Why does Norway have only 66 runestones – and Sweden 2,300?

Gunnvôr gerði brú, Þrýðríks dóttir, eptir Ástríði, dóttur sína. Sú var mær hônnurst á Haðalandi.

‘Gunnvor, Trydrik's daughter, made the bridge in memory of Astrid, her daughter. She was the handiest maiden in Hadeland.’

So reads the Dynna stone from Hadeland, carved in the eleventh century. A mother built a bridge for her daughter and wanted everyone who crossed it to know what the girl could do. Whoever reads it today has something unusual in front of them, and not only because the stone remembers a daughter. A runestone from Viking-age Norway is itself a rarity. The Scandinavian Runic-text Database, Rundata, has 66 Norwegian runestones from the Viking Age. From Sweden it has more than 2,300, and Uppland alone has 1,178. As early as 1989 Birgit Sawyer wrote that only about fifty were known from Norway (“Från norskt område är endast ca 50 kända”).

Enlarge ⤢

This article asks why. We did three things: read what research has already said, tested an old explanation on the whole Swedish material, and looked for what sets the Norwegian stones apart from the Swedish ones. The plan was locked before we counted.

No Norwegian had his stone raised

In Uppland the stones most often say that someone had the stone raised. litu risa stin at tuba, ‘had the stone raised in memory of Tobbi’, reads U 1177 at Harbo, and below the text the carver signed his name: Öpir. The little word lét, ‘had’, holds a division of labour: one who ordered and one who carved. We have written about how it sets Uppland apart from Västergötland in The man from Västergötland raised it himself – the man from Uppland had it raised.

On the Norwegian stones it is almost never there. Of the 66 stones, a single one has something Rundata translates as ‘had made’, and there, on a fragment from Bore in Rogaland, the decisive runes are reconstructed. Rune Palm saw this in 1992 in his thesis Runor och regionalitet: of 29 Norwegian inscriptions, none had the auxiliary verb – “this feature is not attested at all in Norway” (“Som framgår av tabell 37a är denna företeelse inte alls belagd i Norge”, p. 217).

Of the Swedish stones 29 per cent have lét, and among those Rundata dates to 1050 or earlier, the ones contemporary with the Norwegian stones, 24 per cent. A difference this large would arise by chance in fewer than one case in a hundred million – about twenty times less likely than the top prize on the Swedish Lotto with a single row. But the chart also shows that Norway is not alone. Västergötland is just as low, 3 stones of 131. The difference does not run between the realms but between west and east, from Norway and Västergötland to Uppland, where four stones in ten have lét.

Palm's explanation, tested

Palm had an explanation. It is “tempting”, he wrote, to link the auxiliary verb to the fact that it is in Uppland's runestone districts that the professional carvers seem to have had a quite special status; there one could entrust the making of the memorial to the professionals (p. 221, our translation). In the west, he thought, the patron himself was responsible for the stone: “In Denmark and Norway and in the more southerly provinces of Sweden, carvers are conspicuously absent from the inscriptions” (p. 227, our translation).

That is a hypothesis that can be tested, and as far as we know nobody has. We counted the carvers whom Rundata links to at least ten Swedish runestones. There are 22, with Öpir (112 stones), Balle (76), Åsmund (74) and Fot (68) at the top, and together they made about one Swedish runestone in four. Among stones where someone had the stone raised, 46 per cent were made by one of them; among the others, 18 per cent.

Uppland has both many lét and many professional carvers, and the younger stones have more of both. So we compared only stones in the same province and the same style group – roughly the same district and the same time – and pooled the comparisons (62 such groups). The link remains. When the text says lét, the odds that a professional carver made the stone are 1.6 times as high. Chance would give a link this clear about once in 2,600. Palm's guess holds.

In Norway there was no such trade to be seen. Rundata names no carver for any of the Norwegian runestones, and when a Norwegian stone itself names the one who carved it, it is a new name every time. At Gjerde in Hordaland the son did the work himself: Erlendr reist rúnar þessar eptir Ôlvi fôður, ‘Erlend carved these runes in memory of Ölver, his father’.

Why so few?

That the workshops were in the Mälaren valley does not on its own explain why Norway has so few stones, but it fits what research has already proposed. Palm again: runestones are a phenomenon tied to a landowning class, with their main areas next to the best soils, and the smallness of the Norwegian material must be linked to the topography and the comparatively small area of farmland (p. 255). Where demand was great a trade could arise: the need for memorial inscriptions around Lake Mälaren was so great that a group of people could specialise in making runestones (p. 256). And he points to power: Denmark and Norway came under the control of monarchs earlier than Sweden (p. 257).

Sawyer put the same thing from the other side: in Norway “this fashion never caught on as it did in Denmark and Sweden”, which suggests that raising stones filled functions that “in areas poor in runestones were unnecessary or were met in other ways” (1989, p. 189, our translation).

The Norwegian stones wear the time of kings and mission on their surface. The Kuli stone in Nordmøre says Tolf vetr hafði kristindómr verit í Nóregi, ‘Christianity had been twelve winters in Norway’, and one of the logs in the foundation of the bridge by the stone has been tree-ring dated to 1034, writes James Knirk (2017, p. 94). The Galteland stone in Agder remembers a son who died in the king's host: Sá varð dauðr í liði, þá's Knútr sótti England, ‘he died in the host when Knut attacked England’ – that is, in 1015–16.

Our own contribution to the question is a link, not a proof. Where the stones were many, in the Mälaren valley, they were often made by professional carvers for patrons who had them raised. In Norway there was neither the patron's formula nor the trade. Whether that was because there was no demand, or because land, kingship and church met the same needs in other ways, the numbers cannot tell.

