Earl Thorfinn of Orkney and the forgotten battle of 1058

Abstract

Verses by Arnórr ‘jarls’-poet’, preserved in Orkneyinga saga, describe a battle in which Earl Thorfinn of Orkney fought the English south of the Isle of Man. The thirteenth-century saga-author associated these verses with a campaign supposedly fought in the 1030s or 1040s, but this account is doubtful. Turning to the Norwegian expedition of 1058, this article considers whether the verses might originally have referred to that campaign and later become linked with a different story. New readings of the key stanzas are proposed, and a new sequence, with consequences impacting on the chronology and circumstances of Thorfinn’s life and death. There are ramifications for discussions touching the saga tradition, Arnórr’s career, Malcolm (Máel Coluim) III Canmore’s career, and his marriage to Ingibjorg.

Soon after the death of Earl Thorfinn (Þorfinnr jarl) of Orkney in the middle years of the eleventh century, Arnórr jarlaskáld (jarls’-poet) composed a memorial encomium (erfidrápa) in the late earl’s honour. Employing the chief metre of skaldic verse, dróttkvætt (named after the ruler’s chosen band of fighting men, the drótt), his poem would have been ‘declaimed out loud’ to the dead earl’s hearth-troop. ¹ Like other forms of skaldic poetry, the memorial drápa was a form composed and passed down orally, utilising devices of metre, syntax and diction to ensure the integrity of its transmission over time. Verses thought to derive from Arnórr’s Þorfinnsdrápa (memorial poem in honour of Thorfinn – hereafter Þorfdr) are preserved as inclusions in later prose works, principally Orkneyinga saga, the early thirteenth-century saga of the deeds of the earls of Orkney. ² The poem’s editor, Diana Whaley, an authority on Arnórr and his work, identifies sixteen stanzas and nine half-stanzas (helmingar) as belonging to the poem and as dispersed individually or in groups in Orkneyinga saga and elsewhere. ³ Putting them into a sensible order by reference to their themes and subject matter, and to the context in which they are cited in the saga, Whaley’s reconstruction of Þorfdr is presented with two caveats: that parts of the poem may have been lost – as with any poem reconstructed from fragments – and that arguments can be made for ordering the stanzas differently. ⁴

The subject of this article is a battle which Earl Thorfinn fought in England, against the English, which Arnórr commemorated in Þorfdr. According to Orkneyinga saga, the battle occurred at a time when Thorfinn and his nephew and fellow earl, Rǫgnvaldr, were raiding south of the Hebrides and around the eastern shores of the Irish Sea. ⁵ Rǫgnvaldr’s alleged involvement in the battle serves as a chronological marker, given that he died c. 1045. A second reference point is provided by the saga-author’s statement that Harthacnut ruled England and Denmark at the time, and that he was in Denmark at the time of the battle. ⁶ These pointers suggest a date between 1035 and 1040, during which time Harthacnut maintained a claim to be England’s ruler but was detained in Denmark by his wars with Magnus of Norway (Magnús Óláfsson, d. 1047) and was unable to enforce that claim. Perhaps, however, the episode is supposed to be set during his reign proper (1040–2).

The saga-author tells the following story. One summer, Earl Thorfinn went raiding in the Hebrides and in various parts of Scotland. While he lay at anchor off Galloway ‘where Scotland borders on England’, he sent some of his troops south to raid the English coast, ‘as the people had driven all their livestock out of his reach’. The English assembled, made a counter-attack, recovered all that had been stolen, and killed every able-bodied man among them, except for a few whom they sent back to Thorfinn with an insulting message of defiance. Deciding to seek revenge the following summer, the earl returned to Orkney for the winter, and early in the spring raised a levy throughout his earldom. He also sent word to his nephew Rǫgnvaldr, asking him for help, and Rǫgnvaldr duly mustered troops from his lands. Thorfinn now had an army from Orkney and Caithness, and ‘large numbers of Scots and Irish flocked to join him, as well as a good many from all over the Hebrides’. Leading this force against the English, the earls began pillaging and looting. ‘The commanders in charge of England’s defences’ in Harthacnut’s absence marched their troops against them, and there was ‘a hard, fierce battle’, which the earls of Orkney won. After that, ‘they ravaged far and wide throughout England’, killing, looting and burning.

