monophyletic group of “true 3-row games,” as distinct from essentially
2-row games played on a 3-row board. For example, abalala’ (Courlander 1943,
Pankhurst 1971, p. 163), a type-(d) game (in the sense of Townshend) participates in
the the geometry of a 3-row board, in that capture may be from one or from two holes
3. Townshend’s type-(d). These are games in which captures are made from holes
on the opponent’s side of the board directly opposite the hole receiving the final
counter of a sowing. Townshend’s type-(d) games as he defined them, that is as a group
of two-row and three-row games, is paraphyletic, as it excludes the four-row games
which developed from them. But if we put the four-row double-board games in, then
the group would seem valid. Solo is then a sub-group. Townshend (1979) also defines
type-(d)-ii as the sub-group of games in which the final counter, landing in an empty
hole, is captured together with the counters of the hole opposite. This also seems valid.
His type-(d)-i, on the other hand, consists of type-(d) games with no other special
characteristics, and is paraphyletic.
For convenience, and because the concept is seminal, we will describe capture of this
sort, namely from holes opposite the hole on a player’s own side receiving a final counter
as “Townshendian capture.” Games employing such capture as the principal method will
be called “Townshendian mancala games.” The subgroup of games employing what
Townshend designated as “type-(d)-ii captures” as described above will be called “Gogo,”
after the Mijikenda game kigogo, which is in this group, and such captures will be call
gogolian.” Then both Solo and Gogo are seen as monophyletic groups of
Townshendian mancala games.
4. Solo. Examining the division of Solo into types (a) and (b) according to whether
captures are removed from the board or sown back in, respectively, out-group
comparison with other games shows clearly that (a) is the ancestral condition, and (b)
is derived. This is greatly reinforced by looking at capture methods. As noted above,
all solo games employ Townshendian capture, so the candidate out-group consists of
wari” games employing this method. In general, Townshendian wari games capture
opposite an empty hole on the player’s side receiving the final counter of a sowing.
This is also generally true for solo-type-(a) games, while for solo-type-(b) games capture
is generally opposite a non-empty hole on the player’s side receiving the last counter
of a sowing. Thus we conclude that solo-(a) (which Townshend designates as “Cela
(Townshend 1977a, p. 50) is paraphyletic, and solo-(b) is monophyletic. Borrowing
Townshend’s terminology while perhaps extending its purview, we will use Sombi to
mean “Solo games with captures opposite a final, non-empty hole sown back into the
board.” Thus Sombi will include, not exclude, Townshend’s mangola, kibuguza, cisolo,
and bao.
5. “Intermediate Types”. Townshend (1979, p. 119) reports the existence of games
(he calls them “Intermediate Type D”) in Western Kenya (Nandi kecuek and Kipsigis
ndoto, both on 4 x 6 boards) in which capture is opposite an empty hole receiving final
counter, but where captures are sown back into the board. He does not give complete
descriptions, but it seems likely that the immediate ancester of Sombi would have been
just such a game, retaining the ancestral state of the Townshendian capture method.(3) It
is consistent with this analysis that these “Type D” games have captured counters sown
in starting at the postultimate hole of the sowing which captures them. Sowing in
captures is strongly reminiscent of compound, or relay sowing. If it is indeed essencially
a kind of generalized relay, then its original form might well have been to “relay” the
captured counters as if they had actually occupied the empty, final hole opposite them.
Townshend’s subgroups “Type A-I-(iv), -(v), -(vi) and -(vii)” all employ sowing in of
captures forward of the point of capture, which may be the ancestral state. Let us use
Kecuek to denote “Solo games with captures sown back into the board.”
6. Typology of Solo. Several of Murray’s types of solo-(a) games are based on what
Townshend calls “bonus captures.” That is, a player who captures in the usual way adds
to his winnings the contents of one or more holes of his choosing on his opponent’s side
of the board. As this does not appear outside of Solo, it appears to be possible to use this
character to define a monophyletic group. We shall designate by Nchombwa the group
of solo games employing bonus captures. The name is based on a game described from
Malawi by Sanderson in 1913.
7. Typology of Sombi. Consider first the “reverse-holes” of Murrays typology. For
outgroup comparison, we look at solo games outside Sombi. Most have strictly counter-
clockwise sowing, some sow clockwise, and some allow either sense, but none allow
sense-reversal only to capture or only from certain holes. It thus appears that “Sombi
games with reverse-holes” may constitute a monophyletic group. We shall designate the
group by Alok, which is, according to Driberg (1927), the term employed for the
procedure by the Acholi of Uganda.
Townshend (1977a) defines the group Mangola to designate Sombi games in which
the final counter of a sowing skips over an empty hole, to sleep in the following hole.
This is a distinctive property, not found in any other games outside the Sombi group.
Townshend reports one Mangola game with reverse-holes: the Alur game of Leka. This
game is perhaps a hybrid between Mangola and Alok.
Townshend defines Cisolo to designate Sombi games in which compound sowing is
performed as in pussa-kanawa games(4), by relay from the hole following the final hole of
the previous sowing, and having a distinctive capture method, first described by Driberg
for the game Choro as played by the Lango in Uganda. If a player, whose turn it is to play,
should have an occupied inner-row hole opposite an occupied inner-row hole of his
opponent, then he immediately captures the contents of the latter, together with the
contents of the opponent’s outer-row hole in the same column, if any, and sows them in,
starting in his own inner-row hole of this column. If such a capture is possible, it is
obligatory. We call such captures “Langolian.” The Lango game also has standard Sombi
captures, as well as reverse-holes. The relay-method of Cisolo influences its capture-
method, as captures are from holes opposite an occupied inner-row hole following the
final counter of a sowing, that is, from the hole from which, if capture is not possible, a
relay would begin. If we believe it likely, as Townshend does (1977a, p. 47) that Langolian
captures arose once only, then we may define a monophyletic group Langola, for those
Sombi games in which it occurs. Cisolo is then a subgroup. The fact that the Lango game
has reverse holes may be explained in several ways: 1. Cisolo games have lost reverse-holes
which their ancestors posessed; 2. The Lango game Choro is a hybrid between Ugandan
Alok games and an ancestral Langolian game without reverse-holes; 3. Alok is
polyphyletic, reverse-holes having arisen independently; or 4. Langola is polyphyletic.
The second possibility seems to me the most likely, subject to further evidence.
Langolian capture is reminiscent of Swahili bao, in that, during the first phase of
that game, opposing occupied inner-row holes occasion an immediate capture.
Moreover, in Langolian capture and in Swahili bao, unusual in Solo generally, capture
is obligatory.
Townshend defines the group Kibuguza to include two games having an unusually
generous capture-rule: any final counter landing in the interior row captures from the
two opposite holes.
We summarize the discussion above with the following table (facing page).
A Word on Methodology
It should be emphasized that the remarks above are intended only as a rough
commentary on the groupings of mancala games which have appeared in the literature.
They do not constitute a formal classification. This can only be accomplished “from the
bottom up,” rather than “from the top down.” That is, for each game studied, we pose
the question: what is the closest relative? or, failing that: what games described to date
are most closely related to the game in question in sharing with it certain uniquely-
derived characters. Then, of such a group, we repeat the question. Thus the higher taxa
are constructed out of the lower, rather than being defined by certain properties, like
Platonic ideals. Thus, having proposed Kekuek, above, or Gogo, as valid taxa does not
make them such. Many more games will need description, and many more characters
will be required, if we are to achieve much confidence in our understanding of the
relationships between the mancala games already reported.
Games with capture of holes [Townshend type -(a) captures. e.g. “typical”
East African Maasai enkeshui]
Games with capture of n-tuples [Townshend type-(b) captures. e.g. “typical”
West African wari]
Townshendian games
Solo [four-row double-board]
Nchombwa [bonus captures]
Kecuek [captures sown back in]
Sombi [capture opposite occupied hole]
Alok [reverse holes]
Langola [Langolian captures]
Cisolo [relays from postultimate hole]
Mangola [last counter skips empty hole]
Kubuguza [final inner-row counter captures all opposite]
Gogo [final counter captured together with counters opposite]
Pussa-kanawa games [empty, eat. Townshend type (e) captures]
Polyphyletic groups:
Games with only simple sowing (no relays)
Games with postultimate relays (= Pussa-kanawa + Cisolo)
Mancala III
Paraphyletic groups:
Wari (in the sense of Deledicq & Popova 1977)
Townshend’s two-row type-(d)-i
Mancala-IV Type-(a) ( = Type B of Townshend, etc)
Sombi exclusive of Mangola, etc
[all the grab-bag groupings indicated above by “others”]
Some Newly Described Games from China
In a continuing effort to advertise the richness of Asian mancala play, I report here four
games from South-West Yunnan Province, China. All the venues of play are on the
“Southern Silk Road,” an early trade route between China and India, across Burma.
