The Karlevi runestone

The Karlevi runestone (Fig. 1) stands close to the western shore of the island of Öland, off the south-east coast of Sweden, near to the village of Karlevi. Its runic inscription includes one whole stanza of skaldic verse, a highly cryptic and metrically elaborate form of early medieval Scandinavian praise or memorial poetry. The Karlevi stone is the oldest existing record of a complete skaldic stanza. Skaldic verse has otherwise come down to us in manuscripts written considerably later than the original composition, following a long period of oral transmission. The inscription is also the only known record of a complete skaldic verse in Viking Age runes, although there are some fragments of skaldic verse in runes on small objects. Indeed, Judith Jesch opens her analysis of the vocabulary of runic inscriptions and skaldic verse with a description of the stone and its inscription, calling it ‘one of Scandinavia’s most remarkable monuments from the Viking Age’.See Jesch (2001: 1). The Viking Age is usually held to designate the period of AD 750-1100; Jesch calls c.950–c.1100 the ‘late’ Viking Age. In this paper, I will briefly outline the origins, distribution, and chief features of the runic alphabet; contextualize the inscription on the Karlevi runestone; analyse the actual lettering and layout of the inscription as carved on the stone in relation to the metre, style, and cultural context of the verse it inscribes; and finally describe the character of skaldic verse and interpret the stanza.

Runes

Runes are the letters of an early Germanic alphabet. The earliest inscriptions have been dated to the period AD 150–200, and although there are a few regions in Scandinavia where runes were still being used in the fifteenth century, in most other places their use had declined, and disappeared even earlier on the Continent and in England. It is important to be clear that there is no runic language as such: runes were used to inscribe various early Germanic languages, such as Old English, Old Norse, or Old Frisian. There are some Latin inscriptions using runic characters, and even the odd word in Greek or Hebrew in Christian inscriptions. Most runic inscriptions are found in Scandinavia—in Sweden, Norway, and Denmark, with by far the most in Sweden—but they are also to be found in a number of other places with Scandinavian connections, such as Iceland, Greenland, and England. They certainly originated as a writing system somewhere outside the Roman Empire. Many believe that runes developed amongst Germanic peoples in direct response to exposure to the Roman alphabet, when Germanic tribes came into close contact with Rome in late antiquity, and admired and envied their literacy, perhaps with particular reference to monumental inscriptions.For a clear summary of scholarship on the origins of the runic alphabet, see Barnes (2012: 9). There are clear resemblances to—but also differences from—a number of Mediterranean writing systems, whether Roman, Greek, Phoenician, or northern Italic alphabets such as Camunic, and Lepantic. However, the precise roots of the runic alphabet are unknown, and still disputed by scholars.

Essentially, runes were used to inscribe various materials, though not for conventional writing in ink on vellum, parchment, or paper. Runes are constructed primarily of vertical stems, with either straight or curved side branches; there are very few horizontal lines. Perhaps these shapes were developed so that carved lines would be clearly distinct from the naturally occurring horizontal grain of wood, though wood was only one of the materials used for runic inscriptions. Further, although the Danish runologist Erik Moltke maintains that ‘it is clear that the runes were designed to be carved in wood’, if he is inferring this from their verticality then there is a suspicion of a somewhat circular argument (Moltke 1985: 32). Some inscriptions consist simply of the runic alphabet itself, in a conventional order that gave the alphabet its name: the futhark, or fuþark, after the first six letters in the series: f, u, þ, a, r, and k. The forms of these letters, and the alphabet as a whole, changed over time, and there were also local variations. A late Viking Age form of the runic alphabet—the so-called ‘Younger Fuþark’—is similar to that found on the Karlevi stone:

The Karlevi runestone in context

The portability of many items bearing runic inscriptions means that it can be difficult to ascertain where an inscription originated. This is of course not the case with large and extremely heavy items such as the Karlevi stone.

