A Double-sided Hnefatafl Board from Cathedral Hill, Downpatrick:
Time Consumed in an early Medieval Monastic Enclosure
By Mark Hall

Excavations on Cathedral Hill, Downpatrick, Co Down, took place in 1985-87 (post- excavation work is on-going but the interim summary can be found in Brannon 1988, 61-4). The excavations were prompted by the need to extend the Cathedral's burial ground on the south- west slope of the hill, some 100m away from the Cathedral. The present Cathedral is all that remains of the 12th century Benedictine monastery, itself a re-building on the site of an Early Christian monastic centre. Archaeological evidence for the occupation of the hill dates back to at least the Bronze Age (ibid, 62). The finds from the 1986 season include a graffiti-incised stone gaming board. The implication of the summary report is that the graffiti board was one of the finds recovered from the Early Christian period cemetery (ibid, 62 & 63-4), though its site context is that of 12th century dump material (N. Brown, pers. comm.). Re-use or deposition of graffiti gaming boards in a burial context is not unusual, as work at Tintagel Parish Church (Cornwall) has shown. There later medieval burials (dated broadly to the 12th -14th centuries) included slate- incised gaming boards (of the merels class of game - see Parlett 1999,109-23) re-used as burial covers (Nowakowski & Thomas 1992, 22 & 24-6).

The double-sided hnefatafl board from Downpatrick Cathedral

The discussion of the Downpatrick find will seek to demonstrate that it is consistent with other boards of the tafl group of board games, broadly dateable to the 9th -13th centuries.

Description and Dating
The board is double-sided and graffiti incised into a block of slate measuring 180x220x25mm. It was recovered during the 1986 season of the excavations and subsequently numbered DPAT 86 277 and 86/583. It is now in the collections of Down County Museum. The board has previously been illustrated in Rankin (1997, 44) where it is described as a 12th century gaming board for an unknown game. The graffiti boards fill the available space on each side of the block and do so deliberately. The first board (Illus 1) measures 160x130mm and is laid out as a grid of 7x7 lines (which can also be read as a grid of 36 irregular cells). The central crossing-point is demarcated with an irregular circle and each of the four corner cells has its outer-angle demarcated by a graffiti arc. The second board measures 113x110mm and is laid out as a grid of 8x8 lines (which can also be read as a grid of 49 irregular cells). The centre point appears to be
marked with a cross and the corner points have no obvious demarcations. I am grateful to Mike King for the description of this board and his observation that the profusion of incomplete lines on this side make it look as if it was a failed attempt to execute the design, replaced by the more readily useable version on the other side.

The dating of the Downpatrick board is problematical. The excavator, Nick Brannon has (pers. comm.) described it as coming from no meaningful context (that is, dumped) on the chronological divide between the late Early Christian period and medieval strata. This gives a 12th century date for the dumping but perhaps suggests a pre 12th century date for the board itself with subsequent dumping after a period of use and possibly reuse or abandonment. The following discussion supports a 10th -12th century date.

Hnefatafl: the Game

The design of the Downpatrick double-sided board is entirely consistent with the known forms of the game hnefatafl. This is a variant within the tafl group of games, which involves a contest between two unequal forces. It can be played on the lines (or interstices) of a board (as with the Downpatrick example) in the cells or in points which receive pegged pieces. Board sizes are known of 7x7, 9x9, 13x13 and 18x18 cells/lines or points. The principal, or king-piece, (the "hnefi") of the defending side occupies the central cell or intersection, where his defenders surround him. His objective is to reach one of the four corner cells and so secure victory. The attacking force is arranged along the edges of the board and has to try and capture the king-piece by surrounding it on four sides. All the pieces move orthogonally, as the rook piece in chess (Parlett 1999, 196-204; Murray 1952, 56-64). The more useable 7x7 board from Downpatrick would probably have been played using a "king" and eight defenders against 16 attackers arranged in groups of four along each side, though it is also feasible that there were 20 attackers similarly arranged.