1. A game with no sectors from Tengchong County.
Principal Informants: Zhang Jingyao, male, 16, and his mother Yang Xiuying, 46,
both Han nationality.
Venue: Yunnan Province, Baoshan Prefecture, Tengchong County, Hehua Township,
Xiaozhuang Village. This is on the main road from Tengchong to Lianghe.
Date of interviews: April 1996.
Name of the game: Laomuzhuqi. Qi means “board game”, as in xiangqi (Chinese
chess) or weiqi (called Go in Japanese). Laomuzhu means “the old mother pig”, and
refers to the large stones used in play.
Format: 2 x 5. The board is normally drawn on the ground with chalk or charcoal:
a rectangle divided down the middle and into five compartments on each side, two
compartments at one end marked with big X’s to indicate that the two large stones,
informally called laodao, are placed there at the start of each round. To begin, each
player has five small stones in each of his other four compartments.
Preliminaries: To decide who plays first, the players simultaneously throw out one
of three fingers (huaquan in Chinese): Thumb, called taishan (the mountain); pin-
kie, called xiaogongji (the little cock); or index finger, called mayi, the ant. Then, as
in “paper, rock, scissors” there is a cyclical order determining the outcome. To wit:
The mountain crushes the chicken, the chicken grabs and eats the ant, and the ant
knocks down the mountain (in Chinese: Taishan ya ji; ji ba mayi zhuachi, mayi gong-
dao taishan) This is done before each round.
The Play:
1. On his first move in each round, each player must play from a hole on his own side
of the board. Subsequently, there are no sectors, i.e. both players may play from any
regular hole, on either side of the board (“regular” meaning that the hole is not a
tian- see 4. below).
2. Play is in either direction, but the laodao must be sown first in any sowing contai-
ning it.
3. Relays and captures are standard Pussa-kanawa type. That is, relays are from the hole
following the final hole of a sowing. If this is empty, the contents of the next hole
Map of
are captured; if the hole following that is empty too, then the contents of the next
hole are also captured, and so on. When a stable situation is reached at the end of a
round, the stones are divided evenly between the players. If the number is odd, the
remaining one is awarded by throwing fingers, as in the preliminaries.
4. At the end of a round, if one player has both laodao, he sells one of them to his
opponent for 5 stones. Then both players fill their holes if they can. The filled holes,
and a partially filled hole, belong to the player who fills them, but any holes left
empty become the property of the opponent. They are his tian”, or productive
Subsequent rounds.
5. Normally, the owner of the tian keeps a singleton in each, any surplus being remo-
ved by him. They are entirely disregarded in reckoning relay and capture. In parti-
cular, they are neither played from, captured from nor relayed from, although they
may be sown into by either player.
6. When either player, in the course of play, approaches a series of 1 or more tian, he
has the option either to skip over all of them, or to sow in all of them. If his move
circles the board one or more times, he has the freedom to decide this option sepa-
rately at each circuit.
7. Except that the trailing player, if he has not sufficient stones to sow in all the tian, is
obliged to skip over them, while the leading player may in that case sow into them.
However, if he does so, his move ends, and he may not capture even if the next hole
beyond the tian is empty.
8. Under no circumstances may a laodao be sown into a tian. A laodao lying in a hole
before a tian and being played or relayed in the direction of the tian must skip over
the tian.
Affinities: The game’s use of captured holes shows some similarity with the game mak-
huhai, played by people of Dai nationality in Dehong Prefecture, Luxi County, Fengping
Township (my own notes, recorded 1995, not published). In the latter game, the trailing
player is forbidden to sow into the captured holes (called hem in the Dai language), while
the leading player is free to sow or not to sow in them, and to relay or not to relay from
them. The lack of sectors appears in laomuzhu from Longling county, Longxin Mengmao
(Eagle, 1995, p. 56).
2. A game from Lianghe County with a new method of hole capture.
Principal Informant: Zhao Jiakang, male, 25, Achang nationality.
Venue: Dehong Dai-Jingpo Autonomous Prefecture, Lianghe County, Jiubao Township,
Henglu Village.
Date of Interviews: April 1996
Name of the game: Dong Wo (in Chinese, dong is to move, wo is a hole)
Format: 2 x 5. At one end of the board, a large stone in each hole; at the start of the
game, five small stones in each of the other holes. The large stone is the laomuzi
(“old mother”) or simply muzi (“mother” - pig is understood)
Preliminaries: none. the players agree who plays first; on subsequent rounds, they
The Play: (in the following, “regular hole,” or simply “hole,” is distinct from “captured
hole” as discussed below.)
1. Play is in either direction, with standard Pussa-kanawa relays and captures. A player
plays only from his own side, except as in 14. below.
2. It is not allowed to sow one of the muzi into a hole containing the other one, but
otherwise one is free to sow them in any hole of a sowing. If a muzi is singleton, and
the neighboring hole contains the other muzi, the singleton may not be played in
that direction. Such a singleton may, however, in the course of play, be relayed into
a neighboring hole containing the other one, even deliberately. In this case, one of
the muzi is captured by the player who is moving.
3. If at the end of a move all a player’s regular holes are empty, and his opponent has at
least two stones remaining in his holes (muzi counting as 5), he must place a single-
ton in each of his holes. He must do this whether it is he or his opponent who is
about to move.
4. The round ends if, at the end of a move:
i. All holes are empty, or
ii. One player has a single stone and the other has no stones, or
iii. Each player is reduced to a single stone, and neither player is in a position to cap-
ture the other’s stone.
Winning a hole:
5. In cases i. and iii. above, the round is ping (even); but in case ii, the player having
the last stone in one of his holes is the winner of the round, even though he may
have captured fewer stones, and he is awarded a hole on his opponent’s side. He
chooses the hole, except he cannot take the end-hole containing the laomuzi. He
puts the surviving singleton in it.
6. At the end of a round, after sale of a laomuzi (worth 5) if necessary, the players fill
their holes with their winnings (except for holes which have been won (or purchased
– see below) by an opponent). If necessary, a player borrows from his opponent’s sur-
plus in order to fill his holes.
Buying a Hole:
7. A hole can be bought for 10 stones. In order to buy a hole on his opponent’s side, a
player must have not only the purchase price of 10 (taking into account any accu-
mulated debt), but at least one stone in addition to put in the bought hole. The pur-
chaser chooses which hole to buy (except the hole at the end where the laomuzi are
placed). If his opponent owns holes, won or bought, on the player’s own side, these
must first be bought back before a hole can be purchased on the opponent’s side. To
buy back a hole, a player must have not only the purchase price of 10, but an addi-
tional 5 stones to fill the hole. A player may not refuse to sell, if his opponent has
the wherewithal to buy.
8. NB: the restriction on buying holes on his opponent’s side while his opponent owns
holes on his own side does not extend to winning a hole. That is if a player wins the
round he takes a hole on his opponent’s side without regard to the status of holes on
his own side.
9. The players may, by mutual agreement, exchange holes they own on one another’s
side. Moreover, a hole which has just been won may be immediately bought back by
the loser of the round, if he has the 15 stones necessary for the transaction. After a
hole has been bought, or bought back, the seller may use the proceeds to buy back,
or buy on his own account. Thus a series of purchases might take place between
Captured holes:
10. Holes which are won or bought function exactly the same in play. We shall call them
captured holes. Normally, the owner of a captured hole keeps a singleton in it.
11.Whenever a stone is sown into such a hole, by either player, the owner may remove
it, and normally does. The owner may forget to take such a second stone. But if a
third stone is sown in, the hole-owner’s opponent (the player on whose side the hole
lies) is entitled to remove two stones, leaving the hole singleton.
12. The owner of the captured hole may, at any time during play, and no matter who is
moving, remove the singleton so that the hole is empty, thus causing or preventing
a capture or a relay. If the hole becomes empty through capture, relay, or the owner’s
having removed a singleton, the owner may at any time put a single stone in his hole,
thus causing or preventing a capture or relay. But he must do this quickly if the other
player is moving, as the opponent is not obliged to wait for his decision to remove,
or add, a singleton.
13. The owner of a captured hole may not sow a muzi into it, unless it is unavoidable as
a relay. But if his opponent should sow a muzi in, it is captured. In other words, a
player may not directly play a muzi into his captured hole, and he must avoid, if pos-
sible, relaying a muzi into his captured hole. A player may, however, relay a muzi
into a captured hole if under standard relay procedure it is unavoidable, and thus
capture it. For example, if it is singleton, and the next hole is captured, or if it is dou-
bleton and both of the next holes are captured, etc.