Like other sagas, as mentioned above, Orkneyinga saga includes skaldic verses at intervals to corroborate stories such as this one. To the saga-author and audience, the verses functioned rather like a footnote, referencing the authoritative testimony of poets who were alive at the time of the events narrated in the saga. Having recounted the tale of Thorfinn’s victory, the saga-author introduced stanzas to support his claims, prefixing them with the formula ‘as Arnórr says’. ⁷ The most complete witness to the text of Orkneyinga saga, in Flateyjarbók (GKS 1005 fol., mainly c. 1387–95), attaches only two stanzas to this battle. ⁸ Another work, Uppsala universitetsbibliotek R 702, which is a collection of verses from different sources, adds a third, which Whaley includes after the other two in her reconstructed text of Þorfdr, but with the caveat that ‘it is not possible to tell whether it [i.e. the third stanza] was included in the original saga’. ⁹ In Whaley’s edition(s) of Þorfdr the stanzas are numbered 16, 17 and 18, the first two being the ones associated with this battle episode in Flateyjarbók.

The three stanzas, 16, 17 and 18, are given below, as reconstructed and translated by Whaley (with meanings of the kennings inserted in the brackets in capitals):

Then came the edge-blizzard [BATTLE] which the English remember, and never after will a lofty ring-strewer [GENEROUS RULER] come there with a larger force. Slender-wrought swords bit the mighty troops beneath their shields, and the descendant of Rǫgnvaldr inn gamli (‘the Old’) [= Thorfinn] rushed forth there south of Man. ¹⁰

The jarl bore his standard onto the native soil of the English, and his retinue reddened at once the eagle’s tongue; the leader called for banners to advance. Flame grew; halls collapsed; the war-band drove [men] to flight there; the foe of branches [FIRE] flung out smoke and hurled light close to the sky. ¹¹

Many a blast of horns sounded between the defences, where the banner of the stout-hearted hero waved; the bountiful one stormed into battle. Not a trace of fear seized the grim troop of the thief-feller [JUST RULER = Thorfinn], once it seemed light enough for battle; iron blades quivered, and wolves chewed morning carrion. ¹²

After having cited at least the first two of these stanzas (if not necessarily the third), the saga-author added that Thorfinn ‘fought two pitched battles in England, along with many more sorties and killings’, before returning to Orkney in the autumn. ¹³

Scholars working with saga material have long been aware that the verses cited in such cases may have been all the saga-writers had to go on. Some episodes in the sagas would have been fabricated from clues which the verses offered, while other episodes were embellished either with half-remembered traditions or with pseudo-historical window dressing, borrowed from written sources. Roberta Frank cites the case of a saga-author who took the compound noun skeiðarbrandr (‘prow of the ship’), which he came across in an anonymous stanza, to be the name and epithet of a sea-king called ‘Ship-Brandr’. Having conjured this ghost onto the historical stage by his error, he duly furnished him with brief biographical details. ¹⁴ Other fabrications doubtless escape exposure for want of evidence to reveal the fiction, for most claims made in the sagas cannot be checked. Even so, there is evidence of the ‘notoriously inaccurate dating’ to be found in Orkneyinga saga, ¹⁵ and it has rightly been said that in the case of figures such as Thorfinn, who were remote from the date of the saga’s composition, ‘the possibilities for confusion … were great.’ ¹⁶ Historians, increasingly, are making it their default position to doubt whichever details in the sagas cannot be corroborated (an approach displayed with admirable consistency, most recently, in the work of Neil McGuigan), ¹⁷ but this approach has its difficulties for the editor of skaldic poetry, given that challenging the veracity of the prose context destabilizes a principal basis for reconstructing the poems. ¹⁸

In the past, in a largely neglected article of 2003, only the late Kari Gade has called the story of Thorfinn’s English campaign into question. ¹⁹ Two reasons she gave for doubting it were the lack of any record in English, Welsh, or Irish sources of an attack of that sort and magnitude within the broad timeframe of the 1030s and 1040s, and the lack of support in the poetry for the saga-author’s claim that Rǫgnvaldr participated in the raid. ²⁰ Neither objection is fatal of itself. Raiding expeditions occasionally did go unreported, it would seem, ²¹ and Rǫgnvaldr’s participation may have been referenced in other verses, known to the saga-author but lost to posterity. Still, the tale of Thorfinn’s campaign has fabulous qualities – a classic tale of defeat, followed by plans for revenge, and, ultimately, victory (one thinks of Alfred the Great and Robert the Bruce), in which an earl reported to be mighty meets his match at the farthest edge of his dominion and earns his reputation there by overcoming the mighty English. Swept along by the story, the audience is conditioned to believe that a warlord whose men had been wiped out in a failed cattle raid was able to attract large numbers of Orcadians, Scots, Irish, and men of the Hebrides to join him on a repeat expedition. Typically, military failure might be thought to have the opposite effect.