Nevertheless, the Karlevi stone presents its own issues of dating and provenance. Both the language—a form of West Scandinavian such as a Norwegian or Icelandic poet might use—and the shapes of individual runes have led scholars to date the inscription to around the year AD 1000 (see, for instance, Larsson 2005: 408). The alphabet itself is a variant of the Younger Futhark known as ‘long branch’, which is sometimes referred to as ‘Danish runes’ although it was used in Sweden as well. The placing of the inscription in more-or-less straight lines on the three flat faces of a roughly rectangular stone is also characteristically Danish in style. As Jesch points out, contemporary national boundaries do not necessarily indicate medieval ones, and in the Viking Age parts of southern Sweden were under Danish control. Nevertheless, as she notes, ‘the rune-forms and the style of the inscription are closest to those of Denmark proper, rather than those of the ’Danish’ areas of southern Sweden’ (Jesch 2001: 2). Swedish runic inscriptions are more often run along the extended ribbon-like bodies of snake- or serpent creatures, which either form the outer border around an illustration or intertwine in a complex tangle (Jansson 1962), though such inscriptions tend to be later, and some earlier Swedish inscriptions are more like Danish ones. It is significant, then, that the stone, which is almost certainly in its original position, records Norwegian or Icelandic poetry in typically Danish runes and layout, but is found in Sweden and not in Western Scandinavia.

The key to understanding why a runestone with Danish runes is found so far to the east of Scandinavia may lie in the proposed date and evident function of the stone. The Karlevi stone, as is plain from the text of the inscription, is a memorial gravestone. The section of its inscription which is not part of the stanza is very hard to read and interpret, but one possible reconstruction and interpretation is that the stone was set up in memory of Sibbi the Good, the son of Foldarr—of whom nothing else at all is known—and this information is followed by a damaged section of text which may perhaps be interpreted as reading ‘and his retainer placed on Öland this memorial to honour the dead’.The formulaic nature of runic memorial inscriptions can often aid the interpretation of fragmentary and otherwise difficult-to-decipher inscriptions, such as this one; See Jesch (2001: 2).

A seventeenth-century Swedish antiquarian, Jonas Rhezelius, made reference to the stone in a diary entry from 1634, when he was visiting the antiquities on Öland, and a contemporary sketch of the site shows the runestone and two burial mounds that have now disappeared. The date, and mixed provenance indicators, may further suggest that the stone marks the grave of, or commemorates, a Danish war-leader who fought in the Battle of Fýrisvellir in Sweden, which took place at the end of the tenth century.

The Battle of Fýrisvellir was fought near modern Uppsala. Eiríkr, King of Sweden, was challenged by his nephew, named in Old Norse sources as Styrbjörn, the leader of a formidable and celebrated force of Viking mercenaries called the Jómsvíkings, who were based somewhere in the Baltic.Throughout this piece, I use the conventional Old Norse forms of most proper names—thus, for example, Eiríkr and not Eric or Erik. Some sources claim that Styrbjörn was the founder of the Jómsvíkings, others that he took the leadership over a group founded by the Danish king whose daughter he married: Haraldr blátönn (Bluetooth), the son of Gormr the Old. There are brief allusions to the Jómsvíkings in Old Norse poetry and sagas, but the most prominent account is the thirteenth-century Old Norse Jómsvíkinga saga, which itself exists in different recensions and has an enormously complex textual history. It is impossible to be sure how historically reliable any of the sources are.See Finlay (2014: 63–79) and a number of other articles in this issue of the journal, which is dedicated to the saga and related texts. See also Rowe (2001: 3–28). However, according to Styrbjarnar þáttr Svíakappa (The Story of Styrbjörn the Champion of the Swedes), a short narrative in the Old Norse historical compilation Flateyjarbók, Styrbjörn gathered support in Denmark before taking on Eiríkr of Sweden. He was at first allied with Haraldr Bluetooth, but took Haraldr hostage and demanded a huge Danish fleet to support his attack on Eiríkr. In spite of this, Styrbjörn’s troops were defeated and his attempted takeover of the Swedish throne failed. Haraldr apparently escaped from Fýrisvellir.

