Why do the runes look the way they do?

The runes are straight strokes, almost without horizontal lines, and in the Viking Age 16 runes had to serve a language with many more sounds. We follow the runes from the first finds around AD 160 to the dots of the eleventh century – and measure, across the whole Swedish runestone material, what it cost when the rune row shrank.

By
Digital analyst and founder of Vitki
LinkedIn ↗
siriþr : kiarþi : bur : þosi

So begins the Ramsund carving in Södermanland, the rock face with Sigurd the dragon-slayer. Rundata normalizes it as Sigriðr gærði bro þasi, ‘Sigríðr made this bridge’. Anyone who reads the runes letter by letter stumbles at once. The word kiarþi was pronounced roughly gærði: the k-rune stands for a g, the i-rune for an æ and þ for a ð. In bur the u-rune stands for an o.

That is not because the carver was careless. It is because of how the runes were made. This article follows the runes from the beginning: where they came from, why they are made of straight strokes, and why the Vikings wrote with only 16 signs. And we measure something we have not seen measured before: how much each rune had to carry.

The first runes

The oldest certain runic inscriptions date from the second half of the second century AD. A bone comb from the bog at Vimose on Funen carries the word harja and is archaeologically dated to around 160. The spearhead from Øvre Stabu in Norway lay in a grave from 175–200 (Spurkland, Futhark 1, 2010, p. 77). The runes are thus almost two thousand years old, and they first appear in southern Scandinavia, on portable objects.

Vimosekammen. Photo: Nationalmuseet, Roberto Fortuna og Kira Ursem, CC BY-SA 3.0, via Wikimedia Commons. The comb from Vimose, around AD 160. The runes read harja.

Where they came from is debated, but one explanation has the strongest support. Terje Spurkland writes: “The view of the matter that seems to have the strongest support is 'the Latin theory'” (Futhark 1, 2010, p. 65) – that is, the runes are based on the Latin alphabet. Several runes resemble Latin letters: f, r, s, b and m. In the twentieth century Greek and North Italic models were proposed too, and Gad Rausing lists all three theories (Fornvännen 1992, p. 200). Jonas Carlquist holds it very likely that contact with literate Rome was decisive (Futhark 13, 2024, p. 40). Who made the rune row, and where, nobody knows.

Straight strokes

The first thing one sees is that the runes are made of straight strokes. There are almost no curves and almost no horizontal lines. Where Latin has a horizontal stroke, as in F, H, L and T, the rune has a slanting one.

The usual explanation is wood. Staffan Fridell writes that when you carve runes in wood with a knife it is much easier to cut towards yourself than away from yourself, and that horizontal cuts are troublesome (Futhark 6, 2015, p. 12). A horizontal cut follows the grain and shows poorly. A slanting or vertical cut goes across it.

But it is an explanation without direct proof. Wood from the runes' first centuries is almost never preserved, and the oldest finds are of metal and bone. Spurkland notes that those who argue the wood theory have had to rely on an argumentum ex silencio, an argument from silence (p. 74). Fridell has since refined the picture: the slanting strokes suit wood, but curves turning into angles are better explained by metal – it is easier to carve angles than curves in metal (Futhark 14–15, 2025, p. 205). The shape of the runes may thus be a legacy of both the knife in wood and the point in metal.

Stone has no grain, but the shapes came along. When carving in stone, Fridell argues, people imitated the rune forms that had developed in, and been shaped by, carving in wood (p. 8). This is what the straight strokes look like when cut in limestone, on a grave slab from the end of the Viking Age. Vitki's groove analysis has found the grooves in the scan and measured them: 1,172 cross-sections in 97 runes. The grooves are on average 3.7 millimetres deep and 13 millimetres wide. Every stave and every side-stroke is a straight cut of its own.

Image: Vitki, detail from the groove analysis of the grave slab Ög 240 in Hov church, Östergötland, turned a quarter so that the row lies horizontally. Orange: carved grooves found by the analysis; dots: measured cross-sections along the stave and across the side-strokes. Scanned by Laila Kitzler Åhfeldt in 2008 (ATOS II) in Cecilia Ljung's doctoral project on early Christian grave monuments: Kitzler Åhfeldt, L. 2025, 3D-data Tidigkristna gravmonument Östergötland och Småland, doi:10.5281/zenodo.17571537, CC BY 4.0.