This unlikely tale – a fable even – could well be a fantasy of Orcadian might from the post-viking era, when men from Orkney still mounted occasional cattle raids on the shores of the Irish Sea but could only dream of visiting revenge upon defenders who opposed them. Gade, however, had a more compelling reason for rejecting the tale’s historicity or setting, and that was the match she found between the events described by Arnórr (together with certain details in the prose context) and a recorded attack on England in 1058.

As Gade noted, ²² the attack on England in 1058 is reported in English, Irish and Welsh sources. While no account gives all the details which we should like to have, the various records between them build a picture of what happened. From the English perspective, the story of 1058 starts in 1055 with the English earl, Ælfgar. For unclear reasons, in 1055, he was outlawed and driven from the realm. After recruiting eighteen ships of Irish-Scandinavian pirates, he joined forces with Gruffudd ap Llywelyn, king of Gwynedd. Together, they invaded England across the Wye and attacked the royal town of Hereford, precipitating negotiations which brought about Ælfgar’s restoration. After the campaign, the fleet of Norse-Irish mercenaries went to Chester, the sea-port of Mercia, to collect their pay. ²³ Ælfgar’s father, Earl Leofric, held the Mercian earldom at that time, but, upon Leofric’s death, in 1057, Ælfgar was translated from his own earldom to Mercia. In 1058, he was expelled from England a second time and again made his way to Gruffudd, his old ally.

While Ælfgar and Gruffudd were planning their next moves, a viking fleet, described by the English chronicler as a ‘raiding ship-army from Norway’, came into the Irish Sea. ²⁴ An Irish annalist keeping a chronicle at the monastery of Clonmacnoise regarded the fleet’s arrival as a development of sufficient importance to record that it was led by the (unnamed) son of the king of Norway, and that it comprised ‘the Gaill [i.e. foreigners] of Orkney, the Hebrides, and Dublin’. ²⁵ Welsh sources identify the fleet’s commander as Magnus (Magnús), son of Harald Hardrada (Haraldr Sigurðarson), and state that he ‘laid waste a region of the English with the help of Gruffudd, king of the Britons [i.e. Welsh]’. ²⁶ John of Worcester, who used a lost version of the Anglo-Saxon Chronicle with similarities to the extant C and D texts, echoes their testimony in his statement that Ælfgar recovered his earldom in 1058 with the help of Gruffudd ‘and the support of a Norwegian fleet’ which – John adds – ‘joined him unexpectedly’. ²⁷ The D text of the Chronicle, which mentions the ‘raiding ship-army’, says that Ælfgar ‘came back again, with violence, through the help of Gruffudd’, and adds that the raiding fleet ‘came here’ [to England], implying an attack. ²⁸ The Irish annalist thought it had come ‘to seize the kingdom of England, but God consented not to this’, possibly hinting at perceived failure in the outcome of the viking expedition. ²⁹

An additional piece of evidence, not noted by Gade, indicates which ‘region of the English’ suffered the attack. A record in Domesday Book states that King Edward the Confessor had given Gruffudd all the land that lay across the river Dee but had later taken the land back after Gruffudd wronged him. ³⁰ In the past it was thought that the gift might have been made as a settlement following the campaigns of 1055–6, which began with Gruffudd and Ælfgar’s joint attack upon Hereford. The corollary of this argument was that Edward revoked the gift after Gruffudd’s second attack upon England with Ælfgar in 1058. ³¹ It is now thought to be more likely that the gift of land across the Dee was made as a settlement following the 1058 campaign and revoked in 1063 when Gruffudd mounted a third and final invasion. ³² The reason for linking this gift of land to the 1058 campaign and not the earlier one is twofold. First, the territory targeted by Gruffudd and Ælfgar in 1055 lay in the region of Hereford, and any land granted to the Welsh king in settlement at the treaty of Billingsley that year most likely lay in the adjacent region of Archenfield, rather than to the north in Cheshire. ³³ Second, Cheshire lay in Earl Leofric’s earldom in 1055, only passing to his son Ælfgar in 1057, and the greater likelihood is that it was the son who was forced to part with a portion of it, for Edward’s gift to Gruffudd, as the price for his rebellion.