Hnefatafl: origins, distribution and dating

Hnefatafl (literally "king's table") is generally held to be of Scandinavian origin. The earliest archaeological evidence for the game includes a board fragment dated to the 5th century AD from a grave at Wimose, Funen (Murray 1952, 58) and it seems to have been carried by the Vikings to all the countries they raided and settled. The distribution of boards and playing pieces from Scandinavia to the British Isles and from Scandinavia to the Ukraine certainly supports this idea. However, there are indications that the game, or a variation of it, was known in the North Atlantic Littoral before the Viking Age (which is not to say that the ancestors of the Vikings were not involved in its spread). One of the finest surviving boards is that from Ballinderry Crannog, CO Westmeath, Ireland (Kendrick 1933; Hencken 1936). It is a complete board made from yew wood
and probably made in Dublin. It is a peg-board design of 49 holes (equivalent to a 7x7 grid) with clearly demarcated centre and corner positions. It has an elaborately carved interlace border and two carved handles with faces. A slightly less elaborate, incomplete board, now lost, from Knockanboy, Derrykeighan, Co. Antrim, survives as a drawing in the Ordnance Survey Memoirs
of County Antrim (Simpson 1972, 63-4, pi VII). Both are clearly high status pieces, in contrast to the much simpler and fragmentary example of mid 12th century date from Waterford (O'Keefe 2000112 & fig 49, after Hurley et al 1997) or the graffiti boards from Downpatrick. Further examples of Scandinavian boards (and pieces) can be found in an increasing literature on the subject (see for example Owen & Dalland 1999,128-32; Rosedahl & Wilson (eds) 1992, cat nos 71.321,342,360,572).

A note of caution was sounded earlier in ascribing the spread of the game solely to Viking influence. There are a number of boards from Scottish sites in particular that suggest that the game was known prior to the arrival of the Vikings. Pictish phases at Buckquoy, Howe and Birsay,in the Northern Isles, have produced graffiti hnefatafl boards, incised on stone, and there is asurface find of a board from Dun Chonallaich, near Kilmartin, Argyll which is presumed to be Dalriadic (Ritchie 1987, 61-2). The Northern Isles examples could be indicative of cultural contact with early Viking arrivals. Magnusson (1990, 2) has commented on the possible peaceful initial contacts between Picts and Vikings a century or so before the Lindesfarne raid in 793 AD (see also Graham-Campbell and Batey 1998, 23 and Crawford 1987, 39-47). The broch site at Scalloway, Shetland has also produced gaming pieces that could have been used in a tafl game (Wilson & Watson 1998, 174-5, where the authors are reluctant to accept a tafl identification because of the pre-Viking dating). Evidence from Anglo-Saxon England - before the Viking invasions - includes a large number of playing pieces, some of which were undoubtedly used for tafl games (the fullest review is Youngs 1983, 853-73 and see Hall 2001 for a summary). There was clearly an Anglo-Irish connection, for by the early 10th century it was known in the elite Christian circles of both countries. The manuscript known as Corpus Christi College Oxon.122 is an 11th century copy of a 10th century Irish Gospel book. It includes (f.5b) a drawing of a layout for hnefatafl, heavily disguised as an allegory for the harmony of the Gospels. There it is called Alea Evangelii - Game of the Gospel/Evangelists - and the title caption notes that the game was
brought to Ireland by Dubinsi, Bishop of Bangor (d. 953), from the court of King Athelstan (r. 925- 40). (For a discussion of the document see Robinson 1923, 69-71, 171-81 and the frontispiece illustration; for a discussion of the game see Murray 1952, 61-2).

A further graffiti board is known from Whithorn, Dumfries & Galloway, Scotland (Nicholson 1997, 449) though this is contextually dated to the 13th century or later. Nicholson (op. cit.) follows Ritchie (1987) in suggesting that the game in question is an Irish equivalent to hnefatafl, which they equate with the Irish name "brandubh". It is certainly true that there was cultural/linguistic variation in the naming of similar games - the Welsh name for hnefatafl is "tawlbwrdd". But there is also some degree of fluidity in the way the names were applied. In the
earliest written forms (12th century and later) of Irish and Welsh legends and romances (including the Ulster Cycle and the Mabinogion) the names "fidcheall" and "gwyddbwyll" (a predecessor of tawlbwrdd) are used to refer to chess and chessboards. The Gaelic/ British terms seem to have been so broad, i.e. referring to board games rather than specific board games, that they could be re-applied to whatever game was popular at a given time (a question that is also discussed in relation to the gaming material from Finlaggan, Islay, Scotland in Hall forthcoming). In the later medieval period, chess certainly superceded hnefatafl (Ritchie 1987, 62; Parlett 1999, 202-3). In the later medieval translations of Arthurian legends into Dutch manuscripts all gaming references are to chess rather than to any earlier games of the original tales (Meuwese 1996, 152-3). This evidence then could indicate a tafl game known in the British Isles before the Vikings (though equally the names discussed above could have applied to something markedly different that was displaced by the popularity of hnefatafl, which was then given the older name(s)).