The women who were remembered

What we did not look for, but found, concerns whom the stones remember. Among the 20 Norwegian stones where we can see how the dead was related to the raiser, 4 remember a woman. Dynna remembers a daughter. On two stones a man remembers his wife: Klepp in Rogaland, reisti stein þenna ept Ásgerði, kván sína, and Mariakirken in Stavanger, eptir Jórunni, konu sína. At Eik in Sokndal, Saxi built a bridge fyrir sálu móður sinnar, ‘for the soul of his mother’. In Sweden this is true of 33 stones of 1,067, 3 per cent, and among the early Swedish stones 2 per cent.

Chance would give a difference this large in about four cases in a thousand, and against the early Swedish stones in fewer than two in a thousand. But it is four stones. One stone more or less moves the share by five percentage points, and Öland is almost as high, 5 of 30. The pattern is clear enough to see, but not enough to explain.

It also stands in a strange contrast to who raised the stones. Sawyer showed in 1989 that women as sole raisers are somewhat rarer than average in Denmark and Norway (p. 190), and that men and women raising a stone together belong to Uppland, Öland and Södermanland but occur “in Norway never” (“i Norge aldrig”, pp. 191 f.). Norwegian women thus seldom raised stones, but they were remembered more often. We have not found this described before. Sawyer's book of 2000, with more tables, we have not been able to read, however.

Enlarge ⤢

The red stone from Ringerike

Now to the styles. Rundata uses Anne-Sofie Gräslund's style groups, made for the animal ornament of the Mälaren valley, and they fit Norway poorly: 33 of the Norwegian stones have only a plain rune band without an animal head, and 29 have no style at all. The few ornamented stones belong instead to a style named after a Norwegian district, Ringerike. Haakon Shetelig called it the Ringerike style, “best represented by stone monuments” (“best representert ved steinmonumenter”, 1950, p. 60), and he names the Dynna stone and its picture. Bertil Almgren placed the Vang stone in Valdres “in the front rank of the Ringerike style's many accomplished works” (1944, p. 88, our translation).

Enlarge ⤢

The style is quite literally named after a stone. According to Almgren, the ornamented Norwegian stones of the eleventh century are cut from “red sandstone, brought from Ringerike” – except the Vang stone itself, which is “a dense, grey mica schist” (p. 84, our translation). The Alstad stone from Toten says it in its own words. Jorunn raised it in memory of her husband ok fœrði af Hringaríki útan ór Ulfeyju, ‘and brought it out of Ringerike, from Ulvøya’, and the geologist V. Goldschmidt later showed that the stone's rock is found precisely on Ulvøya (Kleiber 1965, p. 63). A widow in the eleventh century thus carried her stone from Ringerike to Toten so that it would be of the right kind.

Enlarge ⤢

A Norwegian runestone in the later Urnes style, so common on the Uppland stones, we have not found.

What did not hold

We also expected the Norwegian stones to speak of land more often. Three do. At Nørstebø in Lesja, reistu stein þenna, þá er þeir skiptu lôndum sínum, ‘they raised this stone when they divided their lands’. At Oddernes, Eyvind built the church á óðali sínu, ‘on his allodial land’. And at Sele in Rogaland there is an agreement, Máli er á steini þessum, ‘the agreement is on this stone’, about something that is to belong to someone at eign ok at óðli, ‘as property and as allodial land’. But three of 66 against 30 of 2,321 in Sweden is not enough. Chance gives a difference like that in about six cases in a hundred, too often for us to call it a finding.

The weaknesses

The Norwegian stones are few, and many are fragments; when the pattern about women rests on four stones, a few new finds could change it. The Norwegian stones lack the Swedish style groups, so we could not hold time constant as we usually do; instead we compared with the Swedish stones Rundata dates to 1050 or earlier, and both findings held there. The features are counted from Rundata's English translations, which are made the same way in every country; the Norwegian hits we read one by one, the Swedish ones not. It was Palm who first saw that Norwegians did not have their stones raised – our contribution is to test his explanation and the finding about women. And the link between lét and the professional carvers shows that they go together, not that one caused the other.

Next time you stand by a runestone in Uppland, read the verb. If it says litu risa, ‘had raised’, someone paid a carver. If it simply says that someone raised the stone, they did as people nearly always did in Norway – where a widow at Toten carried her stone all the way from Ringerike.

Sources

• Scandinavian Runic-text Database (Rundata), Department of Scandinavian Languages, Uppsala University. Open Database License.
• Palm, Rune 1992. Runor och regionalitet. Studier av variation i de nordiska minnesinskrifterna. Runrön 7. Uppsala. Open on DiVA.
• Sawyer, Birgit 1989. Det vikingatida runstensresandet i Skandinavien. Scandia 55:2, pp. 185–202. Open at Scandia.
• Almgren, Bertil 1944. Vangstenen. Viking 8.
• Shetelig, Haakon 1950. Religionshistoriske drag fra vikingetidens stilhistorie. Viking 14.
• Kleiber, Boris 1965. Alstadsteinen i lyset av nye utgravninger ved Kiev. Viking 29.
• Knirk, James E. 2017. Kuli-steinen og landsnamnet Noreg. Namn og nemne 34, pp. 93 ff.
• The analysis plan (locked before the run) and the code are open on GitHub. On the formulas within Sweden: The man from Västergötland raised it himself. On women as raisers: She inherited from her child.
• Photos: Wikimedia Commons, with photographer and licence in the captions.