Haraldskvæði ‘Poem about Haraldr’ (Þhorn Harkv), traditionally dated c. 900, is the earliest of the three early eddic praise-poems in the skaldic corpus, the others being the anonymous Eiríksmál (Anon Eirm) and Eyvindr skáldaspillir’s Hákonarmál (Eyv Hák), though see also the anonymous Darraðarljóð (Anon DarrV (Nj 53-63)). The title Haraldskvæði and the alternative Hrafnsmál ‘Words of the Raven’ are both modern, and indeed, there is no certainty that these twenty-three stanzas formed parts of a single composition. The poem was first assembled from the ms. sources in Munch and Unger (1847, 111-14). The metre of the stanzas is one unifying consideration, for they are composed in málaháttr and ljóðaháttr (see below for details), while Þorbjǫrn’s other known compositions are in dróttkvætt. Moreover, most of the present stanzas can be interpreted as part of a dialogue between a valkyrie and a raven. In the poem as presented below, sts 1-2 introduce the two speakers, after which the valkyrie asks questions in sts 3, 15, 18, 20 and 22, and the raven’s replies constitute the remainder of the poem. Yet the sources do not present the stanzas as a unified poem, nor do they agree about authorship.
The stanzas celebrate King Haraldr hárfagri ‘Fair-hair’ Hálfdanarson (r. c. 860-c. 932; see ‘Ruler biographies’ in Introduction to this volume), and fall into three groups, covering: introduction and Haraldr’s court, both warriors and entertainers (sts 1-6 and 15-23); his role in the battle of Hafrsfjǫrðr (Hafrsfjorden) c. 885-c. 890 (sts 7-12, though st. 12 could be unrelated); and his marriage to Ragnhildr (sts 13-14). Fsk is our only source for most of the first group, and those stanzas are all quoted together there, though brief prose links follow sts 6, 19 and 21. These stanzas must stem from a single composition (though cf. Vogt 1930a, 184), especially as sts 15-23 retain the question-and-answer format established in sts 3-6, and the address to the raven in st. 20 shows that the speakers are the same. Stanzas 7-12 are linked to the stanzas about Haraldr’s court by their metre and by being devoted to praise of Haraldr, as well as by the fact that (as noted by Guðbrandur Vigfússon, CPB I, 255), st. 7 begins with a question, presumably posed to the valkyrie who continually interrogates the raven or ravens of sts 3-6 and 15-23. Finnur Jónsson (LH I, 428), with several others, notes that the promise in the first stanza to tell of íþróttir odda ‘feats of weapon-points’ goes unfulfilled if sts 7-12 are extraneous. He also remarks that it would be unlikely, and without skaldic parallel, that Þorbjǫrn should have composed two poems, both in málaháttr and both in the same dialogic form between the same ‘persons’. Stanza 12 is in fact ascribed by Snorri in SnE not to Þorbjǫrn but to Þjóðólfr ór Hvini (Þjóð), but the closing line of this helmingr describing a great slaughter, fǫgnuðum dôð slíkri ‘we welcomed such doings’, seems particularly appropriate if spoken by a raven (as remarked by Kershaw 1922, 78), providing another reason for assigning it to Harkv. The third group has smaller claim to unity with the rest. Fidjestøl (1993d, 669) says that they are ‘possibly out of place in this context’; Flo (1902, 69) says that if they belong to this poem, there must have been more stanzas on the topic of the women mentioned in these stanzas; Sueti (1884, 13-14) orders st. 14 after st. 6 and regards st. 13 as a lausavísa. These stanzas are composed in málaháttr, and st. 14 is attributed to Þorbjǫrn hornklofi in Hkr but to Þjóðólfr in Flat, where another version also appears, ascribed to Þorbjǫrn.
The evidence for the unity of all twenty-three stanzas is thus inconclusive, and they have not always been regarded as a unit. For example, de Vries (1964-7, I, 137-9, supported by Harris 1985, 97) argues that Þorbjǫrn later added sts 13-23 to an earlier composition of his, perhaps on the occasion of Haraldr’s marriage to Ragnhildr (see also Olsen 1942b, 30, though Genzmer 1926, 126-7 supposes only st. 13 was composed on that occasion, and st. 14 is a later poet’s addition). Von See (1961b, supported by Weber 1967a, 1467, Fidjestøl 1976c, 9, Fidjestøl 1982, 31 and Ehrhardt 2002b) argues that the same stanzas were composed by an incompetent imitator in the early twelfth century, influenced by Atlamál. Jón Sigurðsson (SnE 1848-87, III, 409) regards the stanzas on Hafrsfjǫrðr as having been composed immediately after the battle, and the rest later; see also Sueti (1884, 12-16) and Boyer (1990a, 193-4). Metcalfe (1880, 383-5) and Larsen (1943-6, II, 246-50) treat the stanzas on Hafrsfjǫrðr, excluding st. 12, as one poem and the remainder, excluding sts 13 and 14, as another. By contrast, Reichardt (1926) and Wolff (1952) see here a unified composition, as does Finnur Jónsson (LH I, 430), supposing the whole was composed as much as ten to twenty years after the battle of Hafrsfjǫrðr. See also Würth (1999). Scholarly views are thus widely divergent, and especially since there is nothing like a consensus about an alternative arrangement, less inconvenience to scholars will result if all twenty-three stanzas are treated here as a single work and numbered in accordance with Skj and Skald. Regardless, even if the stanzas all belong to one composition, it is certainly only a fragment (Noreen 1926, 163).