It stands to reason that a campaign combining the northern Welsh forces of Gruffudd and a viking fleet in the Irish Sea would have attacked England in that self-same area, approaching up the Dee. In 1055, Gruffudd and Ælfgar had targeted a major town, namely Hereford. In 1058, the obvious target was the walled town of Chester or its adjacent suburbs. Chester was a rich and cosmopolitan port. At that date, shipping could sail up the Dee to the city walls. ³⁴ Trading links to Norway and Ireland, coupled with Norse and Irish settlement, ensured that Chester was already well known to the attacking force. The surrounding area of the Wirral peninsula and the Dee valley afforded further hope for a raiding army to feed itself, since it comprised much the best farming land in the county. ³⁵ It was probably this part of Cheshire that had suffered a viking attack in 980, though it is not clear whether the raiders had come from Norway, the Northern Isles, or the shores of the Irish Sea. ³⁶

Gade rightly noted the parallel between the large composite force ascribed to Thorfinn in the saga account (comprising a host from Orkney and Caithness, with troops also from Scotland, Ireland and the Hebrides) and the large composite force described by the contemporary Irish annalist, who mentions participants in the 1058 campaign from Orkney, the Hebrides and Dublin. ‘It is striking’, she says, ‘that the constituency of the 1058 fleet is almost exactly the same as that of the naval force he allegedly summoned for the attack on England’ which the saga-author describes. ³⁷ This may be an instance of an authentic detail that had been handed down and preserved in the medley of traditions and fictional elements. It is no obstacle to Gade’s case that Arnórr’s verses make no mention of the Dubliners, Hebrideans, or Welsh, since the participation of allies was seldom referenced in accounts which were designed to concentrate praise on one ruler or one people. Neither the Welsh nor the Irish sources for 1058 mention Ælfgar; nor does the Irish annalist mention Gruffudd; ³⁸ and Welsh sources ignore Ælfgar in their descriptions of Gruffudd’s campaign with him in 1055. ³⁹ A further point to be made is that Arnórr’s assertion that Thorfinn’s forces attacked England ‘south of [the Isle of] Man’ fits very well, geographically speaking, with an attack in the Chester area, where Edward gave land to Gruffudd, extending the latter’s holdings eastwards to the Dee.

Having linked Arnórr’s verses to the Norwegian expedition of 1058, Gade speculated about its aims (on which there will be some comment later). Though differing in their opinions on why it might have been launched, one point on which scholars agree is the strangeness of this expedition not getting a mention in the sagas. Judith Jesch noted, in 1993, that ‘there is no indication in Orkneyinga saga, or any Icelandic source, that Thorfinn was involved in this [1058] raid on England, indeed it is not even mentioned’. ⁴⁰ ‘Curiously’, Alex Woolf remarked in 2007, ‘Magnus Haraldsson’s expedition to the west is completely ignored by the Icelandic saga tradition. There is absolutely no mention of it in either Orkneyinga saga or any of the sagas of Norwegian kings.’ In Woolf’s view, ‘the fact that the sagas were able to completely ‘forget’ such a major event, the first reliably attested Norwegian royal expedition to the British Isles, serves as a warning about their reliability in general’. ⁴¹ William Thomson in 2012 and Barbara Crawford in 2013 also noted the strange omission. ⁴² Most recently, Neil McGuigan (in 2021) has repeated the point that the fleet of 1058 ‘seems to be unknown to saga tradition’, ‘damningly’ for that tradition, he adds. ⁴³

Woolf, McGuigan and others are right, of course. It is an indictment of the sagas’ pretensions to having captured at least the most significant events for posterity, that none of them refers to the Norwegian expedition of 1058. It is equally strange that none of the scholars debating this topic cites, or seems to have been aware of, Gade’s 2003 article, and its argument that a corrupted and mislocated account of the expedition is concealed in Orkneyinga saga’s story of Norse attacks on England, supposedly in the time of Harthacnut. Her argument should have served as a stimulus to future discussion, putting historians on the scent. Instead, it seems to have dropped off the radar, rather like the fleet that sailed to England in 1058.

On the strength of the case she had built in that paper, Gade concluded that ‘the order of the stanzas [in Þorfdr] as presented by the standard editions of skaldic poetry (which follow the chronology of events in Orkneyinga saga) is incorrect’. ⁴⁴ For Gade, this meant that the three stanzas connected to Thorfinn’s English campaign, stanzas 16–18, ‘ought to be placed after stanzas 19–20, which detail the battle of Rauðabjǫrg in 1044’. ⁴⁵ Apparently, Diana Whaley disagreed, for the introduction and discussion accompanying her edition of Þorfdr for the corpus of Skaldic Poetry makes no mention of Gade’s article or the argument it advances, despite Whaley’s new edition having appeared in 2009, six years after the publication of Gade’s case, and in a volume under Gade’s general editorship. In her introduction to this edition, Whaley reasserts that ‘there is no reason to challenge the chronology of the saga, which, together with the internal evidence of the stanzas, is the main evidence available’. ⁴⁶ The stanzas thus remain in the order in which they appear in her DPhil thesis of 1979, and likewise in her 1998 edition, notwithstanding the case Gade had since put forward.