14. A captured hole is added to the sector of its owner. That is, he is allowed to play a
singleton from it.
15. The singleton in a captured hole, except it be removed by its owner, is captured or
relayed normally.
Victory: A player who is reduced to a single regular hole has lost the match.
Affinities: The functioning of the captured holes is something like that in a game from
Tengchong County, Wuhe Lianmeng (Eagle, 1995, p. 58), but here the leading player
has much more flexiblity in their use. The awarding of a hole as a bonus to the player
with the last stone in play has not been reported before.
3. A game from Baoshan municipality.
Principal Informant: Yang Guichang, male, 60, Han nationality.
Venue: Baoshan municipality, Xinjie Township, Shanjiao village, about 10 km south of
Baoshan city.
Date of interview: April 1996
Name of the game: Laomuzhukeng. Chinese laomuzhu, as elsewhere, is “the old sow”;
keng is a hole
Format: 2 x 5. One large stone in each row, the owner of the row free to place it in any
of the holes. Large stones are called laomuzhu, small stones called zhuer (piglets).
The game may also be played with 3 or 4 rows, and by 3 or 4 players, respectively.
1. Play is in either direction, with usual Pussa-kanawa relays and captures. A player
plays only from his own side, except as in 4. below.
2. Any zhuer together in a hole with a laomuzhu belong to the player on whose side the
hole lies, and may be immediately removed. Thus the laomuzhu is always singleton.
She is played, relayed and captured normally. If two should fall together, just one is
captured, again by the player on whose side the hole lies.
3. When both laomuzhu have been captured the move capturing the last one ends nor-
mally, but then the round ends, each player capturing what remains in his holes. If
at the end of a move the only stone still in play is a singleton laomuzhu, the round
ends and the laomuzhu is taken by the player on whose side she lies.
4. If a player’s holes are empty and a laomuzhu is still in play, he plays from his oppo-
nent’s side of the board.
5. Laomuzhu are not sold back. If a player has captured both, he keeps them, and puts
each of them in one of his holes. Both players fill as many of their holes as they can.
The trailing player loses the holes he cannot fill. They become the shuitian (paddies)
of the leading player. He places a singleton in each. They must be consecutive, if
there are more than one, and they must start at an end-hole. The leading player plays
Captured Holes:
6. The shuitian are sinks, that is, any stone sown in a shuitian becomes the property
of its owner, and is out of play. They are sown into normally by both players, but
neither played from, captured from or relayed from. Laomuzhu are sown into them
normally by either player. In reckoning capture and relay they are entirely
Victory: The game is played until a player has no holes left. If a player is reduced to a
single hole, but has a laomuzhu, he may battle back.
Terminology. An ordinary hole is a wo. When a laomuzhu is moved or relayed it is said
to tiao (leap). Zhuer do not leap, they simply zou (walk). For a player to move is dong.
When capturing pussa-kanawa, one may say: ou wo chi, ou wo chi, where wo means hole,
chi means to eat, but the informant isn’t clear what character to write for ou, which is
pronounced with a high level tone. It’s meaning is “to scoop out”, and when the player
says ou wo chi” he performs the motion of scooping out the empty hole with his fingers
to show that it is empty. Alternatively, one may say: ge wo chi, ge wo chiwhere wo and
chi are as above, and ge means “empty,” but Mr. Yang has no idea how to write it. Thus
both ou and ge are local dialect.
Affinities. The game is similar to yucebao, described from a Bai nationality village in
Lijiang County (see Eagle 1995). The principal difference is that in yucebao captured
holes function like regular holes, while in laomuzhukeng they are sinks. Also, ending the
round when both laomuzhu are gone hasn’t been reported before.
4. Piggyback. A game from Baoshan municipality.
Principal Informant: Tao Rusong, male, 68, Han nationality.
Venue: Baoshan municipality, Hetu Township, Xiacun (lower village), several kilometers
east of Baoshan city.
Date of interview: April 1996.
Name of the game: Laomuzhuqi “Old sow chess,” as above.
Format: n x 5, where n is the number of players, and may be 2, 3, 4, or 5. One large
stone, the laomuzhu, in each row, and five small stones, the zhuer, in the other four
holes. Each row belongs to one player, and each is free to put his laomuzhu where
he likes. Boards are drawn in the ground with a stick, and are rectangular, divided
into squares, not a series of holes. (Mr. Tao states that he himself has played with
five players and rows, but this is maximal. He is positive that he has not played since
the age of eight, when all children in the village played, and that no-one plays any-
Preliminaries: Throw fingers (huaquan) to see to see who goes first, in subsequent rounds
the trailing player plays first.
1. Play is in either direction, with standard Pussa-kanawa relays and captures. Each
player plays only from his own row.
2. The first player is free to choose the direction of play, and if there are more than two
rows, and the first player is on an inner row, he decides which way to turn on rea-
ching the end of the row. But once the direction is established it cannot be changed.
Adjacent rows are sown in opposite directions.
3. Any stone sown into a hole with a laomuzhu, and any stones in a hole into which a
laomuzhu is sown, stay together with the laomuzhu, and if two laomuzhu are sown
together, they stay together, are relayed and captured together: the piglets stay with
their mother. In effect, the zhuer travel piggyback, and whoever captures their
mother captures them too.
4. When a player whose turn it is to move has only empty holes, the game ends. The
stones remaining in the other players’ holes belong to no-one. The captured stones
are counted, the laomuzhu counting five, and the player with the greater number of
stones is the winner.
Affinities: The use of the laomuzhu is quite unusual: in no other reported mancala game
does a group of stones travel as a group, not “spreading out” as sowing normally
Courlander, Harold. 1943. “The Ethiopian Game of Gobeta.” The Negro History
Bulletin. October: 21-23.
Deledicq, André and Assia Popova. 1977. Wari et Solo: Le jeu de calculs africain. Paris.
Driberg, J.H. 1927. “The Game of Choro or Pereaüni.” Man, No. 114 (September):
168-172; Man, No 127 (October): 186-189.
Duncan, Thomas and Tod F. Stuessy, eds. 1985. Cladistic Theory and Methodology.
New York.
Eagle, Vernon A. 1995. “On Some Newly-Described Mancala Games from Yunnan
Province, China, and the Definition of a Genus in the Family of Mancala Games.”
in: Alexander J. de Voogt, (ed.), New Approches to Board Games Research: Asian
Origins and Future Perspectives. Leiden.
Minelli, Alessandro. 1993. Biological Systematics: The state of the art. London.
Murray, H.J.R. 1952. A History of Board-Games Other than Chess. Oxford.
Panchen, Alec L. 1992. Classification, evolution, and the nature of biology. Cambridge.
Pankhurst, Richard. 1971. “Gabata and Related Board-Games of Ethiopia and the Horn
of Africa.” Ethiopia Observer, 14 (3): 154-206.
Pankhurst, Richard. 1982. Gabata and Other Board-Games of Ethiopia and the Horn
of Africa; Azania, 17: 27-41.
Russ, Laurence. 1984. Mancala Games. Reference Publications, Algonac, Michigan.
The Folk Game Series: No. 1.
Sanderson, Meredith G. 1913. Native Games of Central Africa; Jour. Royal Anthrop.
Inst. Gt. Brit. & Ireland, 43: 726-736.
Santos Silva, Elisio Romariz. 1995. Jogos de quadrícula do tipo mancala com especial
incidência nos praticados em Angola. Lisbon.
Townshend, Philip. 1976. Autour du jeu de Mankala. Zaïre-Afrique, No. 105 (May):
Townshend, Philip. 1977a. Les jeux de Mankala du Zaïre, du Rwanda et du Burundi
(Cahiers du CEDAF [Bruxelles] no. 3): 1-76.
Townshend, Philip. 1977b. Mankala Games. International Committee on Urgent
Anthropological and Ethnological Research, Bull. No. 19: pp 47-54.
Townshend, Philip. 1977c. The SWA game of ||hus (das Lochspiel) in the wider context
of African Mankala. Jour. SWA Scientic Society, 31: 85-98.
Townshend, Philip. 1978. Review of Deledicq et Popova 1977. Journal des Africanistes,
T. 47, fasc. 2: 207-209.
Townshend, Philip. 1979. Mankala in Eastern and Southern Africa: a Distributional
Analysis; Azania, 14: 108-138.
Townshend, Philip. 1982. Bao (Mankala): The Swahili Ethic in African Idiom;
Paideuma, 28: 175-191.