There are many examples of runic inscriptions on stone cenotaphs (tomb-like monuments commemorating people whose remains are buried elsewhere) all over Scandinavia, and more particularly, other runestones in Sweden that may commemorate Danish warlords who fell at Fýrisvellir, or at least a battle near or at Uppsala. A stone at Sjörup commemorates Tóki’s son Ásbjörn, who ‘fought as long as he had a weapon’ and points out proudly that he did not flee from Uppsala (Jesch 2001: 225). The inscription on one of the Hällestad runestones states that it was raised in memory of Tóki the son of Gormr, by Áskell, one of his retainers, and it too boasts that its subject did not flee Uppsala (Jesch 2001: 222–23).

Jesch notes that, as Moltke demonstrated, part of the latter inscription is in the traditional Old Norse poetic metre fornyrðislag (old story metre), with its short, two-stress, two- or three word lines linked by alliteration. The verse is not laid out on the stone in poetic lines, but the metre is obvious from the layout below (Jesch 2001: 222, reproducing Moltke):

However, the major part of the Karlevi inscription constitutes one whole stanza of skaldic verse which praises the bravery and success in battle of a Danish chieftain. Before considering the unique features of the extraordinary skaldic genre, I turn to the physical form of the artefact—the runestone itself—to describe the actual layout of the inscription, which follows neither metre nor meaning.

Layout and lettering

The stone itself is grey granite, stands about a metre and a half above ground, and is about 70cm in diameter. As Jesch describes it, ‘the inscription falls into two parts, clearly indicated by cross- or hammer-like marks at the beginning of each. From adjacent starting points at the bottom of the stone, both texts proceed boustrophedon (“as an ox turns in ploughing”) in different directions, the first going off to the right and occupying three lines, the second going off to the left, occupying six lines’ (Jesch 2001: 1). The figures accompanying this paper show the three-line section that comprises the corrupted dedication, while the six-line section comprises the skaldic stanza.

I have made a conventional transrunification of the stanza and arranged it in lines corresponding to the bands on the stone itself, based on the edition in Ölands runinskrifter (see Söderberg and Brate 1900–1906: 24–25). This arrangement is no longer common ; most accounts of runic inscriptions offer transrunifications and transliterations as continuous prose, with appropriate divisions sometimes marked by a vertical line (|). However, the banded lines on the stone do not correspond at all to poetic lines as defined by skaldic metre:See Jesch (2001: 320). The asterisks and colons represent actual punctuation marks on the stone, which are often extremely difficult to distinguish from naturally occurring flaws. Letters in round brackets are uncertain; unreadable characters are indicated by a dash (–). See Barnes (2012: 224) for a full account of the conventions of transliteration.

ᚠᚢᛚᚴᛁᚾ : ᛚᛁᚴᚱ : ᚼᛁᚾᛋ : ᚠᚢᛚᚴᚦᚢ : ᚠᛚᛅᛁᛋᛏᚱ
* ᚢᛁᛋᛁ * ᚦᛅᛏ * ᛘᛅᛁᛋᛏᛅᚱ * ᛏᛅᛁᚦᛁᚱ : ᛏᚢᛚᚴᛅ *
: ᚦᚱᚢᚦᛅᚱ : ᛏᚱᛅᚢᚴᚱ : ᛁ : ᚦᛅᛁᛘᛋᛁ * ᚼᚢᚴᛁ :
ᛘᚢᚾᛅᛏ : ᚱᛅᛁᚦ : ᚢᛁᚦᚢᚱ : ᚱᛅᚦᛅ : ᚱᚢᚴ : ᛋᛏᛅᚱᚴᚱ
* ᛁ * ᛏᛅᚾᛘᛅᚱᚴᚢ : – (ᚾ)ᛏᛁᛚᛋ : ᛁᛅᚱᛘᚢᚾ *
* ᚴᚱᚢᚾᛏᛅᚱ : ᚢᚱᚴᚱᚬᚾᛏᛅᚱᛁ * ᛚᚬᚾᛏᛁ