Whaley’s decision to preserve the status quo is not without implications in the matter of how these verses are to be reconstructed and understood. By choosing to accept the outline of the narrative presented in Orkneyinga saga, Whaley was inclining against readings that were at odds with the prose context. One such reading had been suggested by Hermann Pálsson and Paul Edwards in their translation of Finnbogi Guðmundsson’s edition of the saga, for those two translators had taken the first of the relevant stanzas (Whaley’s stanza 16) to give the sense that the Orcadians were ‘cut down’ in the battle. ⁴⁷ Whaley noted, however, that the kenning which they apparently took to refer to men of Orkney collectively – ‘Rǫgnvalds kind … ins gamla’ (the child/descendant of Rǫgnvaldr the Old) – appears to refer to the leader Thorfinn himself, and that this fact problematised any reading implying that the referent was ‘cut down’, since the saga claimed that Thorfinn triumphed in the battle, surviving many more years before dying in peaceful retirement. ⁴⁸

The troublesome passage occurs in the second helmingr of stanza 16:

Bitu sverð, en þar þurði,

Þunngor, fyr Mǫn sunnan

Rǫgnvalds kind, und randir

ramlig folk, ins gamla. ⁴⁹

Whaley comments:

The role of the noun phrases ramlig(t) folk ‘mighty troops’ and kind Rǫgnvalds ‘the descendant of Rǫgnvaldr’ is ambiguous: each could be the nom[inative], the subject of þurðu/þurði ‘rushed’, or acc[usative], object of bitu sverð ‘swords bit’. ⁵⁰

Finnbogi Guðmundsson had struggled with the same problem and sought to get around it with the suggestion that kind was dative, producing the reading ‘swords bit for Rǫgnvaldr’s descendant’ (from ‘bitu sverð kind Rƍgnvalds’), but Whaley rightly remarks that ‘there is nothing in the syntax to show that kind is not acc[usative singular], which would yield the sense that the hero was wounded’ [sic]. ⁵¹

Whaley also discusses four key manuscripts which preserve stanza 16 or its second helmingr. These manuscripts are Flateyjarbók (Flat), Codex Trajectinus (Utrecht, University Library, MSS 1374 (T^x)), Codex Upsaliensis (Uppsala, University Library, De la Gardie 11 (U)), and Codex Wormianus (Copenhagen, Arnamagnæan Collection, AM 242 fol. (W)). ⁵² She comments:

it appears that the scribes of mss T^x and W, and probably of U and Flat, took ‘mighty troop’ as the subject of ‘rushed’, since the two mss (T^x and W) which read pl[ural] þurðu also have pl[ural] ramlig folk, while the two which read [singular] þurði also have [singular] ramligt folk. ⁵³

Despite the combined evidence of those scribal readings, which would make the mighty troop/s the subject of ‘rushed’, and Rǫgnvald’s kind (i.e. Thorfinn) the object bitten by swords, Whaley concluded that ‘the hero Þorfinnr is more likely to be depicted rushing forward than being pierced by swords’, and she therefore adopted singular þurði in her edition of the verse, with the kind-phrase as its subject. ⁵⁴ There is a slight textual basis for Whaley’s rejection of these four scribal readings, in that the manuscript Uppsala, University Library, R 702 (s. xvii^1/2) supplies the variant þorði (singular, ‘dared’) with plural ramlig, ‘which could make sense’, but, as she notes, ‘the scribal evidence is against it’. ⁵⁵ Alternatively, þorði could be taken as a transcription error for þurði in support of Whaley’s reconstructed reading.

Rejecting the weight of scribal evidence in the manuscript witnesses, Whaley preferred to reconstruct the text in this instance because it did not seem likely to her that the poet was describing injuries to his hero. Having restated her position, that there was ‘no reason to challenge the chronology of the saga’, she began from the premise that the stanzas were describing a campaign early in Thorfinn’s career, from which he emerged victorious. ⁵⁶ Whaley’s epistemological premises for reconstructing Þorfdr, where evidence in other sources was lacking, relied on two kinds of evidence: ‘the chronology of the saga’ and ‘internal evidence of the stanzas’. ⁵⁷ Yet there is now mounting evidence, at least in this instance, that something is amiss. Not only is the chronology likely to be wrong, with Thorfinn’s attack on England having been mislocated; there is also a problem with the internal evidence of stanza 16, which – on the weight of the scribal readings – has the protagonist being cut by swords. If we take ‘mighty troop/s’ as the subject of ‘rushed’, according to the scribal evidence, the second helmingr may be translated as follows:

‘The mighty troop/s rushed forth there beneath their shields, south of Man, and slender-wrought swords bit the child/descendant of Rǫgnvaldr the Old’.