On balance then a case can be made for tafl games being played in the British Atlantic zone before the Viking Diaspora. Indeed it is likely that there were several variations, possibly interchangeable and evolving through contact via the Atlantic seaways - some of the pegged boards would be ideal for play on board ship - a clear implication of Ritchie's paper in establishing the commonality of the game amongst Irish, Picts and Vikings.

The Downpatrick example (if we assume it was not simply a failed board followed by a more usable executed design on the other side) is a double-sided board and this is a common format, particularly for graffiti boards (usually wooden examples). The wooden boards from the 9/ 10th century Gokstad ship-burial; from Trondheim (two, both dated to the late 11th-mid 12th century) and from Toftanes, Faroes are all carved with both hnefatafl and merels and it has been suggested that the Vikings were the first to combine two games on one board. In addition to the archaeological evidence for this, there are written sources that describe such double-boards. The late 13th century Kroka-Refs Saga describes a double-board for hnefatafl and chess (made of walrus ivory and sent as a gift from Greenland to Harald Hardrada) (McLees 990, 81; Rosedahl &Wilson 1992, cat 321, for the Toftanes board). The date of this saga is compatible with that of the graffiti board from Whithorn (already mentioned) and further demonstrates that continued medieval survival of the game alongside the seemingly more popular chess, backgammon and merels.

The author knows of no other double-sided boards from Ireland (but would be happy to here of them) where the other examples known are single-sided. The wooden boards from Ballinderry and Waterford have been mentioned above and a third example, similar to that from Ballinderry, was found in the 19th century at Knockanhoy, Derry Keighan, CO Antrim, but is now lost (Edwards 1990, 78). A further stone-incised graffiti board is known from Garryduff 1 (Edwards 1990, fig 45f) and like Downpatrick it is of 6x6 cells but has no corner cell demarcation. It is unlikely to have had the central point demarcated but this is uncertain because the central area is damaged. The focus of the preceding discussion has been on hnefatafl boards but the other key element of the archaeological evidence is the playing pieces. These are known made in wood, stone, bone, glass, ivory and amber and are often found as complete or near-complete sets from Viking graves in Scandinavia and elsewhere (Owen and Dalland 1999, 128, gives a useful summary in the context of the example from Sanday, Orkney). However, the simple graffiti board from Downpatrick is unlikely to have had purpose-made pieces other than of the most prosaic kind. Small stones or pebbles would have been readily adaptable and available for such a board, as would sherds of broken pottery (the types of pieces easily missed in an excavation).

Conclusion
The preceding discussion has outlined the nature of the double-sided board from Cathedral Hill, Downpatrick and identified it as for the game of hnefatafl. It is readily comparable to a number of other graffiti boards from the British Isles and Scandinavia. The other Irish sites that have produced hnefatafl boards are all secular, as are the majority of hnefatafl find-sites (and I include pagan burials here). The two notable Christian, ecclesiastical exceptions are Downpatrick and Whithorn Priory (a third may be Tintagel Parish Church, where one of the graffiti boards may be for a tafl game). Graffiti boards in general do have strong associations with church sites, be they parish churches, cathedrals or monastic houses so there is nothing untoward in the two hnefatafl examples. Whilst the monks of these establishments are not unlikely to have played at hnefatafl, as at other board games, the boards in question are presumably those of lay workers or visitors to the monasteries. The nature of the boards proclaims an expedient, ephemeral use of a readily available raw material. These were
presumably the possessions of someone of lower social or material status, whether monk or lay: they may not have been possessed as such but simply made use of and abandoned or reused (as with the Tintagel grave coverings) as necessary. The context of the Downpatrick board is of little assistance here: its dumping supports ephemeral, low status but does not permit a more detailed glimpse of use and possible reuse Its broad dating is consistent with other known examples. The board from Downpatrick is clearly not the property of a secular lord, a well-to-do merchant or craftsman or of an abbot or senior cleric, but it is no less deserving of our attention as a reminder of the human, cross-cultural desire to play.