fulkin : likr : hins : fulkþu : flaistr (:)
* uisi * þat * maistar * taiþir : tulka *
: þruþar : traukr : i : þaimsi * huki :
munat : raiþ : uiþur : raþa : ruk : starkr
* i * tanmarku : –(n)tils : iarmun *
* kruntar : urkrontari * lonti

Most obviously, there are only six lines rather than the usual eight of a classic skaldic stanza, and the original punctuation does not represent poetic line breaks. Neither the metre nor the meaning as evidenced by prose word-order are apparent at a glance from this layout. Fig. 1 displays the first part of the strophe (direction of reading). To see the second part of the strophe in full, one has to look at the stone from a different angle (Fig. 4).

As a reductive pragmatic point, monumental texts laid out boustrophedon are notoriously awkward to read, involving a good deal of neck-craning. This suggests that the primary purpose of the inscription was to exist and endure, not to be read. There is an intriguing link here to the Ruthwell Cross, an eighth-century Anglo-Saxon runic monument with an inscription that closely corresponds to a Christian poem, ‘The Dream of the Rood’, which is also found in in a tenth century codex.It has been suggested that the runes were added to the monument in the tenth century; see Conner (2008: 25–51). That stone cross is approximately 5.5m high, and while the runic inscription does not reach to the very top of the cross, it would nevertheless be very difficult to ‘read’ the inscription from ground level. But an even more intriguing similarity between these two runic monuments is that both of them also carry Christian inscriptions in Latin, in Roman letters.

On the opposite face of the Karlevi stone from the runic stanza is a fragmentary inscription which seems to read: INONIN, and underneath that, HE. Both of these scraps are prefaced with a cross, though the shape could conceivably be that of Thor’s hammer. It seems as though the inscription was an attempt to carve ‘In nomine Jesu…’. It is uncertain whether this inscription is contemporary with the runic one, but if it is, then it is an example not only of an attempt to include both Christian and possible pre-Christian references in a memorial text, but also of a remarkable concurrence of runic and Latinate literacy. One can only speculate as to precisely whose literacy these inscriptions speak to, and of.

Skaldic verse

I turn now to consider the unique features of the skaldic genre. Skaldic metre, or dróttkvætt (court metre), is much more complex than the relatively simple old story metre, or fornyrðislag, of the Hällestad stone. Dróttkvætt stanzas consist of eight short lines (or four long ones, each with a break in the middle), each short line having six syllables, with the final syllable being unstressed. The first and every following odd line contain two alliterating syllables, and two internal half-rhymes. The second and every following even line pick up on the alliteration from the previous line on one syllable (usually the first syllable of the even line), and contain two internal full rhymes. Each eight-line stanza is divided into two halves, or helmingar, which are syntactically self-contained, but are often linked together by a conjunction such as ‘when’, or ‘because’. It involves a considerable departure from the norms of prose word-order, and the cryptic nature of the stanza is further increased by the use of poetic synonyms and, particularly, poetic periphrases known as kennings.