Now Arnórr uses various kennings to conjure up Thorfinn in his poem. This is the only one that links him to his ancestor (called ‘the Old’ to distinguish him from Thorfinn’s nephew). The fact that Thorfinn, in this one stanza alone, is linked to Rǫgnvaldr ins gamla reflects a unique decision by the poet to highlight their connexion, specifically in the context of this battle (or more specifically its conclusion, since the kenning wraps up the stanza). In chapter 3 of Orkneyinga saga, Rǫgnvaldr, earl of Møre, as he is styled, makes his appearance as ‘inn riki ok inn ráðsvinni’ (‘the powerful and wise-counselled’). He joins Harald Fairhair (Haraldr hárfagri) on an expedition to the Irish Sea, subduing Shetland, Orkney, and the Hebrides on the way, before raiding the Isle of Man. There, Ivarr, his son, is killed in the fighting, and, in compensation, Harald Fairhair gives him Orkney and Shetland, thereby making Rǫgnvaldr the first earl of Orkney (though he relinquishes the islands into the hands of his brother Sigurd). For an audience familiar with the story, Arnórr’s references to Thorfinn as the ‘child/descendant’ of Rǫgnvaldr and to a troop rushing forth ‘south of Man’ would have conjured up the Man expedition from their origin myth, and the death in battle of Rǫgnvaldr’s son, Ivarr.

The parallel is too cleverly constructed not to be deliberate, and its purpose would appear to extend beyond signalling the fact that Ivarr and Thorfinn both fought battles near Man. For when taken together with the ostensible statement that Thorfinn was bitten by swords, the poet’s message is clear. What Arnórr is doing is describing Thorfinn’s death in battle, and honouring his end by making an obvious connexion with the death of Rǫgnvaldr’s earlier kind, Thorfinn’s prototype, Ivarr, similarly slain on an expedition to Man. The resonance proceeds from the double meaning of kind, as both ‘child’ and ‘descendant’. Thus while Arnórr speaks of Thorfinn, he simultaneously evokes Ivarr.

Pursuing this thread of ambiguity, the dramatic elements in the first helmingr of the stanza can be read as preparing the audience for the end. It begins by introducing the fight as the battle ‘which the English remember’, a motif which Arnórr employs in another memorial drápa to imply a victory for its subject (in that instance, Magnus of Norway, crushing the Wends of Wollin), ⁵⁸ but one which also allows the possibility that the battle might just as likely have been memorable to the English because they won. Arnórr’s next remark is as finely balanced: ‘and never after will a lofty ring-strewer come there with a larger force’. Here he speaks with two voices, one celebrating the peak of his subject’s heroic career, another tinged with nostalgia (or even sorrow), that such glories will never come again. Suspending the release of emotion for an audience who knew how it would end, Arnórr completed his stanza with the troops’ valiant charge and Thorfinn’s death in the sword-blizzard. The audience, his drótt, is reminded, in his death, of the sacrifice of Ivarr, his ancestor – a sacrifice linked, at least in later legend, to the acquisition of Orkney itself.

Such are the double grounds for interpreting stanza 16 as an account of Thorfinn’s death, a reading which explains why its hero is bitten by swords, and which upholds Gade’s case for linking the stanza to the campaign of 1058. Further to this, the present ordering of stanzas 16 and 17 makes no sense and should be reversed. Stanza 17 begins with the statement that ‘the jarl bore his standard onto the native soil of the English, and the retinue reddened at once the eagle’s tongue’. Thorfinn then calls for his banners to advance. Stanza 16 begins, ‘Then came the [battle] which the English remember’. The more logical sequence is that the earl begins raiding England, bearing his standard onto its soil (stanza 17), before fighting a memorable battle (stanza 16). As well as being more logical, such a sequence would also better fit with his death at the end, and it should override any concern an editor might have about keeping the stanzas in the order in which the saga-author put them. Stanza 18 may not belong here at all. Unlike the other two stanzas, it makes no mention of England or the English, and it refers to ‘defences’ (i.e. fortifications of some sort) which are unaccounted for and unanticipated in the other verses and in the prose. Stanza 18 is not in Flateyjarbók (which nevertheless has stanzas 16 and 17 together). Nor is it in the verse collections in the other manuscripts listed above. The only significant manuscript in which the stanza appears is R 702. ⁵⁹ Although Whaley did include the stanza in her edition(s) of Þorfdr, she adds the caveat, ‘it is not possible to tell whether it was included in the original saga’. ⁶⁰