The rune row

The runes were ordered in a fixed sequence, the futhark, after the first six: f, u, þ, a, r, k. The older futhark had 24 runes in three groups. The oldest complete rune row is on the Kylver stone from Gotland, usually dated to around AD 400. Whether it was meant to work as magic or was an aid to memory is debated: Bernard Mees describes the traditional magical reading (Futhark 7, 2016, p. 15), Carlquist thinks an aid to memory more likely (Futhark 13, 2024, p. 9).

Kylverstenen. Photo: Helena Bonnevier, CC BY-SA 4.0, via Wikimedia Commons. The Kylver stone from Gotland, with the whole older rune row.

The older futhark was used throughout the Germanic world. In England it gained new runes for new vowels and became the Anglo-Saxon futhorc. In Scandinavia things went the other way.

From 24 to 16 runes

Around 700 the Scandinavians changed their rune row. Hanna Åkerström sums it up: around 700 there was a transition from the Proto-Norse 24-rune row to the Viking Age's 16-rune row (Futhark 8, 2019, p. 8). The surest find with the new row is a skull fragment from Ribe with an inscription from around 725–750.

The strange thing is that the row shrank while the language gained more sounds. Runes for g, d, p, e and o disappeared. Why is an open question. Kurt Braunmüller sees a deliberate reform, Michael Schulte a gradual development, and Michael Barnes doubts that the question can be settled: “I suspect lack of data from the period of transition between the older and younger fuþark would make it difficult to judge which theory, if either, had the greater explanatory power” (Futhark 4, 2013, p. 12).

What the shrinking cost

How much, then, did each rune have to carry? We have measured it. We set each rune on the Swedish Viking-age runestones beside Rundata's normalization of the same word – the editors' reading of how the words sounded. That came to 6,424 words and more than 20,973 runes. We did the same with the Proto-Norse inscriptions, written with the older futhark.

How often a rune stands for another sound than its own

the k-rune stands for g

Younger futhark15.9 %of 932 runes
Older futhark2.6 %of 39 runes

the t-rune stands for d

Younger futhark7.9 %of 2,217 runes
Older futhark0 %of 32 runes

the u-rune stands for o

Younger futhark21.4 %of 1,619 runes
Older futhark0 %of 72 runes

the a-rune stands for æ

Younger futhark21.9 %of 2,880 runes
Older futhark0.7 %of 153 runes

All runes except þ: another sound than the rune's own

Younger futhark12.6 %of 20,973 runes
Older futhark4 %of 695 runes

The scale runs from 0 to 30 %. Source: Scandinavian Runic-text Database (Rundata). Younger futhark: Swedish Viking-age runestones; older futhark: Rundata's Proto-Norse inscriptions. Each rune compared with Rundata's normalization, word by word. Percent. Chart: Vitki, after Rundata (ODbL).

With the older futhark a rune almost always stood for its own sound: 4.0 per cent of the runes stand for another sound than the rune's own, about one in twenty-five. With the younger it is 12.6 per cent, about one in eight. The k-rune stands for a g one time in six, the b-rune for a p one time in four, the u-rune for an o one time in five and the a-rune for an æ almost as often. In the older futhark g, d, p and o had runes of their own, and there the k-rune stands for g only 3 times in a hundred.

For the reader this means that every runic text is a small riddle. kiarþi can only be read as gærði if one already knows which word it ought to be. It is like writing English without g, d and p, and trusting the reader to understand.

Long-branch, short-twig and staveless

The younger futhark came in several forms. Magnus Källström writes that what we now call long-branch runes seem to have dominated in Denmark in the first two centuries of the Viking Age, while the short-twig runes are typical of Sweden and Norway (Futhark 4, 2013, p. 106). The short-twig runes have shorter side-strokes and are quicker to carve. The Rök stone in Östergötland, from the early ninth century, is carved with them, and the same stone has a passage in the runes of the older futhark. That passage is generally seen as a cipher, write Fridell and Henrik Williams (Futhark 11, 2021, p. 169) – an older writing system that only a few could read.

Ög 136. Photo: L'imaginaire, CC0, via Wikimedia Commons.