Townshend, Philip. 1986. Games in Culture: A Contextual Analysis of the Swahili
Board Game and its Relevance to Variation in African Mankala. PhD. Thesis,
University of Cambridge, Department of Social Anthropology.
Wiley, E.O. 1981. Phylogenetics: The Theory and Practice of Phylogenetic Systematics.
New York.
1. For descriptions, terminology and bibliography, see Murray 1952, Deledicq & Popova 1977,
Townshend 1979, Russ 1984, Santos Silva 1995, Eagle 1995.
2. On the taxonomic terms and concepts discussed here, and for extensive references, see Duncan
and Stuessy 1985, Minelli 1993, Wiley 1981.
3. Townshend reports at second-hand, but does not confirm, the contrary possibility: a solo-(a)
game with capture from opposite a non-empty hole receiving a final counter. He calls this
“Intermediate Type C”.
4. This would appear to be a notable example of convergent evolution, as Cisolo is otherwise
quite unlike Pussa Kanawa. On Pussa-kanawa games, see Eagle 1995.
Chinese character glossary
The Development of the English Board Game,
1770-1850 / Caroline G. Goodfellow
Throughout the 18th century and well into the 19th century, the size of the
middle income group of merchants, solicitors, doctors and industrialists grew.
Trade flourished and unknown areas were explored. The adventurers who were
prepared to open shipping routes and establish trading agreements reaped rich financial
rewards. This was an age of enlightenment, invention, innovation and scientific
discovery. Games were a part of the industrial and social life of entire nations, reflecting
changing ideas and ideals, particularly during periods of major upheaval.
The upbringing of children within these middle class families changed dramatically.
Education became essential, covering not only the `three Rs’, but sensible grounding in
national and international affairs. National pride and achievements were stressed, as were
faults. In general, everyone seemed to be looking outwards, to try to understand new
concepts. We must, however, when viewing the games of this period, remember to set
them against their own time rather than to evaluate them in the light of modern history,
knowledge or ideals.
The publishers of many of these games were already established producers of maps
and books, many of which were aimed at children. The idea of creating an educational
tool was, in a way, a novelty. The Game of Goose was already well known and it required
few changes to create The Game of Human Life or the History of England. The games were
well received by parents who appreciated the educational aspects, the children’s resulting
enjoyment and possibly that the games could be played in relative peace and calm.
The early games stressed learning through play, but this aspect was gradually
dropped in favour of sheer enjoyment of play. However, not all games were that
enjoyable despite the claims of their titles. Perhaps the whole logic of such games was
summed up by John Harris in the introduction or Advertisement as it was called to his
game Historical Pastimes which was published in 1810.
The utility and tendency of this Game must be obvious at first sight; for surely there
cannot be a more agreeable study than History, and none more improving to Youth,
than that which conveys to them, in a pleasing and comprehensive manner, the Events
which have occurred in their own Country.
The little prints, which are regularly numbered from 1 to 158, represent either
Portraits of principal Personages who have signalised themselves as Kings, Statesmen,
Churchmen, Generals, Poets etc., or some remarkable Occurrence in our Country. This
will naturally excite a curiousity in the youthful mind; and that curiousity will be gratified
in the short account of each reign subjoined. On the whole, the writer flatters himself,
that the public approbation will convince him, that the hours he has devoted to the
formation of this little Scheme, have not been spent in vain.”
It is known from surviving records, paintings and artifacts that games of the period
(and today) are played in a similar way to those of ancient civilisations in the near and
far east. There are only a few basic methods of playing games and over the past 250 years
many thousands of variations have been created. The basic methods of play come down
to four types – race games, strategy games, table games and card games. Many were
originally developed for adults and were later adapted for children. Today I will be
discussing only board games developed in the United Kingdom.
It was the race game which above all became the basis for the educational game of
the late 18th century and first half of the 19th century. The aim was to win but it was
played with an element of chance and normally gambling was included by the means of
a central kitty or pool into which players placed an agreed number of counters.
Throughout play, rewards were given and penalties paid; sometimes these were the
receiving or forfeiting of counters and at other times forward or backward movement on
the board itself, though not the removing of markers from the board unless there was a
provision for retirement from the game.
Why was this format produced for the wealth of educational games, including those
teaching morals and behaviour, even though they retained the element of gambling and
chance? Quite simply they were exciting to play. Astute publishers could use this
characteristic to great advantage to encourage learning and reading skills.
The Game of Goose is generally regarded as the prototype of the modern race game.
Devised in Italy and taken from much early formats of games played in the Middle and
Far East, it was first noted in England by John Wolfe in 1597 as “the Newe and Most
Pleasant Game of the Goose”. There were usually 61 or 63 squares or compartments
which stressed good and bad behaviour. Of these squares, a number were plain and a
number decorated with either a scene or a goose. Landing on a goose was a good move
and rewarded, often with the words “double his chance forward”, while if landing on a
bad square such as the Ale House a double penalty was paid usually in the form of a
counter to the pool and waiting until all others had taken a turn. There were also very
severe penalties such as death or overthrowing the end of the game, which required the
player to either start again or withdraw from the game completely.
One of the earliest versions of this game was published about 1725 by John Bowls
& Son and one of the latest during the 1980s. In general the format was a spiral or a
flattened spiral but occasionally a new design was made, for example a one published by
Richard Holmes Laurie, November 22, 1831. It shows a huge goose with three golden
eggs set against a country scene.
Variations of a theme were numerous. Laurie & Whittle, successors to Robert Sayers
(they in turn were succeeded by Richard Holmes Laurie) continued to publish The Royal
Pastime of Cupid Or Entertaining Game of The Snake which was based on the ancient
Egyptian Game of the Snake, however, possibly in design only and not in rules. The
reissuing of existing and popular games, often by successors to a business, was common.
In many cases, no changes were made; in other case’s additions were included to update
a game, particular ones on history or geography. All that a publisher needed to do was
change the name and date of publication.
Perhaps one of the best games of a goose was The New and Favourite Game of Mother
Goose and The Golden Egg. It shows episodes and characters from the pantomime of the
same title and includes Clown and Harlequin and places in London. It was published at
Fig. 1: Wallis’s New Game of Universal History and Chronology, 1814. This game has
as its centre illustration George Prince of Wales, later George IV.
Fig. 2: The second example of the game, published about 1840, has the centre replaced
with medallions showing George IV, William IV and Queen Victoria. The last
scene shows a railway bringing the game right up to date.
the height of the career of Grimaldi the Clown, in September 1808, by John Wallis.
Many of the first publishers of games were in fact cartographers and they quickly
included the race game idea into game of Geography. The spiral format was not used; it
was replaced by a map – of England and Wales, Scotland and Ireland (collectively or
individually), of Europe or of the world. One of the first publishers was Robert Sayers,
working from 53 Fleet Street. He published much for children including puzzles and
other games. Having started in 1745, he was succeeded in 1794 by Laurie and Whittle.
The ideas for geographical games are summed up in a statement in the booklet of
Geographical Recreations or A Voyage around the Habitable Globe, 1809.
The game, consisting of 116 little prints of the most interesting objects in
Geography, is designed to familiarise youth with the names and relative situations of
places, together with the manners, customs and dresses of the different nations in the
habitable globe; and, as curiousity will naturally be excited by the scenes which present
themselves, and the observations likely to occur, it is presumed that these, with an
occasional reference to the Synopsis herewith given, will prove a continual source of
amusement to young people of both sexes, and will furnish such a fund of geographical
knowledge, as may prove equally beneficial in reading and conversation.”
This statement shows the high hopes that many publishers had for their games. The
geographical ones could be boring to look at, they were simply maps, often surrounded
with vast texts. One wonders now as we play these games, were the youths of both sexes
excited, amused or entertained by any of them. Often the titles were more exciting than
the games. There were The Tours of Europe, England and Wales by John Wallis, The
Royal Geographical Amusement of The Safe and Expedious Traveller, The Royal Geo-
graphical Pastime. By the mid-19th century, many geographical games were devised on a
pictorial note rather than simply maps.
The historical and geographical games are quite illuminating when studying them
for the context. To the modern reader, one may be surprised to see how well informed
many of the comments could be. For example in A Tour Through The British Colonies
and Foreign Possessions, published by John Betts, there is expressed disapproval in the
selling of alcohol and weapons to North American Indians and the accepted fact that
Sebastian Cabot rediscovered Newfoundland in 1496. William Darton, on the other
hand, with his beautiful game The Noble Game of The Elephant and Castle, Or Travelling
in Asia, 1822, seemed quite mystified by the strange ways and creatures found there for
which there was no ‘rational’ explanation; this included a woolly mammoth.