The conclusion now emerging is radically different from the saga story of Thorfinn’s career. According to the saga, the earl got his fighting out of the way in the 1020s–40s, went on a pilgrimage to Rome c. 1050 to expiate his sins, returned to Orkney, founded a church, and settled down quietly to live out his days. His story arc thus assumes the moral and didactic pattern in which sin was to be atoned for by a penitential journey and the endowment of a house of God. The truth was not at all like that. Like other war-leaders of his age, the earl continued campaigning to the end and met his death in battle. Either this history was soon forgotten (and the verses were misinterpreted), or a different story was preferred, or both processes occurred. According to the saga, Thorfinn died ‘towards the end of the reign of Harald Hardrada [i.e. Haraldr Sigurðarson – a claim which has led scholars to assign his death to c. 1064–5]’. ⁶¹ The evidence discussed in this paper is not incompatible with that statement, but it also indicates that he died in battle in the Chester area in 1058. ⁶²

Arnórr – ironically, as it may seem to us today – describes the fight as the battle ‘which the English remember’. This was, of course, a motif, but it is worth noting in passing that even memorable battles sometimes failed to register in the chronicles. A battle, immortalized by Shakespeare, in which an English royal army led by Earl Siward defeated the forces of the Scottish king Macbeth, in 1054, was of no interest to the compiler of the E text (although it appeared in parent or sibling chronicles which he must have seen). ⁶³ Rather, he devoted his annal for 1054 to recording a change of popes and a great pestilence among cattle. A fight between the forces of William the Conqueror and the English at sea, in 1066, is known to historians only from a chance notice in Domesday Book and a record in a chronicle from Niederalteich in Bavaria (of all places). ⁶⁴ There is no mention of it in English chronicles. Still, the D chronicler knew a lot more detail about the events of 1058 than he cared to divulge, for he breaks off with the tantalising comment: ‘it is tedious to tell how it all happened.’ ⁶⁵ It is not known whom King Edward had put in charge of the Mercian earldom, upon the occasion of Ælfgar’s second exile, and it is therefore difficult to guess which commanders the invaders faced. Since Chester lay at the end of Watling Street, a central artery of the road system since Roman times, troops might have been deployed there to oppose this considerable threat from almost anywhere in the realm. The Irish annalist’s opinion that the viking fleet had come ‘to seize the kingdom of England’, and that God ‘consented not to this’, favours the idea that he had heard report of an English victory, referencing God’s part in the affair to signify divine adjudication against the invaders, presumably by the verdict of battle. ⁶⁶ Yet if Edward beat off the vikings, he still had to accommodate Gruffudd, hence the gift of land across the Dee. It is possible at least to perceive how the saga of 1058 might have seemed tedious and complex to an annalist, concerned more with other matters.

Assigning Thorfinn’s death to 1058 is not without implications for Scottish history, as it rules out several theories about the careers of Malcolm (Máel Coluim) III Canmore and Ingibjorg, his wife, who was also Thorfinn’s widow. For a start, Thorfinn’s death on the expedition of 1058 rules outs the theory that the dispatching of a Norwegian fleet to Orkney was an intervention in a power vacuum which the earl’s death had created, ⁶⁷ and likewise the argument that Malcolm married the earl’s widow, Ingibjorg, before taking the throne of Scotland that same spring or summer (since Thorfinn would have still been alive, unless Malcolm snatched her up very quickly after his death). ⁶⁸ Similarly, there would now be difficulties with the suggestion that a Norwegian force went to Orkney in 1058 to assist Malcolm in an invasion bid which he is sometimes thought to have launched from that location, with or without Ingibjorg’s assistance. ⁶⁹ Thomson’s theory that it was, in fact, Ingibjorg’s death (assigned to 1058 in his argument) that prompted the Norwegians to dispatch a fleet to Orkney also cannot co-exist with Thorfinn’s presence on that joint campaign, since the theory depends on Thorfinn having died at an earlier date. ⁷⁰

As for Ingibjorg, historians remain ambivalent, being wary of accepting the solitary late claim that she married Malcolm Canmore after Thorfinn’s death, and yet disinclined to abandon it, for want of better evidence. ⁷¹ If the marriage is accepted, Thorfinn’s survival to 1058 need not be problematic. As Thomson has observed, Ingibjorg was most likely born in the early 1030s, married to Thorfinn in the mid-late 1040s, and bearing his children – at least two sons – before the decade was out. ⁷² She should still have been capable, then, of bearing children to Malcolm through the 1060s, when she would have been in her thirties, before the king of Alba took Margaret for a wife at the tail-end of that decade. ⁷³ Fixing Thorfinn’s death in 1058 still allows a span of years in which Ingibjorg might have borne children to Malcolm, including Duncan, who did not necessarily have to have been very old in 1072 when he was taken hostage by William the Conqueror. ⁷⁴

