The prayer fell silent – but the cross remained

Tófa and Hemingr had this stone raised in memory of Gunnarr, their son. He died abroad among the Greeks. May God and God's mother help his soul.

So reads a runestone at Åshusby in Uppland, carved some time between 1020 and 1050 to judge by its ornament and attributed to the carver Åsmundr. The last sentence is a prayer, and it is not unusual: about one in five runestones with a legible text asks God, Christ or God's mother to help the dead. But when we followed the prayer through time we saw something that surprised us. The prayer comes and goes – while the cross stays.

Enlarge ⤢

The prayer comes – and goes

We went through the 1,292 runestones in Sweden that have a legible text in the Scandinavian Runic-text Data Base and looked for prayers: God help his soul, God help his spirit, God's mother and their many spellings. 804 of the stones can be roughly dated by their ornament. Runestone art changed style during the eleventh century, and scholars have put the styles in order, from RAK around the year 1000 to Pr5 in the early 1100s.

The prayer becomes more common in the first half of the eleventh century and peaks with the style group Pr2, around 1035, when 37 % of the stones carry a prayer. Then it falls steeply: 10 % in Pr4 and 3 % in Pr5, among the last runestones. Our model, which takes into account who carved the stone and where it stands, puts the peak around 1038 (between 1023 and 1048) and shows a decline from about 30 to 2 per cent in the early 1100s.

It was not the faith that disappeared

One might think that the prayer disappears because the stones become less Christian. The cross speaks against that. It is found on 46 % of the earliest stones, on 70 % around 1035 and still on 62 % in Pr4 and Pr5, long after the prayer has begun to fall silent. With the same model as for the prayer the cross does decline after its peak around 1028, from about 70 to 40 per cent, but that decline does not hold in both halves of the material. The prayer falls to a fifteenth; the cross remains on a large share of the stones. It is the prayer, the plea to God to help the dead, that almost disappears.

At the same time something else happens: the carvers begin to sign. The carver's name appears on 3 % of the earliest stones and on 38 % of the last. Œpir carved, says the stone at Morby in Uppland, where a mother had a bridge built for her daughter's soul – without a prayer.

Enlarge ⤢

Our hypothesis: the prayer moved into the church

That runestones carry prayers is well known, and scholars have long seen the runestones as part of the conversion. What we add is the measurement: the rise and fall of the prayer can be followed with carver and area held constant, and it differs from the cross, which remains. That calls for an explanation.

Our hypothesis is that the prayer on the stone was the laity's own intercession during the mission period. In the first half of the eleventh century there were still few churches and priests in the Mälaren valley. Whoever wanted someone to pray for the soul of the dead could put the prayer where it was seen – on the stone by the road, perhaps in the hope that passers-by would pray. When the church grew in the second half of the century, with priests, consecrated ground and masses for the dead, intercession became the church's task. The stone no longer needed to pray. It went on carrying the cross, as a sign of faith and family, and increasingly became a memorial – signed by its carver.

A first test: a shift at the same time everywhere

If it was the local church that made the stone's prayer superfluous, the prayer should fall silent earlier where churches came early. That we could test. We took the medieval churches in Wikidata built by the year 1200 and compared stones near and far from them, using only stones still standing where they were raised (504 dated stones).

The result was clear, but not what we expected. The prayer is somewhat more common near the early churches, but it falls silent just as fast everywhere: after about 1075, 4 % of the stones within three kilometres of an early church carry a prayer (25 stones), and 3 % of those further away (118 stones). We see no sign that the prayer disappeared earlier near the churches (p 0.28 for the difference in how fast it disappears).

That speaks against the simplest form of the hypothesis – that each farm stopped praying on its stone when it got its church – and for a common shift that took hold at about the same time throughout the Mälaren valley. Such a shift could come from above: in the second half of the eleventh century the church began to be organised in dioceses, and Sigtuna is mentioned as a bishop's seat around 1060. We therefore state the hypothesis more sharply:

The prayer fell silent when intercession for the dead became a matter for the organised church, not for the individual family.

The test has a limit: the style groups divide time into steps of about twenty-five years, so a difference of a decade or so would not show.

How the hypothesis can be tested further

1. Finer dating. If the carvers' careers are dated more finely, through their collaborations and chains of masters and pupils, it can be seen whether the prayer falls silent at the same time everywhere or whether the shift spreads from the centres of the dioceses.
2. The words of the prayer. If the church took over intercession, the wording of the prayers should approach the church's language before they disappear. God's mother is named in 27 % of the prayers in Pr2 and 25 % in Pr3, against 12 % in RAK and 5 % in Pr4 – a lead to follow, even though the number of prayers per style group is small.
3. Other provinces and countries. Where the church was organised earlier or later, the prayer should have fallen silent earlier or later. A comparison with Denmark needs better records of the crosses on the Danish stones than we have today.
4. Gifts for the soul. If intercession moved into the church, gifts for the soul of the dead – to the church or as bridges – may have taken the prayer's place. Bridge building does not clearly increase in our material, and gifts to the church do not show on the stones; other sources are needed here.

Caveats

• The dating rests on the ornament and is approximate, and the style groups were worked out for the stones of Uppland and Södermanland.
• Prayers are found by searching the normalised text in Rundata. A damaged text can hide a prayer; we therefore tested stones with a preserved end on their own.
• An association in time is not a cause. The hypothesis explains the pattern, but other explanations are possible – for example that a new generation of carvers had other ideals. That the pattern holds with the carver held constant speaks against it being merely a change of generation.
• While working on this we also found an error in how the app leaves out stones that have been moved to churches. It does not affect the main finding, which does not depend on where the stones stand, but the test against the early churches uses the corrected selection. The error has been fixed in the app and is described in the methods.

See for yourself

Enlarge ⤢

Enlarge ⤢

Enlarge ⤢

The maps, curves and models come from Vitki, an open tool for runestone research that you can download and run on your own computer from GitHub. The method is described in METHODS.md.

Sources and data

• Scandinavian Runic-text Data Base (Samnordisk runtextdatabas), Department of Scandinavian Languages, Uppsala University. Open Database License.
• Gräslund, A.-S. 2006. Dating the Swedish Viking-Age rune stones on stylistic grounds. In: M. Stoklund et al. (eds), Runes and their Secrets: Studies in Runology. Copenhagen: Museum Tusculanum Press, pp. 117–139.
• Sawyer, B. 2000. The Viking-Age Rune-Stones: Custom and Commemoration in Early Medieval Scandinavia. Oxford: Oxford University Press.
• Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics 6, pp. 65–70.
• Wood, S. N. 2017. Generalized Additive Models: An Introduction with R. 2nd edn. Boca Raton: CRC Press.
• Medieval churches: Wikidata (CC0). Photos: Wikimedia Commons, with photographer and licence in the captions.