The games of geography essentially looked at the world from a British viewpoint
but the publishers did acknowledge that other places and people might be interesting.
The history based games, however, were altogether more inward looking and usually
strictly based on events as they affected the British Isles. During the period under
discussion, King George III was on the throne for most of the time, from 1760 to 1820.
He and his son became the focal points for the games whether or not the events actually
happened during their lifetimes. The second rather amazing fact is that for many of the
history games, nothing happened before 1066 and the invasion of William the
There are a few exceptions. One was The New Game of Universal History and
Chronology, published in 1814 by John Wallis. This game starts with Adam and Eve,
Anno Mundi 1. All the dates are given with great conviction, for example the Universal
Deluge occurred in AM 1636. One now wonders how this knowledge was lost between
then and today.
The games of history had many squares, often more than 150, and of course by
reducing the period of history covered more details could be included. The games were
often re-issued and reworked with changes being added in or adapted, especially with
the death of the kings. George the Third reigned for so long, that a game was published
by John Harris to celebrate his Jubilee in 1810. The squares showed events of his lifetime
and the text in the accompanying booklet was very flowery when in praise of the king
but was equally robust when describing some of the more disagreeable events. Included
also were scientific and geographical discoveries.
Ten years later John Harris reworked The Jubilee as The Sun of Brunswick to
celebrate the new king, George IV. The last 20 odd compartments were redone to
include the deaths of Queen Charlotte, wife of George III, Charlotte Augusta the
daughter of George IV, who was in a fact heir to the throne and known as Princess of
Wales, and of George III himself. By the 1840s, Queen Victoria was on the throne and
one of the most important new inventions was the railway, both were added to reissue
Gradually both history and geography-based games were replaced by other forms of
games and of course, methods of learning. But these subjects were not the only ones to
be treated in this way. Mathematics, natural history and languages were also subjected to
the race game format though generally these games were not as attractive to look at.
However, if played correctly, they were enjoyable ways of learning.
An Arithmetical Pastime was published in 1791 by C. Taylor. It had 100 circles, some
of which contained illustrations while others had directions and forfeits to be paid. The
forfeits were repeating tables, some of which were obvious such as the times tables,
others were less so, for example the wine measures. Other tables are not now used – the
avoirdupois, for example, measured pounds and ounces. If a player could not meet the
forfeit, he had the choice of missing turns or moving backwards.
A later version, with the same title, had quite different rules and used two teetotums
for the moves and to learn the mathematical disciplines. The players had to subtract,
multiply or divide the two numbers given by the teetotums and the result formed the
move. Used to accompany the game was a set of verses to be read out and a list of further
rewards or forfeits. This version was the ultimate teacher as it taught everything - morals,
history, geography and arithmetic.
inked to mathematical games were those of astronomy. Science in Sport or The
Pleasures of Astronomy, published 1804 by John Wallis, had 35 compartments with the
portraits of astronomers and representations of astronomical phenomena. At the time
nine planets and their movements around the sun were known, however, these were
added to with fictional representations, for example The Man in the Moon. Often
behaviour traits, The Studious Boy and the Blockhead, signs of the zodiac, comets,
Fig. 3: The Noble Game of the Elephant and Castle is a geographical trip through
Asia, showing the latest archaeological finds including a woolly mammoth.
Fig. 4: Virtue Reward and Vice Punished is probably the best known of the “moral”
games. Many of its teaching are as valid today as they were in 1818.
rainbows and even known astrologers were shown.
The Circle of Knowledge, published about 1845 by John Passmore, included the
zodiac around the central compartment. Unlike many games of the time, it was arranged
in four concentric circles. As well as the zodiac and the four cardinal compass points,
Europe, Asia, America and Africa were shown, together with the four seasons, the four
houses of the zodiac and the four sciences - electricity, chemistry, optics and astronomy.
The illustrations too were unusual, for example fire was illustrated by a volcano, a
burning farm and a pit explosion while optics were shown as a giant telescope, a magic
lantern show and the perspective of a tunnel.
One of the best games of natural history was William Dartons 1820 version of
British and Foreign Animals.The subtitle again gave graphic details about the aims of the
publisher with “A New Game, Moral, Instructive and Amusing, designed to allure the
Minds of Youth to an Acquaintance with the Wonders of Nature.” Both domestic and
wild animals were included, even the Australian kangaroo. A companion game published
by Darton in 1822 had the wonderful title The Delicious Game of The Fruit Basket or
The Novel and Elegant Game of The Basket of Fruit. Sadly only the design and title were
delicious as the game dwelt mainly on penitentiaries and trial by jury, the Royal
Academy, hospitals, national schools and the School for the Blind, sciences and religion.
Both games while following the principles of rewards and forfeits were very much a
teaching tool.
William Dartons games seemed to have a sombre side as many were very moral and
rather disagreeable in tone yet beautifully executed. He later produced one which should
have cheered everyone up, called A Voyage of Discovery or The Five Navigators, 1836.
However, it dealt with the dangers and incidents likely to befall sailors.
We have touched on the games teaching history, geography and so on but the final
main group was that of morals; these games were extended forms of the original Game
of the Goose. Within many of the previously mentioned games, morals were included,
however, the publishers devised many more which were very strict. Often they had good
titles and possibly the best was The New Game of Virtue Rewarded and Vice Punished for
The Amusement of Youth of Both Sexes, published in 1820 by William Darton. A
generation earlier, a similar subtitle was “Designed for the Amusement of Youth of Both
Sexes and Calculated to Inspire their Minds with an Abhorrence of Vice and a Love of
Virtue.” Taken from The New Game of Emulation, published by John Harris in 1804,
this game actually showed many images a child would actually come across, a shepherd
with his flock, a church, school, even a rocking horse while the almost cryptic words
cheerfully exert themselves to obtain an honorary prize” while being “perfectly aware of
the consequences of disgrace and naturally dread it” rather deadened the pleasing
Two methods of teaching morals or behaviour were produced, though both used the
same games format. One method was based on the passage of life from infancy to old
age and death and showed the temptations one might meet along the way. These tended
to have straight formal titles such as The Game of Human Life and more often than not
were based on the male life to the exclusion of females. However, the games were not
restricted to boys playing them, as the subtitle to The Game of Human Life states “...most
Agreeable and Rational Recreation ever Invented for Youth of Both Sexes.” Parents were
encouraged to instruct their children on each of the characters, usually 84, with “a few
moral and judicious observations... and contrast the happiness of a virtuous and well
spent life with the fatal consequences arising from vicious and immoral pursuits”. Most
of the characters have the same meanings today although some of the pursuits would not
now draw penalties, for example The Romance Writer must pay two and move back to
the Mischievous Child and the Dramatist must pay four and begin again. The Tragic
Author has the harshest as he advances to the Immortal Man and dies but to compensate
him he actually wins the pool or kitty by finishing first. The game quoted was published
in 1790 by John Wallis and Elizabeth Newbery. (While we are discussing only English
games of this period, there are some very fine French and German examples of The
Game of Human Life, often far more intricate in design.)
The second method, developed slightly later, was stricter and based on the
fundamental principles that if you behave properly all bodes well, but if you slip from
the straight and narrow the penalties are severe. These titles were much more lucid with
The Cottage of Content or Right Road and Wrong Ways, The Journey, The Mansion of Bliss
and The Mansion of Happiness. (This latter one is also the title of one of the first
American board games, published by W & S B Ives in 1843.)
Penalties were paid for often routine bad traits – straying, boasting, wasting time,
idleness, self-indolence, obstinacy, ignorance, pride, conceit and forgetfulness; other
traits were considered very bad - theft, lying, drunkenness, cheating. In most of the
games there were more bad than good traits – recollection, repentance, patience,
kindness, exertion and diligence. Snakes & Ladders was the game which took over the
teaching of morals, and it was a version of a game introduced from India. Like the earlier
versions, there were more snakes, the baddies, than ladders, generally 12 to 8.
By the middle of the 19th century new processes were being introduced,
lithography had been developed which was cheaper than the engraving and etching
processes, even though the publishers had been one of the first groups to involve
themselves in new technology – the use of static steam presses instead of hand operated
ones. Many of the original group had died. The men and women who developed these
early teaching games seemed to be inspired and that inspiration seemed to die with
them. New publishers, new methods of production, new methods of teaching and new
games formats changed the look of the board games and the very reasons for their
As we study them today, however, the games remain excellent teaching tools. We
learn from these games the history and social life of the people, what they felt was
important and how they regarded themselves in relationship to the rest of the world and
other cultures.