Kennings use two or more elements to denote a single referent, these elements being linked in a genitival relationship. Thus, a skaldic poet might call a Valkyrie ‘a goddess of battle’, that is, with the base word ‘goddess’ (and sometimes the name of an actual goddess) standing for any supernatural female, and then defined by the determinant ‘of battles’: the supernatural female associated with battle is thus a Valkyrie. Similarly, the poet might call battle ‘the storm of Odin’, that is, the base word ‘storm’ standing for some sort of injurious (outdoor) tumult, but defined by the determinant ‘of Odin’, the god of battle: the injurious tumult associated with Odin is thus a battle.I use familiar English versions of the names of Odin and Thor, although all other names are in normalized Old Norse forms. This can lead to infinite regression if the elements are linked together—‘the goddess of the storm of Odin’, and so on—though skaldic poets were encouraged to not go beyond five elements in extended kennings. However, it should be noted that in an extended kenning, all the elements except for one (the base word) are in the genitive case, so once the word order is disrupted—as it almost always must be to accommodate the alliteration and rhyme—the order of these kenning elements can be very hard to determine. Furthermore, a degree of familiarity with Old Norse myth is vital in order to understand some of the allusions, which refer to many personages and events across the whole corpus of Old Norse mythological traditions—even including some names and episodes for which we have no other surviving evidence.

The Karlevi stanza

Sven B.F. Jansson, in his description of the Karlevi stone, rightly noted that the stanza it records ‘could serve as a model example of a skaldic poem in dróttkvætt. It complies with all the strict rules of the metre’ and, moreover, ‘the imagery and phrasing are also completely characteristic of skaldic style’ (Jansson 1962: 125). Therefore, to illustrate the character of a typical skaldic stanza, it seems pertinent to use the Karlevi stanza as a demonstration model, with the language of the stanza here converted into normalized Old Norse:Although I am not at this stage concerned with the content of the stanza, but rather with its form and metre, I nevertheless offer here one possible simplified prose translation, which, I should stress, is not the only way of construing the stanza: ‘A warrior—[one] whom [the] greatest deeds followed—(most knew that) lies hidden in this mound. A battle-strong, more unharmful sea warrior will not rule land in Denmark’.

Folginn liggr hinns fylgðu,
flestr vissi þat, mestar
dæðir dolga Þrúðar
draugr í þeimsi haugi.
Munat reið-Viðurr ráða
rógstarkr í Danmörku
Endils jörmungrundar
ørgrandari landi.

The insistent regularity of the metre is plain even to those who do not know the original language. The eight short lines of six syllables are evident enough, as perhaps is the way each short line ends on an unstressed syllable. The alliteration and rhyming scheme are also apparent. For example, the alliteration of folginn and fylgðu in line one, the first odd line, is picked up by the first letter of the first syllable of the first even line (line two) with the word flestr, and so on throughout the stanza. The alliteration is complemented by a subtle internal half-rhyme where there is double alliteration—folg- and fylg- —and a ringing internal full rhyme—flest- and mest- —in the lines with only a third alliterating syllable. In this stanza, the two helmingar are free-standing and separated by modern editorial punctuation which echoes the sense of the stanza. The stanza is not laid out in poetic form on the stone, but essentially—as with skaldic stanzas in manuscripts containing prose texts of Old Norse sagas—the poetic lines on the stone are run together continuously as if they were prose. In saga manuscripts, apart from the occasional marking-up of the initial letter of a strophe, one cannot distinguish the stanzas from the prose of the saga. They are effectively invisible on the page, but once the text is read the metre is unmissable.

As is true of all dróttkvætt stanzas, even if the metre of the Karlevi stanza is obvious, the meaning is not. Despite its layout, anyone who can decipher runes and is familiar with the rules of dróttkvætt can identify what is indisputably a stanza of skaldic poetry, but extracting meaning from it is a challenge. This may well have even been the case for a Viking Age audience. Although it is theoretically possible that skaldic strophes might have been composed impromptu, and understood at first hearing by practised connoisseurs of the art, it is telling that one of the most celebrated Old Icelandic family sagas—which are renowned for their apparent naturalism—tells of a woman who heard a skaldic verse recited, committed it to memory, and then interpreted it at her leisure (Þórólfsson and Jónsson 1943: 58–59). The difficulties presented by dróttkvætt stanzas mostly arise from the uncertainties of word order and the interpretation of kennings. Moreover, skaldic stanzas may be rich in association and ambiguity beyond one unitary and arguably reductive meaning. The Karlevi strophe can be used to demonstrate this, taking it one helmingr at a time.