Acknowledgements
I am indebted to Mike King, for bringing the board to my attention, and to Nick Brannon, for agreeing to its publication in advance of the final report on Downpatrick Cathedral.

References
Brannon, N 1988 'Life and Death at an Early Monastery - Cathedral Hill, Downpatrick, Co Down', in Hamlin, A & Lynn, C (eds.) 1988
Pieces of the Past: Archaeological Excavations by the Department of the Environment for Northern Ireland 1970-1986, 61-4, (Belfast).
Crawford, B 1987 Scandinavian Scotland, (= Scotland in the Early Middle Ages 2), (Leicester).
Edwards, N 1990 The Archaeology of Early Medieval Ireland, (London).
Graham-Campbell, J and Batey, C 1998 Vikings in Scotland An Archaeological Survey, (Edinburgh).
Hall, M A forthcoming in 2001 'On the possible Merels Board Incised on the Pre-Conquest Cross-Base at Addingham St Michaels,
Cumbria', in Trans Cumberland & Westmorland Antiquarian & Archaeol Society (2001).
Hall, M A forthcoming, 'Finlaggan at Play: the Gaming Equipment', in Caldwell, D forthcoming, Finlaggan final excavation report.
Hencken, H O'N 1936 'Ballinderry Crannog no. 1', in Proc Royal Irish Academy 43 (1936), 103-240.
Hurley, M F Scully, 0 M B and McCutcheon, S W J 1997 Late Viking Age and Medieval Waterford Excavations 1986-1992, (Waterford).
Kendrick, T D 1933 'A gaming board of the Viking period found in Ireland', in Acta Archaeol 4 (1933), 85.
Magnusson, M 1990 The Viking Road', 1-12 in Simpson, G G (ed) Scotland and Scandinavia 800-1800, (= University of Aberdeen Mackie
Monographs I), (Edinburgh).
McLees, C 1990 Games People Played: Gaming-pieces, Boards, and Die from Excavations in the Medieval Town of Trondheim Norway,
(= Meddelelser Nr 24), (Trondheim).
Meuwese, M 1996 -Arthurian Illuminations in Middle Dutch Manuscripts', 151-66 in Busby, K 1996 Word and Image in Arthurian Literature,
(New York and London).
Murray, H J R 1952 The History of Board Games other than Chess, (Oxford).
Nicholson, A 1997 'Gaming Boards', 449 in Hill, P 1997 Whithorn & St Ninian -The Excavation of a Monastic Town 1984-91, (Stroud).
Nowakowski, J and Thomas, C 1992 Grave News from Tintagel -An Account of a Second season of Archaeological Excavation at Tintagel
Churchyard, Cornwall, 1991, (Truro).
O'Keefe, T 2000 Medieval Ireland An Archaeology, (Stroud).
Owen, 0 and Dalland, M 1999 Scar A Viking Boat Burial on Sanday, Orkney, (East Linton).
Parlett, D 1999 The Oxford History of Board Games, (Oxford).
Rankin, J F 1997 The Church of St Patrick of Down, (Ulster Historical Foundation).
Ritchie, A 1987 The Picto-Scottish Interface in Material Culture', 59-67 in Small, A (ed) 1987 The Picts, A New Look at Old Problems,
(Dundee).
Robinson, J A 1923 The Times of St Dunstan, (Oxford).
Rosedahl, E and Wilson, D (eds) 1992 From Viking To Crusader - Scandinavia and Europe 800-1200, (Udevella).
Simpson, W G 1972 'A Gaming Board of Ballinderry-type from Knockanboy, Derrykeighan, Co. Antrim', in Ulster Journal of Archaeology
35 (1972), 63-4 & pi VII.
Wilson, G and Watson, P M 1998 The conical gaming pieces', 174-5 in Sharpies, N 1998 Scalloway -A Broch, Late Iron Age Settlement
and Medieval Cemetery in Shetland, (= Cardiff Studies in Archaeology / Oxbow Monograph 82), (Oxford).
Youngs, S M 1983 The Gaming-Pieces', 853-73 in Bruce-Mitford, R 1983 The Sutton Hoo Ship-Burial, volume 3ii (ed. A Care-Evans),
(London).