Publishers Mentioned
John Wallis, with sons John and Edward; one of the most prolific publishers of games
and dissected puzzles between 1775 and 1847. Also worked with John Harris and
Elizabeth Newbery.
Elizabeth Newbery, part of the leading publishing family of children’s literature during
the 18th century; John Harris managed her establishment. Worked with John Wallis
John Harris, took over Elizabeth Newbery’s business in 1801 and worked until it was
sold in 1843.
Robert Sayer, one of the earliest publishers of childrens games, 1745-1794.
Laurie & Whittle, Robert Laurie and James Whittle; acquired the business of Robert
Sayer in 1794. Richard Holmes Laurie succeeded Robert on his death in 1812 and
James Whittle on his death in 1818. Also worked with other publishers including
William Darton.
William Darton, established 1787 and under went many name changes depending on
the partners and sons. The William referred to here was the son who opened his own
establishment in 1804.
John Betts, leading 19th century publisher between 1827 and 1874.
John Passmore, published from 1840 to 1869.
Bul: A Patolli Game in Maya Lowland / Lieve Verbeeck
comparative ethnographic research on colonial and contemporary Mesoameri-
can(1) board games revealed that the Maya board game, called bul, played by the
Mopan and K’ekchi’ farmers in Southern Belize(2), is a native American game.
There is a well-marked affinity and relation between the pre-Conquest “game of the mat
and patol beans” of the Aztec, called patolli, and its various twentieth century manifest-
1. Introduction
In the pre-Conquest times games of chance employing beans or reeds as dice were quite
popular and widespread in Mesoamerica. Although there is an abundance of ethno-
historical documents, no accurate description has been found of how these games were
played. Even the well-known Aztec board game of patolli is still a riddle. Only the
superstitious” aspects of the game and the heavy betting that went with it are well
documented. If we are to believe what the earliest Spanish chroniclers wrote about the
native American games of chance, we must assume that by the end of the sixteenth century
the Mesoamerican games were abolished (fig. 1) and replaced by Spanish or Old World
games (Duran 1967, Sahagún 1981). Besides, the twentieth century ethnographers and
anthropologists do not show much interest in the games of the native Americans either.
The reason probably is that there is no direct demonstrable association between modern
recreational games and divination. No doubt, in ancient times the Mesoamerican games
must have had a mantic significance, but at the eve of the Conquest sheer gambling was
the main objective of the native American gamesters. On the other hand ethno-historians
and archaeologists are still studying the various designs of the patolli boards that have been
discovered in ancient sites all over the Mesoamerican area. Although there are still many
questions unanswered, it is generally accepted that the patolli boards are cosmological
images (Caso 1924-27, Duverger 1978, Swezey and Bittmann 1983).
By now the term patolli has become a generic term. It does no longer signify one
specific Aztec game of chance, played on a mat on which there was drawn a cruciform
board, with four black, marked patolli beans as dice (fig. 2). Patolli now labels any
variant of the square, cruciform or circular game-boards drawn or incised on floors or
benches of ancient Mesoamerican buildings (figs. 3 to 6), or featuring in the multiple
pre-colonial or early-colonial codices (figs. 7 to 9), as well as some of the twentieth
century games of chance that are assumed to be survivals or variants of the ancient game
of patolli. For indeed, in some remote areas, safely away from the surveilling and
punishing Spanish authorities, indigenous groups preserved their ancestors’ games of
chance well into the present century.
Nevertheless, it must be considered a lucky coincidence that this author recently had
the opportunity to observe a Maya board game in the field(3). In the tide of modern
civilization and technology even the most isolated communities are swiftly substituting
their cultural heritage for the “blessings” of westernized societies. And thus, the
Fig. 1: Execution of a patolli player. His patolli board, dice, counters and bundle with
supersticious objects are being burnt (Relaciones Geográficas: Tlaxcala, Tomo
I. 241v 11).
Fig. 2: Patolli, codex
Florentino, lam.
XLVIII, nr. 63,
reproduction from
the work of
Sahagun, Madrid,
Fig. 3: Patolli,
El Tajín, Mexico
(after Duverger,
1978, fig. 4c)
Fig. 4:
patolli Type I
(after Mackie,
Fig. 5.
Mexico, patolli
Type II (after
Bernal, 1963,
lam. 9:3)
Fig. 6: Chichen
Itzá, Mexico,
patolli (after
Ruppert, 1943,
fig. 4c)
Fig: 7: Patolli in Codex
p. 19, facsimile edition,
commentary by Hamy, Paris,
traditional games are being dismissed by the younger generations. The registration of the
ancient games is becoming an urgent issue, not only for native American folklorists and
ethno-historians but also for the indigenous groups themselves, if they are really
concerned about safeguarding various aspects of their cultural heritage.
The study of Mesoamerican board games comprises a large and still unexplored field.
In the scarce literature on contemporary Mesoamerican board games most descriptions
date from the first half of the twentieth century. The ethnographers seldom gave an
accurate reflection of the rules or the playing context and often did not bother to make
a distinction between Old World and New World games. Ventur’s structural description
of the Mopan dice games stands in contrast to the vague accounts in most colonial and
even modern sources (Ventur 1980: 257). The purpose of this paper is twofold: first, to
present the Belizean version of the game of bul, as it was observed in its natural context
by this author, and second, to prove that the board game bul indeed is a modern variant
of the ancient Mesoamerican patolli game. For that purpose the data on bul will be
compared with what is known about a few other Mexican board games. At the same time
this comparison should result in a tentative typology of the patolli games.
2. Bul, a Ceremonial Game
One day in the month of May, my Mopan hosts invited me to the customary vigil
ceremonies they perform before planting their corn. For that purpose the helpers at the
next day’s planting use to come to the hut of the farmer to spend the night with him.
After the habitual prayers and incense-offerings to the God of the Earth and the Wood,
Santo Witz, Santo Hook, Santo Che’, in front of the house altar, the men told me that
now I was going to witness an important part of the ceremony: they were going to “play
corn”. One Mopan man took some grains of corn out of one of the corn bags in the hut
and put them on the floor in a straight line. In the meantime the others went outside to
look for suitable counters, each of them returning with five similar small pieces of twig,
leaf stem or grass, different from the counters of the other players. They formed two
teams and squatted on the floor, in front of each other with the corn “track” between
them. One player then looked for four suitable grains of corn to make the lots, the corn
dice. One side of each grain was blackened with charcoal he took from the cooking fire.
And then the game could be started. It appeared to be a kind of a war-game. The players
moved their men up and down along the corn-track by throws of the four corn-dice,
called bul, which is also the name of the game. During a break the players drank large
cups of their traditional cocoa. At a certain moment the farmer took the incense burner,
lighted its fire and went outside the hut to pray again to the God of the Wood. I learned
later that, in order to enforce their supplication for a rich corn harvest, the farmers
mention the playing of the bul-game in their prayers as another ritual obligation that is
being fulfilled (Verbeeck 1996: 84). Notwithstanding the ritual character of the game
the atmosphere among the men is very joyful and exceptionally loud. It is very unusual
to hear the retiring Maya laugh and shout boisterously. Women never play or even watch
the bul-game. But they follow the proceedings of the game in the kitchen with great
interest. Judging from the men’s exclamations and remarks, they know who are winning.
Fig. 8:
Patolli in
Borgia, pl.
62 (Anders
& Jansen,
1988: 54)
Fig. 9: Patolli,
fol. 60,
loubat, Rome
1904; Graz
They enjoy imitating their excited cries for a favourable throw of the bul. That evening
the whole play took more than three hours. By then it was midnight and the due
moment had come to close the vigil with the ceremonial meal that consists of wah tel
chicharron, corn tortillas with pork rind in broth. The bul and the counters were thrown
away, nobody cared about saving the game instruments.
It was confirmed by other Maya informants that the bul-game is an essential part of
the rituals and ceremonial obligations of the “vigil of the maize”. In the richer villages
however, there are music and dancing besides or instead of playing bul. What seems to
be important is that the corn, which is going to be planted the next day, should be
surrounded by bright joyfulness the night before it will go down into the “dark earth”
(Pacheco 1981: 104). This is probably the reason why I observed so little
competitiveness during the game. The general atmosphere of that bul- evening radiated
harmony, joy and fun. It did not matter at all who won or lost; what was important was
the cheerful playing together. The function of the bul-game is to create a foreshadowing
of the ‘alegria’ that will reign at the harvest of the corn.