The first helmingr makes clear the function of the stone and its inscription as a grave-marker or cenotaph, and heaps conventional praise on the leader buried in the mound:

Folginn liggr hinns fylgðu,
flestr vissi þat, mestar
dæðir dolga Þrúðar
draugr í þessu haugi.

Translated following the original word order, however, the sense is wholly obscure:

Hidden lies whom followed
most knew that greatest
deeds of hostilities of Þrúðr
log/ghost in this mound.

Apart from the clarity of the phrase ‘most knew that’—and a short lucid intercalary phrase in a helmingr is a common stylistic feature in skaldic stanzas—the word order is jumbled not only beyond conventional prose, but also beyond any recognizable sense. I propose the following prose order (one of several possibilities):

A log/ghost of Þrúðr of hostilities—[one] whom [the] greatest deeds followed (most knew that)—lies hidden in this mound.

The kenning still needs explication. Þrúðr is the name of a goddess in Old Norse myth, said to be the daughter of the god Thor and his wife Sif. Here, her name is used to denote any supernatural female—a poetic device noted above—and the supernatural female associated with hostilities was traditionally a Valkyrie. The name itself means ‘strength’ in Old Norse, which is appropriate in this martial kenning. But it may also be that certain stories in Old Norse myth made the use of her name in this kenning particularly appropriate. In a very early skaldic poem—the first stanza of Bragi’s Ragnarsdrápa (Clunies Ross 2017)—the giant Hrungnir is called ‘the thief of Þrúðr’, which implies a story about her abduction though that story has not survived. One might even wonder if there is a distant echo of the story about King Haraldr being held hostage.

The full kenning includes the base word draugr: ‘the draugr of Þrúðr of hostilities’. The reference is to the fallen warrior commemorated by the runestone, so the whole kenning denotes a warrior. But draugr has two distinct meanings in Old Norse. In kennings, a draugr is a log—one of a large number of words associated with trees that serve as base words for warriors. A warrior may, for example, be called ‘the tree of battle’. One might understand that a tree-word denotes any upright living thing outdoors, and when associated with battle could refer to a warrior. But one might go further, and reflect on one of the Old Norse creation myths in which the first humans were fashioned from driftwood. I see this kenning as part of a sustained intertextual body of poetic imagery in which kennings such as the ‘storm of Odin’—with arrows and projectiles figured as lashing rain and driving hail—create the image of an actual storm lashing the trees of a forest, which in turn stands for close-packed warriors in a battle. Accordingly, the meaning of the word draugr in this sense would be ‘log’—that is, a tree which has been felled, or cut down, like a dead warrior.

The second sense of draugr in Old Norse is even more suggestive. In Old Norse sagas, communities are sometimes terrorized by a zombie-like supernatural creature, a malevolent and physically violent revenant coming back from the dead to wreak havoc on the living.See, most famously, the rampages of the revenant Glámr: Jónsson (1936: 7:107–23). Folginn (hidden) is a euphemism for ‘buried’ when employed elsewhere in Old Norse poetry (Jansson 1987: 134–6). Perhaps the poet of the Karlevi stanza imagined the leader he was praising as vengefully coming back from the dead—and out of his burial mound—to posthumously pursue his enemies. Tellingly, the simple past tense plural fylgðu (‘followed’, its plural subject being famous deeds) is etymologically related to the word for the Old Norse concept of a guardian spirit, often manifesting itself after death, the fylgja.

Similar analysis can be performed with the second helmingr:

Munat reið-Viðurr ráða
rógstarkr í Danmörku
Endils jörmungrundar
ørgrandari landi.

Again, this makes little sense unless the word order is rearranged. Translated word for word it reads:

Will not wagon-Viðurr rule
powerful in Denmark
Endill’s monstrous ground’s
more unharmful land.

Rearranged, it can be read:

A battle-strong, more unharmful Viðurr of the wagon of Endill’s monstrous ground will not rule land in Denmark.