All of these theories have been tied up with discussion of the possible aims of the Norwegian expedition. Comparing several such expeditions, Caitlin Ellis found the thread linking them to be a desire ‘to capitalise on political weakness in the Insular world’ (meaning Britain, Ireland, and the surrounding islands). ⁷⁵ This theory could certainly apply in the case of Magnus’s expedition of 1058, since there has been a return to the traditional chronology surrounding Malcolm’s campaign for the Scottish throne, and to the view that his enemy Macbeth died in 1057 (after many years on the throne). ⁷⁶ So the Norwegians and men of Orkney might have thought to exploit a phase of upheaval. While this might not explain Magnus’s fleet ending up in the Irish Sea, there may, for all we know, have been trouble around Man and the Hebrides, ⁷⁷ or word may have reached them that war was brewing between the English king and his Welsh neighbour.

One key point in understanding the 1058 expedition is the fact that its titular leader Magnus was a boy: at most a teenager, on the cusp of his majority. ⁷⁸ The question therefore arises as to who provided the strategic and military leadership behind him. Two candidates that have been suggested are his father’s marshal, Ulf Ospakson (Úlfr Óspaksson) and Earl Thorfinn himself. ⁷⁹ Neither can be ruled out. Yet there may be a more salient observation to make in reference to Magnus’s (or Harald Hardrada’s intentions), in that the heroic tradition, as mirrored in skaldic praise poems, required a would-be ruler to go on his first campaign at an early age. ⁸⁰ Hardrada, in other words, may have entrusted his young heir to a capable commander with the intention that Magnus should prove his mettle.

A final word is needed on Arnórr and his poem, and the middle-to-later stages of his career. Prior to the concluding eulogy and prayers (found in stanzas 23–5), it now appears likely that Þorfdr’s last episode concerned the earl’s final campaign, to England, and his death in battle (as described in stanzas 17 and 16, ordered that way around). The mystery of why Þorfdr ‘does not specifically describe any events from the last twenty years of Thorfinn’s life’ ⁸¹ is solved by the discovery that, in fact, it does. Arnórr does not indicate – at least in surviving verses – that he was present on the earl’s last campaign, but it is a fair guess that Þorfdr was composed in or shortly after 1058, for Thorfinn’s hearth-troop in Orkney. The celebrated earl was survived by his young sons, Paul (Pál) and Erlend (Erlendr), who both followed in their father’s footsteps by joining the next Norwegian expedition which used Orkney as its staging point. That was of course Harald Hardrada’s expedition to England, eight years later in 1066.

Hardrada’s expedition, like Thorfinn’s, ended in tragedy and the composition of a drápa by Arnórr to commemorate its fallen leader. According to Icelandic tradition, Hardrada’s son Olaf (Óláfr), with other survivors, who included Paul and Erlend, returned to Orkney and to Hardrada’s widow Elisabeth (Ellisif) – who had come with him to the island – and wintered there before leaving for Norway in the spring. ⁸² It may have been for that circle, in Orkney, rather than in Norway, that Arnórr composed his Haraldsdrápa in the autumn or winter of 1066. ⁸³ Such are the grounds for placing Arnórr quite possibly in Orkney in 1058 and 1066, though he may have travelled, as he seems to have composed verses for Icelanders too. ⁸⁴

Earlier, I noted the parallel between the failure of Gade’s article of 2003 to leave an imprint in the scholarship, and the failure of the 1058 expedition to leave any recognisable imprint in the sagas. Its imprint was there in disguise, as Gade pointed out, but her argument undercut all sorts of foundations, not least the integrity of a saga narrative and the editorial methods underpinning the reconstruction of fragmented poems. The 1058 campaign might similarly have been forgotten for the reason that it undermined the Norwegians’ or Orcadians’ ideas of their superiority in war. If the campaign on which the young Magnus was supposed to prove his credentials ended in defeat and the death of one of its leaders, it could be conveniently forgotten without leaving a scar in the history books, unlike Hardrada’s campaign of 1066, which resulted in the death of a king and the destruction of his army in detail. Whether or not this was the reason, the story behind the relevant verses had fallen into obscurity by the late twelfth century and, with it, the details of Thorfinn’s death in a long-forgotten battle.