Many thanks go to Peter Tunstall who has generously allowed us to post this saga here. This saga translation is copyrighted. © Peter Tunstall, 2003. This saga may be freely copied, stored and distributed for noncommercial use.

Note on Translation

The Manuscripts

The Manuscripts

Hervor's Saga exists in many manuscripts, six of which are useful for establishing the text. There are three distinct versions of the story, known as R, H and U. Of these, R is thought to be closest to the 13th century original, but lacks an ending. It also has a gap covering the end of Chapter 5 (of the present translation) and the beginning of Chapter 6, and breaks off towards the end of Chapter 12. The R version comes from the early 15th century vellum Gl.kgl.sml. 2845, 4to, of the Royal Library in Copenhagen.

U is preserved in the corrupt paper manuscript U (R:715 of the University Library, Uppsala), mid 17th century, and a part also appears in AM 203 fol. of the University Library, Copenhagen, written by Síra Jón Erlendsson of Villingaholt (died 1672).

The H version was made by Haukr Erlendsson (died 1334) in the Hauksbók, AM 544, 4to. H is closer to U than to R. From the answer to the second riddle, the ending is missing, but two 17th century copies, AM 281, 4to (h1) and AM 597b, 4to (h2), preserve the H version to the end of the riddles.

Which Versions I Have Used

This translation is based mainly on R, for which I have used G. Turville-Petre's edition of Hervarar saga ok Heiðreks, published by the Viking Society for Northern Research (Text Series: Number 2), which follows R "as far as it goes". (Turville-Petre fills in the first gap from H, and the end of the saga, including the Epilogue, is taken from U, with some readings from 203).

In a few places (see below), I have supplemented this with material from H and U. For the former I used Finnur & Eiríkur Jónsson's 1892-96 edition of Hauksbók (Vol. 2, pp. 350-369), and for the latter, Hervarar saga ok Heidrekskongs, ed. Stefán Björnsson, 1785, which actually uses the U-type 17th century paper manuscript AM 345 4to. Both are available online at Saganet.

Divergences from the Turville-Petre edition are as follows:

The story of the king and the dwarves in Chapter 1 is from H, with nods to U. (This incident is only alluded to in R).

The opening to my Chapter 6, introducing Gudmund, is actually the very beginning of U and H. Again, I made an amalgam.

Extra riddles have been included from H, together with much additional repartee between Gestumblindi and the king. The order of riddles varies between versions, and in a few cases I have altered it further to accommodate asides from both R and H. In the process, some asides in my translation have been shifted to follow a different riddle. From H too come several verses in The Waking of Angantyr which are not in R.

In R Arngrim stays peacefully with King Sigrlami. In H & U, Arngrim never stays with the king (who is called Svavrlami) but only attacks him. I have combined these two stories, thus altering somewhat the meaning of both. Having included Dvalin's curse (which doesn't feature in R), it was necessary that the king should die by his sword, but I wanted to keep the story of Arngrim's service in Reidgotaland as this prefigures Heidrek's. In R, the king just retires.

The introduction of Arrow-Odd in Sweden is from U. Heidrek's life in the woods is from H, as is the mention of the retainer slain inadvertently at the end of the riddles, and the people's desire to expose the baby in Chapter 5; so too the names of the six berserks from Brami to Bui inclusive (since only six names are given in R), and the detail regarding the relative weakness of the Haddings. I stuck to R however for the order of the first four.

For the story of Starkad as told in U, see Appendix B.

Names

Names are transcribed in the standard way, with some exceptions: Harveth Fells (Harvaðafjöll), rather than Harvad, for instance. This is the Germanic form of the Carpathian Mountains, and is evidence for the antiquity of the legend's oral transmission; I wanted to give it the sound it might have had had the word survived in English. The Icelandic ð is actually a voiced fricative as in English "the".

Gardariki (Garðaríki) 'realm of fortified towns' is left as such, but identified with Russia the first time it appears. In Old Norse it referred to the kingdom of the Rus, based on Novgorod, Kiev, etc. Its king is the Garda King.

Sometimes a geographical component is translated, e.g. Mounts of Jass, and Jassar Fells, both from Jassarfjöll. I anglicised the unidentified River Graf as 'Grave', preserving possible semantic associations in the original (in this word, the Icelandic f is pronounced like English "v"). And Myrkviðr 'the Dark Forest' becomes its well-known English equivalent, Mirkwood.

Other names have generally been left in Old Norse form, minus vowel-marks, non-English letters and inflectional endings - even where they have been identified with modern places. Thus Samsey, rather then Samsø; Bolm, rather than Bolmsö. The former is an island between Jutland and Zealand (now more famous for its music festival than for ghosts!), and the latter on Lake Bolmen in southern Sweden - although the H and U versions locate it in Halogaland in Norway.

On Continuity

Various contradictions in the story remain. So, for instance, the troop numbers attributed to the Huns in Gizur's verse, Chapter 14, differ from the amounts given in the prose of Chapter 13.

Since in other versions, Heidrek kills his brother with Tyrfing, I suspect I may be one "vile deed" short of the prophesy (depending how you count them) - troubled readers are invited to commit one of their own. After Heidrek's death, Tyrfing seems to lose its quality of killing a man each time it is drawn - unless in some lost version the pike was really a person? (cf. The Saga of the Volsungs, where Andvari lives in the form of a pike). In one verse, late in the saga, "Tyrfing" seems to refer not to the sword, but to a Gothic tribe, the Tervingi of Latin historians.

Confusion exists as to the names of the foster-fathers of both Hervors. That of the first is called Bjarmar by the saga. But a single obscure reference to Frodmar in the first verse of Chapter 5 may imply a different version, unless it is the name of Hervor's supposed father in the thrall's insult. The second Hervor is said in Chapter 10 to have been fostered by a Jarl Frodmar in England, yet by Chapter 13, her foster-father has become Ormar, a Goth - all this according to the R-version. (Actually, for continuity, I have added to Chapter 10 a mention of Ormar from H, thus: "with Jarl Frodmar and later with a man called Ormar." The words "and later" are my own).

Similarly puzzling is the name Soti, mentioned in Hjalmar's Death Song but nowhere else in the story. Virtually the same verse occurs in Arrow-Odd's Saga, which sheds no light on Soti either, although Odd does kill a viking of that name earlier in an unrelated incident. The verse implies that he was a comrade of Hjalmar's who either came along with Odd, or replaced him, or was him, in an alternate version.

However, these inconsistencies hardly seem to matter in the raggedly stitched but intense thing that is Hervor's Saga. In fact, such detached names give a clue to the composition of the saga, confirming that the poetry existed in oral tradition long before it was incorporated into a written prose narrative.

The Riddles

Tafl is a general term for various board games, including chess and the Viking board game hnefatafl or hnettafl, in which one player must protect the king or hnefi.

Note that the "hawk bearing eider-duck into crags" riddle relies on very cryptic wordplay, only partly explicable anymore. For example, the word segl is used here for 'sail', but another Norse word veggr has the double meaning 'sail' and 'wall'. "Dead men" is a pun on valr, which also means 'hawk'. Interpretations of the last two lines are speculative.