Only the catholics among the Belizean Maya still maintain the old costumbreof the
vigil of the maize. As they are becoming a minority, the bul-game is gradually falling into
oblivion and with that another element of the old Maya traditions threatens to
3. Description of the Game of Bul
3.1. Players and tools
Bul can be played with any even number of players above six. They play in two teams
inside the house, squatted in front of each other around the game board. The board is
marked on the clay floor of the hut by twenty grains of corn. The grains are placed in a
straight line, some 5 cm apart, the intervals being the points of play. The board is called
bej, the ‘road’, which is the circuit the players have to run up and down from their
starting point. Depending on the number of players, more than ten or sixteen, the road
is lengthened with five or ten more grains respectively (fig. 10).
Every player has selected his own five counters, recognizable by their specific
material, colour or length. They consist of five similar pieces of equal length of twig,
leaf stem, grass or any other oblong object measuring between 4 and 10 cm, which can
be found in the surroundings of the hut. Bulb shaped counters, like berries, cause
hilarity among the players, first of all because the person who introduces these irregular
counters proves himself to be lazy or not well acquainted with the rules of the game.
But secondly, there is a humoristic linguistic aspect to the deviant shape of counters,
because during the game these objects must be referred to as tziit, the Mopan numeral
classifier for oblong objects. The numeral classifier for bulb-shaped objects, kuul, is
used for the grains of corn, which are the other game tools: they form the track and are
used as dice. In the course of the game the players constantly shout to their partners
the number they should throw in order to land on the “right place”. In their
exclamations “one!” (hun kuul), “two!” (ka’ kuul), “three!” (ox kuul), etc., the word kuul
is always in the air. This intentional linguistic confusion of the dice with the counters
is typical for Mopan humour.
The counters are moved by the throws of the four bul. These dice are four flat-sided
grains of corn, so that they only have two sides to fall on. The grains are prepared by
digging out with the thumbnail the eye on one side of each grain. This is called k’o yik u
päsäk’al a ixi’imi, “to pick the heart of the corn”. Then the hollow of each grain is marked
with a black dot, either by rubbing charcoal in it or by using the live end of a glowing
stick. This operation is called bonik tel butz, “to give colour with charcoal”, or jo’ochtik
tel butzor “rub with charcoal”. The black-spotted side of the dice is called u wich a bul,
the face of the dice”, the unmarked side is called u yit a bul, “the bottom of the dice”.
The value of the throws is determined by the number of black dots that fall upwards:
Fig: 10. Mopan
Maya playing bul
(photo L. Verbeeck
ka wila’ bon a jäwa’ana, “look how many are lying on their back”. This may be one, two,
three or four. If the four unmarked sides have fallen upwards, le’ek wa laj päklaji, when
all lie face downwards, the score is five(5). The bul are simply thrown from the hand onto
the ground. While a player is preparing his throw, by shaking the bul in his hands, the
others are anxiously following his movements whispering tun kaxäl (“they are falling”).
3.2. Rules
Each player has two throws in a turn. He moves his counter after the second throw,
advancing it according to the score of each throw, in arbitrary order. This is impor-
tant because it enlarges the possibilities of capturing an adversary.
The home fields of the teams lie at their left end of the “road”. Thus the teams enter
their men from opposite sides and move in opposite directions. When they have rea-
ched the opponents’ field, they return to their starting point along the same “road”.
It is not necessary to throw an exact number to enter the home field.
The first men of both teams make a throw to decide who starts the game. The
highest throw wins. After the first player entered a marker, the other members of his
team, from left to right, each throwing twice, enter one counter. Then the oppo-
nents get their turns to advance their men from the opposite side of the board, etc.
Each player can only have one movable man at a time on the board. When he has
reached his home field safely, he re-enters that man.
But it is the hope of every player to land into a space occupied by an opponent. In
that case he starts returning back to his home field carrying his opponent as his cap-
tive. The opponent loses his man and enters another counter at his next turn.
Whenever a player captures an opponent he moves directly backwards towards his
home field. But this shortened track does not guarantee his safely passing out, becau-
se the combined men remain vulnerable. If any player of the opposite side plays his
man to the point occupied by the reversing men, he puts his counter on top of the
little stack and moves all of them back to his own home field. This man in his turn
may be taken, losing himself and his prey. They will be reversed again in the oppo-
site direction towards the last captor’s home field. There the captives are retained.
The counters belonging to the partners of the winner are “liberated” and returned to
their owners, who enter them again. The number of these takes and retakes is in fact
unlimited. The accumulation of counters increases the excitement of the players. A
stack may be captured by another stack.
Doubling a space occupied by a partner is permitted and does not change the play
of either.
Players never throw more than twice. If the first throw takes an enemy’s counter, the
second one counts towards carrying him home. If the first throw brings a player safe-
ly home, the second can be used for re-entry on the board.
No player loses his throw. If he has lost his fifth counter, he continues to throw the
bul to help his partners. However, Ventur presents a restriction in the Guatemalan
bul game. If all the markers of a player are ‘immobilized, he is temporarily “paraly-
zed”; his turn is passed, and he cannot again throw the dice until the outcome
of these captures is determined’ (Ventur 1980: 251).
The game ends when a team has no counters left to enter. Winner is the team that
captured most of the enemy’s men.
The goal of the game is to capture as many men of the adversaries as possible. The whole
idea shown by the terms of the game and especially by the exclamations of the players is
that of the pursuit and safe transporting of captured load or prey back to the home field:
in mächaj u kuch (“I grabbed his load”), watak ta pach (“he is coming after you”), tak ti
kol (“and now straight to the field”), jobi (“he is killed”).
3.3. Variants of the bul game
The game of bul consists of a set of five variants, played in a fixed order and which differ
from each other in the way of running or attack. Four variants are inspired on local
animals and their specific ways of catching their prey. This determines the rules of each
variant and its name.
The first game, called aj sayil (wee-wee ant) follows the general rules as described
above in 3.2. and is indeed regarded as the basic game.(6)
The second game, aj t’iwil (the eagle), is the quickest variant. The player who takes
an opponent immediately leaves the road with his prey. He does not re-enter his counter.
In the third game, aj sina’anil (the scorpion), a man can move forward and backwards
to capture an enemy. Retakes are possible in this variant. The winning counter is re-
With the army-ants, the fourth variant, aj sakalil, the men keep on moving straight
ahead to the other end of the road, even when they are carrying one or more opponents.
They do not run back to their home field. The winner re-enters the road from his
starting point.
At the start of the last game, a k’aak’il (the fire), there is a small circle drawn with
charcoal in the middle of the road. The player who lands into that circle is burnt by the
fire and his prey will burn with him. If a player captures an adversary before reaching the
fire, he may return immediately. His counter may be re-entered after his safe arrival
After each variant the teams count how many opponents they ‘ate’ or ‘killed’. But in
the end the outcome of the game is not important.
3.4. Some general remarks on bul
This Maya war game obviously does not require much mental skill or calculation of its
unsophisticated players. The only “clever” move a player can make, is to count his two
throws in the appropriate order, when there is an opportunity to take an opponent. In
Ventur’s description this possibility is non-existant, but he points to the sina’an variant
as the most complex of the five games. This game requires some strategic insight as the
player can move his counter forward or backward to capture an opponent, but is not
obliged to do so (Ventur 1980: 253). As was explained before, the purpose of the bul
games is to pass the time during the vigil and ultimately, it is not important who are
the winners. The game is entertaining, not only for the players who can all take part
until the end of the game, but also for the spectators, who love to see the moves that
lead to captures and recaptures of stacks of counters. Although the playing of bul occurs
in a ceremonial, religious context, the elements of the game do not bear any specific
religious meaning. Nor did the players indicate any connection between the game
circuit and the cardinal points and the centre, which traditionally have a strong
symbolic value in the Maya area. The use of grains of corn as game implements is
simply obvious as these kernels are always ready to hand in a Maya hut. The symbolism
of the game expressed in the players’ terminology is not farfetched either: the game
reflects the farmers’ life. Their walking up and down to the field, their carrying a load,
the uncertainties about winning or losing at harvest time: these are the vicissitudes of
life that are fairly familiar to them. The players were quite conscious of the fact that
they were performing an old costumbre”, but they certainly did not bother about the
probably ancient roots of their game. Besides, it is quite possible that the set of variants,
or at least some of them, are the result of a recent, regional development and perhaps
of an Old World introduction. The last variant in particular, in which a ‘fire’ is drawn
in the center of the track, is reminiscent of the game of goose. Another peculiarity, that
might illustrate a development in the game, is the fact that the English translation the
players gave for certain game elements do not correspond with the Mopan word. When
explaining in English certain episodes on the game for instance, the players talked
about ‘bullets’ when referring to the counters, the image of the animals as hunters being
lost. In Mopan the counters have no metaphoric name, they are called che(sticks) or
reference is made to their owner. The ‘fire’ (k’aak’) in the last game becomes a ‘ditch’ in
4. Historical Sources on the Maya Game of Bul
Nearly a century ago Stewart Culin published in his Games of the North American
Indians a K’ekchi’ Maya version of this bul game, called boolik (Culin 1907: 141-143).