The evidence for the authorship of the stanzas is, as already indicated, of a conflicting nature. All but sts 12-14 are preserved in Fsk, in both the A and B versions, though not as a unitary work but in groups of stanzas. There sts 1-6 and 15-19 are explicitly attributed to Þorbjǫrn hornklofi (as is st. 6 in Hkr), but sts 7-11, on the battle of Hafrsfjǫrðr, to Þjóðólfr ór Hvini, as they are in Flat, where sts 13 and 14 are also called his (though, as remarked above, another version of st. 14 in Flat is assigned to Þorbjǫrn). Stanzas 7-11 and 14 are, however, ascribed to Þorbjǫrn in Hkr. Note that while Snorri attributes st. 11 to Þorbjǫrn in Hkr, in SnE (where only the first helmingr is quoted) he calls it Þjóðólfr’s, along with st. 12. To add to the uncertainty, a version of st. 21 in Flat is called the work of Auðunn illskælda (Auðunn). Fidjestøl (1976c, 17-19), departing from his earlier view, reasons that since Snorri made a practice of analysing his sources and drawing his own conclusions, while the compiler of Fsk tended to take his sources at their word, the attribution to Þjóðólfr is more likely to reflect actual oral tradition. Holm-Olsen (1974, 227) also argues for Þjóðólfr’s authorship of the stanzas about Hafrsfjorðr. See also the Notes to sts 9/3 and 10/1.
In many respects, Harkv is more reminiscent of eddic than of skaldic poetry. Genzmer (1920, 149-54) demonstrates that this is so in regard to metre, vocabulary, syntax, and both (in)frequency and (non-)obscurity of kennings, as well as to the poem’s dialogic form and narrative progress. See also Lie (1957, 80-2) for a stylistic analysis. Þorbjǫrn’s other poetry is quite unlike this. Harkv in fact bears such a resemblance to the eddic Atlakviða in these respects that Genzmer (1926, supported by Reichardt 1926 and von See 1961a, 315) would identify Þorbjǫrn as the creator of both. He concludes that a form of panegyric survives in this poem that resembles, more closely than later compositions, what the early Germanic Preislied ‘praise-poem’ must have been like (see also de Vries 1964-7, I, 139). Yet Wolff (1939, 30) argues convincingly that the form has nothing to do with the prehistory of the praise-poem as a genre: see Introduction to Eyv Hák. Kuhn (1939, 220-5) argues that the poem shows the metrical influence of West Germanic verse, and that Þorbjǫrn hornklofi has thus created a new genre of Preislied that supplanted the older, inherited one; see also Vogt (1930a) and Beck (1986) on the Norse Preislied.
The metre is málaháttr (though Kuhn 1983, 336 disputes this terminology), but with an admixture of ljóðaháttr in the last six stanzas (18/3-5, 19/1-3, 20/3-5, 21/1-6, 22/3-5, 23/5-10; see Haugen 1994, 80).
Most of the poem (sts 1-11, 15-23) is preserved in Fsk, and all principal transcripts have been consulted for this edition: the three FskA transcripts (FskAˣ, 52ˣ, 301ˣ), and the three FskB transcripts (FskBˣ, 51ˣ, 302ˣ). Of the FskB transcripts, FskBˣ is generally considered the best, but for Harkv, 51ˣ seems slightly better, and it is employed below as the main ms. wherever no better copy survives. Stanzas 6-11, 14 are recorded also in Hkr, to which Kˣ, F and J1ˣ, J2ˣ are the best witnesses. Hkr here offers a better text than Fsk. Stanzas 11-12 are cited in SnE (st. 11 in R, Tˣ (ll. 1-4), W, U and st. 12 in R, W), and st. 12 is found only there; st. 12 in LaufE (1979, 331) is copied from W and therefore not used in this edition. Stanzas 7-11, 13-14 and 21 are preserved in Flat (ms. Flat), in the section with the editorial title Haralds þáttr hárfagra (HarHárf), and st. 13 uniquely so. In general, where the text in Flat can be compared with other witnesses, it shows itself unreliable. Finally, one stanza (st. 14) is preserved in ÓT, in representatives of all four classes of mss of this saga: 61 (A), 53 (B), Bb, 325IX 1 b (C) and Flat (D). Finnur Jónsson in Skj A makes use of AM 55 fol, a copy of Flat, rather than Flat itself, for the ÓT text. He may have been unaware that the stanza appears twice in Flat. Occasional reference is made in the Notes to the stanzas copied into 761aˣ, some with variants in the margin. The stanzas there are all copied from known mss, but in some stanzas readings from different traditions are intermixed, and since it is more a reconstruction than a transcript, this ms. is not accorded equal status with the other witnesses.
The poem has been edited complete by Möbius (1860) and Jón Helgason (1946 and 1968, 10-21), and selectively in a number of anthologies, which are occasionally cited in the Notes below. The centuries-long interest in this poem has produced a sizeable body of scholarship regarding it, and the Notes can only very selectively represent that scholarship.