Again, explication of the kenning reduces the meaning to manageable levels. The name Endill is traditionally included in a list of legendary sea-kings. In an elemental paradox entirely characteristic of skaldic kennings, the ‘ground’ of a sea-king—that is, his natural habitat—is the sea. The sea is immense and unbounded (unlike the lands of Scandinavia, which were so ferociously fought over), and may therefore be termed Endill’s ‘monstrous’ ground. Viðurr is a name for Odin, the wagon of the sea is a ship, and the ‘Odin’ of a ship is the commander of a fleet, or sea warrior—who will eventually be interred in a mound. Calling him ‘unharmful’ is probably a litotes—common in Germanic poetry—for ‘benevolent’, or ‘beneficial’, but there is also perhaps a faint martial paradox at work here: a successful war-leader might be expected to cause considerable harm to his enemies.

But, as always in such poetry, there are further allusions at work. In Old Norse mythology, the god who is most often associated with a wagon or chariot—the reið—is Thor, so there are sub-textual connections to Thor as the father of Þrúðr in the first helmingar, and to his association with a wagon in the second. The Odin-name, Viðurr, plays on its near-homonym viðr, which means ‘tree’ in Old Norse, thus picking up the nexus of kennings that associate trees with battle, the storm of Odin. It is also worth noting the Old Norse lexical element I have translated here as ‘monstrous’: jörmun-. This word is cognate with the Old English element eormen, which seems to mean ‘immense’. Jörmungrund- is found in the Old English epic poem Beowulf as eormengrund, where it is used to designate the wide earth, as it often does in Old English poetry (Fulk, Bjork, and Niles 2008: 31 [l. 859]). Old Norse poets called the ‘World Serpent’ by the name Jörmungandr, apparently meaning ‘huge monster’, long before the Karlevi stone was inscribed. The second element of this term, -gandr, also means sorcery or magic elsewhere in Old Norse, and Jörmunr is one of the many names of the god Odin. It may even be cognate with the first element of the Old Saxon name for a sacred ‘world-pillar’, the Irminsul, which itself might contain the name of an otherwise unknown god, Irmin.See Vries (1977). Simek (1993) is more sceptical. The Karlevi stanza may share in an ancient, and evidently pan-Germanic, poetic lexis.

It is possible that some asyntactic juxtapositions were forced on the poet by the demands of metre. And yet they demand further scrutiny. For example, in the first helmingr, the identity of the commemorated leader is contained in a frame constituted by the first two words—Folginn liggr (lies hidden)—and by the final three—í þessu haugi (in this burial-mound). In the third line of this helmingr, dæðir (deeds) is next to dólga (the genitive plural of the word for hostility). These two words do not belong together syntactically, because the first is part of a statement that great deeds accompanied the chieftain, while the second is part of a kenning for a warrior (the log/ghost of the Þrúðr of hostilities). But the two words together echo a phrase used elsewhere in Old Norse poetry—dauðir dólga—which means ‘ghosts’ or ‘fiends’. These two phrases mean different things, but to those attuned to the persistent use of half-rhyme in skaldic verse this would not disguise a sly reference to the word draugr (which can mean revenant) in the next line. The echo is not obvious, but skaldic verse is full of such word play.

The lament in the second helmingr, that a better chieftain will never [again] rule in Denmark, is a conventional expression of praise in Old Norse formal memorial poetry. It might be compared to the ending of a celebrated praise poem, of around the same date, lamenting the passing of King Hákon. Its poet, Eyvindr skáldaspillir (meaning either ‘the plagiarist’ or ‘the one who shows up other poets’) maintains that it will be a good day when another leader as good as Hákon is born, and asserts that ragnarök, the Old Norse apocalypse at the end of the world, will be here before the world sees his like (Aðalbjarnarson 1941). But if the words are taken together as they originally appear in the stanza, the chieftain is rógstarkr í Danmörku (battle-strong in Denmark), though this is poignantly set against the outcome of the battle, meaning that he munat … ráða (will not rule). If Sibbi the Good died in battle, then the inscription related a touching blend of promise and disappointment; of praise and resignation. He was an impressive warrior, but he was defeated.

