Early Medieval Board Games: Hnefatafl

This is the first in a series looking a Viking-Age board games.

Hnefatafl is classified as a war game by board game scholars. There are variants of this game, which often fall under the category of ‘tafl’ games. It is a game that is played with unequal sides or forces: the smaller force (the defenders) has a king piece whose aim is to escape to a corner square of the lattice board. The larger forces (attackers) task is to capture the king to stop it from escaping. The rules were not written down contemporarily with the origins of the game, however this it is most likely the game that was played in its most basic rule set in Scandinavia, Britain and Ireland. Early 20th Century scholar Harold Murray recognised a game recorded by 18th Century Swedish botanist Carl Linnaeus called tablut as a ‘tafl’ game.

Linnaeus’s drawing of tablut (1811)

Linnaeus noted tablut in his diary in July 1732 during a tour of Lapland, he commented that it superficially resembled chess, played on a board of 9×9 cells with two forces: the white or Swedish with eight defenders and a king piece, and the dark opposing side or Muscovites with sixteen attackers. The aims are as described above, each piece moves the same, orthogonally, and all are captured by being sandwiched between two opposing pieces in the same row or column.

It is not disputed by academics that tablut is hnefatafl because it matches both archaeological and literary evidence. The sagas give references to playing board games. Most do not describe the rules, as this was not the purpose of the story, it was assumed people knew the rules already. If a character in a story written today is playing chess the rules are not described because the assumption is made that the reader knows how to play.

Hervarar saga contains two riddles that relate to gaming. The riddle goes :

‘Who are the maidens who fight around their defenceless lord? The darker ones defend all the time, the fairer ones advance’.

The second riddle goes:

‘What is that beast all girdled with iron which kills the flocks? It has eight horns but no head’

The first riddle’s answer is hnefatafl, the second answer is the hnefi (the king piece is often referred to at the hnefi, although it means ‘fist’ in old Norse). These riddles share elements of the tablut game described by Linnaeus with two sides, one involving a king piece, one side attacks and the other defends. Fridthjof’s Saga gives reference to a double-attack being possible, which supports tablut, as a double attack suggests to surround a piece on two sides to take it out of the game. The archaeological finds of gaming pieces particularly from Scandinavia support Murrays claim as elaborate hnefi pieces are found alongside game pieces that are designed for uneven sides.

For example the image below shows the game pieces from Birka grave 523 has one anthropomorphic king piece, fourteen decorated with spirals and five plain dark glass.

Blue glass gaming pieces, including anthropomorphic king piece discovered in Birka grave 523.

There is evidence for different sized variations of hnefatafl throughout Britain and Ireland. In Ireland the Gaelic terms Brandubh and Fidcheall are mentioned. Brandubh means black raven and could signify the shape of the king piece, although no archaeological evidence supports this so far. Fidcheall means ‘wood sense’ and could be the equivalent to the term for table being used as the name of the game in old English. The old Irish poem Scela Cano mac Gartnain tells of King Cano and his retinue sailing to Ireland in AD 668 with ‘fifty well armed warriors, fifty well-dressed ladies and fifty liveried gillies each with the silver leads of two greyhounds in his right hand, a musical instrument in his left and a fidcheall board on his back along with the silver and gold playing men.’

The Ballinderry board found in Ballinderry, Co. Westmeath, Ireland in 1932, supports this. It was found broken in two parts 8ft apart and has been dated to the 10th century. What is interesting about this board is that it is built with the intention to be travelled with. It is a pegged board so the pieces stay in place when played on a rocking boat or carriage. It also has two decorated projecting heads one human one animal that can be used as handles again in conjunction with unstable terrain. Both fidcheall and brandubh are listed in Irish law of the 7th and 8th centuries.

The Ballinderry board

The game tawlbwrdd occurs frequently in ancient laws of Wales, however, it was not described until 1587 by Robert ap Ifan. He provides a drawing of an 11×11 board and includes a description of the rules that are very similar to Linnaeus’ tablut. Ifan, includes extra rules, including moving one of your pieces between two of the oppositions pieces and not being captured, you have to say ‘I am your liegeman’ for your piece to safely move into that space though!

Tawlbwrdd means throw board and could suggest the use of dice with the game, however there is limited evidence of dice in conjunction with a ‘tafl’ game. One example was found at Keythorpe Hall in Leicestershire where two dice were excavated with forty-six bone playing pieces. The majority of gaming sets that could be a ‘tafl’ game include no dice. This does not mean dice were not used as there have been games known to use dice that have been found archaeologically missing their dice, such as the Gloucester Tables Set, a late 11th century backgammon board.

The etymology of the name tawlbwrdd may possibly answer the question about dice. Tawlbwrdd is possibly a misunderstanding for the name of the original game by borrowing and confusing taefl in old English and tafl in old Norse meaning board with the welsh tawl. Thus making tawlbwrdd mean board-board and therefore not indicating the use of dice within the game. Ifan’s account supports this, he states that when a piece is captured it is thrown from the board ‘ai daflu or gwarau’ [and he is thrown from the game]. Taflu is the lenited or softening form of daflu .

Reproduction bone 6-sided dice with ring-and-dot scoring decoration

The Anglo-Saxon evidence for ‘tafl’ games comes from an Irish gospel manuscript of the 11th century and can be seen below. It is played on an 18×18 lattice board with forty-eight attackers and twenty-four defenders and is called alea evangelii or ‘game of gospels.’ It was known in English court during the reign of Athelstan AD 925-39. Here the game has taken on Christian symbolism: there are seventy-two men because the number of items in the harmony is seventy-two and it is played on an 18×18 board because, four evangelists, four gospels and ten canons equals eighteen.

Corpus Christi College Manuscript 122

The earliest date for hnefatafl is from evidence of the 5th century AD board fragment found in a grave from Wimose, Denmark. Murray states that it was played by Scandinavian people from 400 AD and brought by the Norsemen to Iceland, Britain and Ireland, where it then spread to Wales. However, there is limited evidence of hnefatafl being played earlier. This evidence is that of conical playing pieces that were not used in the Roman period except for one exception, dating from 1st to 4th centuries AD found amongst an assemblage of Roman gaming pieces in Spain. An excavation at Cnip has a conical piece that dates from first century BC to the first century AD. Pieces from the Shetland isles have also been found, however, because they are all pre-Viking in date (before 6th century) a hnefatafl identification has been avoided.