A certain Thomas J. Collins had provided him with a detailed description of a corn
game, that was in common use among the K’ekchi’ Mayan Indians in Alta Verapaz,
Guatemala. The ethnographer Karl Sapper described in 1906 a similar game, called
puluc(7), which he had observed not only with the K’ekchi’, but also with other “tribes
of Northern Middle America” (Sapper 1906: 284). The extensive description given by
Culin corresponds more or less with the basic game as described above in 3.2.. The
testimony of Culin’s informant from 1899 only differs on the length of the track. In
that K’ekchi’ version a player only has to run to the opponent’s field, at the other end
of the board. When he has completed his passage of the line without capturing an
opponent, he immediately enters again at his own end of the board. He does not have
to run back to his home field along the line, as today’s Mopan players have to do. None
of the one century old descriptions mention the five variations, which might indeed
indicate a recent development. Nor did the authors make any reference to ceremonial
It is unclear why Murray (1952: 149, number 6.7.6.) classified Culin’s version of
boolik as a race game and not as a war game, as he concludes his description with the
remark that “the game ends when all the men of one side are taken”. Bell based his
interpretation on Karl Sapper’s very brief description of puluc and constructed his own
rules to create a playable corn game. Bell classified puluc as a “running-fight game”, which
is one of the subclasses he differentiates among the war games category (Bell 1960: 89).
Following de Voogt’s classification, all four descriptions of the Maya board game (Sapper
1906, Culin 1907, Ventur 1980, Verbeeck 1996) indeed fit in the class of war games,
their subclass, based on the purpose of the game, being: destruction (de Voogt 1995: 15).
5. Bul, a Mesoamerican Board Game?
How old is this Belizean board game? It is certainly not a recent invention of some
playful Maya Indian. But how to prove that it is not a game the native Americans learned
from their European conquerors? There is no direct evidence of its pre-Colonial origin(8).
From the sixteenth-century chronicler de Landa (1566 [1985]) and the Mayan sacred
book, Popol Vuh (Tedlock 1985), we only learn that the Maya played with dice. But it
remains unclear what kind of dice they used and how they played. The etymology of the
Yucatec Mayan word bul evidences that in the sixteenth century, the Mayans knew dice
games and that gambling was associated with them (Barrera Vasquez 1980). Also the
K’ekchi’ cognate buul, bears the complex meaning of playing a board game and of
winning in a game of chance or a lottery (Haeserijn 1979). Ventur’s exploration in
colonial and modern dictionaries or vocabularies reveals that bul and its cognates gloss
the native dice games as well as all European games of chance and their artifacts (Ventur
1980: 244-46). But this linguistic evidence of the existence of Maya native dice games
does not prove that the Mopan bul game indeed is a variant of an ancient Mesoamerican
game. Comparison of the data on bul with what is known about other Mesoamerican
games of fortune should solve the problem. This will be treated in the remainder of this
paper. At the same time it will be attempted to define the most salient features of those
board games in order to establish a tentative list of Mesoamerican characteristics. This
should distinguish them as a regional subclass from the general class of American race
games presented by Murray (1952:150). Murray based his ‘general characteristics of the
American race-games’ on Culin’s catalogue (1907), which describes one single Central
American dice game. Murray in fact only typified the North American board games.
5.1. Patolli in Mesoamerica
The game of patolli was most popular in Aztec times. According to the chroniclers the
Aztecs had a passion for gambling. In his Historia de las Indias, Diego Duran (1575-81
[1967]) mentions that professional gamesters travelled from town to town with dice,
tied in a cloth, and play-mats, with a cruciform board painted on it, under their arms.
The dice were four large, black beans, called patolli, marked with white dots. The early
descriptions of the game unfortunately are unclear and confusing. The ancient Mexicans
apparently played various games of chance. But we only know the name of the most
famous game, patolli. Following Culin’s classification, Murray and Bell presented this
particular “game of the mat” as a race game (Culin 1898: 844; Murray 1952: 147 no.
6.7.1.; Bell 1960: 6). But, the earliest Spanish sources in fact referred to both war and
race games, when they tried to compare patolli with “alquerque”, or “castro”, or “tablas
reales” (Sahagún 1981: VIII, c 10, p 300; Lopez de Gomara 1552: fol. 42).
Just when and where the game of patolli originated is not clear. The bean, also known
as the mescal bean, was found in archaeological sites in Texas and Northern Mexico and
is said to have been used in prehistorical divinatory cults, long time before the Aztecs
settled in the valley of Mexico in the fourteenth century. The Aztecs would have brought
the hallucinogenic patol beans from the north and named the game after the beans they
already used with oracles and divinations (Duverger 1978).
From archaeological sources it can be deduced that the patolli boards already
occurred in the Classic times, at least some ten centuries ago, in the Maya area as well as
in Central Mexico (Swezey and Bittmann 1983). The design of patolli boards varies
considerably, as is illustrated in figs. 2 to 9, and, from the ancient pictographical
manuscripts we infer that also stick dice were used (Codex Vindobonensis Mexicanus,
sheets 13 and 20). Moreover, the first board game Durán described in his Chapter XXII,
on Aztec games and gambling, appears to be a kind of war game played with cane dice
instead of patolli dice. He accounts as follows (Durán 1575-81[1967: 197]):
There was another game, which was that they made in a plaster floor little hollows
after the manner of the game called “fortuna”, and one person took ten stones and the
other ten stones, and the one put his stones on the one edge and the other on the other on
opposite sides, and with some reeds split down the middle, they cast them on the ground
so that they sprang up, and as many reeds as fell with the hollow side upwards so many
places he moved his stones forward, and thus one followed the other, and all the stones he
overtook he took away until he left the other without any and it happened that five or six
were taken and with the four that were left, he could tell the reeds to turn against the other
and he would still win the game.”
In the colonial period dice games, played with beans or with reeds or sticks, were
observed all over Mexico. In many cases the word Patol labelled stick-dice games too(9).
Culin mentions several of these dice or board games in his catalogue. Their variations
are more in the materials employed and the circuit than in the object or method of
play(10). Summarizing the data presented above this author complies with the use of the
name patolli as a generic term. As will become clear in the next paragraph, patolli labels
both race and war games of Mesoamerican origin.
5.2. In search of the Mesoamerican link
The board games still extant in Mexico, which were studied for this comparative
presentation, are the Nahua petol, played in Puebla (Caso 1924-27: 203-211), the
Purehpecha kolitza or kuiliche, played in Michoacan (Beals and Carrasco 1944: 516-22;
Soto Bravo 1992), and the Chinantec los palos, played in Oaxaca (Weitlaner and Castro
1973: 189, 191). They will not be presented here at length. Only the common
characteristics with bul will be highlighted. An important contribution to the study of
patolli was paid by the Mexican ethno-historian Caso. Seventy years ago he discovered
in the Mexican state of Puebla a race game called petol, that he considered to be a
regional variant of the famous ancient game (Caso 1924-27: 203-211). The Nahua-
speaking descendants of the Aztecs now use four short stick-dice, made of split reeds,
two of which are marked differently with crossed lines in their hollow insides (fig. 11).
Caso also refers to Durán to prove the ancient origin of the stick-dice. This set of cane
dice resembles the various North American sets described by Culin (1907), not only in
their markings but also in the throwing and scoring method. However, the resemblance
of this Nahua petol game with other Mesoamerican games is quite striking and offers good
evidence for a common origin or development within the Mesoamerican culture area.
First of all there is the use of four two-
sided lots [1], which corresponds with
the number and characteristics of the
patol beans used in the ancient times.
As to the cane dice, it appears that in
the modern games also the hollow
sides indicate the score, just as in the
game described by Durán, above in
5.1. (Durán 1575-81[1967:197]). We
may assume that also the pre-Hispanic
scoring method has survived: as a rule
every marked side counts one [2].
Fig: 11: Petol
cane dice
(Caso, 1924-
27, fig. 5)
Fig. 12 (right):
Petol game
board (Caso,
1924-27, fig.
Fig. 13: Kolítza
scoring method
(Beals and
Carrasco, 1944
fig. 5)