This extended explication of the Karlevi stanza demonstrates how a wealth of meaning can be contained in a tightly limited format containing a mere 48 syllables. In fact, even the lengthy interpretations above are limited by my own knowledge, and my critical ingenuity. These denotative and connotative meanings cannot be encompassed in one reading (or hearing) of the stanza. It is tempting to suppose that the dense word-play in skaldic verse is a response to the time-consuming and laborious effort of inscribing runes—that the laconic polysemy of the typical skaldic stanza developed because of the need for inscriptions to be as short as possible. But the elaborate metre and relatively brief span of dróttkvætt stanzas can perhaps be more plausibly understood as a mnemonic aid—a means of dramatically minimizing the likelihood of textual corruption in oral transmission in a period of limited or virtually absent literacy (Jesch 2017). It is significant that thirteenth-century Icelandic historians quoted skaldic strophes in their prose as corroborative evidence of historical facts. Iceland’s leading medieval historian, Snorri Sturluson, made this fully explicit in a prologue where he pointed out that although he used many oral sources, such as old songs or the testimonies of older people with long memories, skaldic verses are the most reliable historical sources because if they are composed in regular metre, and are understood correctly—something not always achieved even by medieval historians—then they are resistant to corruption.See, for instance, the thirteenth century misinterpretation of the kenning skeiðar brandr (the blade of a ship, that is, its prow) as a proper name: Einarsson (1985: 4). And, of course, once engraved in stone, and even allowing for the erosion of centuries and the damaging growth of lichen, inscriptions can be extraordinarily durable. Swedish scholars have shown that runic inscriptions were used to literally ‘set in stone’ property and inheritance claims (Sawyer 2000).

However, the mnemonic function of skaldic verse is not relevant to the Karlevi strophe, which is essentially an elaborate and highly metaphorical statement of almost entirely non-specific praise. Apart from the reference to Denmark (í Danmörku) the only proper names are to mythological personages, and there are no details of the battle or its participants, beyond a vague maritime reference. The stone and its stanza ‘mark the spot’ of an inhumation, but record no facts. The striking multiplicity of metaphorical meanings not only calls into question the function of the inscription as a record of information, but also challenges the assumption that it offers itself as a text to be read in the usual sense of the word. It is, in any case, far from certain whether the runic alphabet was accessible to contemporary readers. Recent scholarly work has suggested that the very appearance of the runic script may deliver a visual message in itself—rather like the decorative use of fake Chinese or Japanese characters to give objects an aura of orientalism—quite apart from the content of the inscription (Birkett 2017). It also remains uncertain whether poet and runecarver were the same person. It seems likely that this would have been the case, but no-one knows how widespread these skills were amongst Scandinavian warriors, and it is theoretically possible that groups of Vikings would have included semi-professional rune carvers or poets (Moltke 1985). Clearly, the poet of the Karlevi strophe was accomplished, not a simple foot-soldier attempting to produce a passable stanza. The runes themselves were accurately and skilfully carved, and very clearly not the work of a hard-pressed amateur. It is thus tentatively possible to conclude that a stanza in a metre designed both to withstand corruption in oral transmission and to live in the minds of its audience was an almost meditative text, one that gradually released and created a multitude of parallel meanings, and which was only later transformed into a skilfully crafted and durable material artefact. Although a present-day audience may assume that the stanza was inscribed simply to be read, perhaps this was not its primary function at all.

Acknowledgements

I am indebted to Matthew Roby, of Exeter College, Oxford, for help with this article—not only for delivering it in my place at the conference organized by Yegor and Lesley, but also for invaluable help with this online version of the conference paper.

Bibliography and further reading