Furthest went the staveless runes in Hälsingland, where the runes have lost their main staves and only the side-strokes remain. Fridell sees them not as a shorthand but as “a cryptic script and a writing experiment” (“en kryptisk skrift och ett skriftexperiment”, Futhark 6, 2015, p. 13). That is debated. In the late Viking Age the long-branch runes took over in Sweden too, and they are what we see on most runestones.

Hs 14. Photo: Bengt A. Lundberg, CC BY 2.5 se, via Wikimedia Commons. The Malsta stone in Hälsingland, carved with staveless runes.

The dots

The answer to the shrunken rune row came from within. Carvers began to put a dot, a prick, in a rune to show that it stood for another sound: a dotted i became e, a dotted k became g and a dotted u became y. According to Källström the dotted runes seem to have been established at the end of the tenth century (p. 114), and in the Viking Age it was mainly e, g and y that were used (p. 116). The dots thus won back part of what was lost around 700.

In the Middle Ages the system was expanded into a rune row for all the sounds of Latin, but at different speeds. In Västergötland the medieval system seems to have been established in the later twelfth century, while in Uppland it probably took until after 1200 (Källström, Futhark 6, 2015, p. 135). On Gotland and in Dalarna the runes lived on very long, in Dalarna into the nineteenth century.

The weaknesses

  • The normalization is an interpretation. Rundata's normalization is the editors' judgement of how the words sounded, not an independent measure of the sounds. We have compared the runes with it.
  • Only some of the words can be paired. We compared only sections where the runic text and the normalization have the same number of words, and without uncertain runes and personal names: 6,424 words. The older futhark rests on 217 words from 284 inscriptions, and the figures there are therefore less certain.
  • The þ-rune is left out of the summary. The normalization distinguishes þ and ð by a convention, not by the carver's choice.
  • We tried to measure when the dots caught on, style group by style group, but Rundata's normalization writes most e-sounds as æ, and the material became too thin to use.
  • Most of the story is known. The origin, the shapes and the transition are research results, with sources in the text. Our contribution is the measure of how much the runes had to carry.

This is the first part of our series on the runes, where each rune gets an article of its own in the order of the futhark, and next time we begin with fé, the first.

Sources

  • Scandinavian Runic-text Database (Rundata), Department of Scandinavian Languages, Uppsala University. Open Database License.
  • Spurkland, Terje 2010. The Older Fuþark and Roman Script Literacy. Futhark 1, pp. 65–84.
  • Rausing, Gad 1992. On the origin of the runes. Fornvännen 87, pp. 200–205. CC BY 4.0.
  • Carlquist, Jonas 2024. Runskrift på fibulor: Ett tidigt germanskt skriftbruk. Futhark 13, pp. 5–54. CC BY 4.0.
  • Fridell, Staffan 2015. Tendenser i skrifttecknens utveckling: Alfabet och runor. Futhark 6, pp. 7–19.
  • Fridell, Staffan 2025. De äldre runornas bildning. Futhark 14–15, pp. 201–206. CC BY 4.0.
  • Mees, Bernard 2016. The Hogganvik Inscription and Early Nordic Memorialisation. Futhark 7, pp. 7–28. CC BY 4.0.
  • Åkerström, Hanna 2019. Det tidigvikingatida runmaterialet: En inventering. Futhark 8, pp. 7–88. CC BY 4.0.
  • Barnes, Michael P. 2013. What Is Runology, and Where Does It Stand Today? Futhark 4, pp. 7–30.
  • Källström, Magnus 2013. Vikingatida och medeltida skrifttraditioner. Futhark 4, pp. 102–128.
  • Källström, Magnus 2015. Gravhällsfragmentet från Tornby. Futhark 6, pp. 107–142.
  • Fridell, Staffan and Henrik Williams 2021. De äldre runorna på Rökstenen. Futhark 11, pp. 167–171. CC BY 4.0.
  • Futhark 1–6 carry the licence CC BY-NC-ND; we quote briefly with the source. All volumes are freely available at futhark-journal.com.
  • The analysis plan (locked before the count) and the code are open on GitHub.
  • Photos: Wikimedia Commons, with photographer and licence in the captions. The groove image: Vitki, after the open scan named in the caption.
Share this article
Email
All articles →