Diana Whaley 2012, ‘Reconstruction of skaldic poems’ in Diana Whaley (ed.), Poetry from the Kings’ Sagas 1: From Mythical Times to c. 1035. Skaldic Poetry of the Scandinavian Middle Ages 1. Turnhout: Brepols, pp. xxxix-xliv.
Some skaldic poetry, above all much of that in SkP VII, Poetry on Christian Subjects, is preserved in complete and continuous form in manuscripts sometimes not very distant in date from the time of composition. For the majority, however, preservation is fragmentary and scattered, and the attempt to reconstruct, as far as possible, the original poetic structures is a major aspect of the editorial brief, which involves determining authorship, distinguishing lausavísur (freestanding, occasional stanzas, literally ‘loose verses’) from stanzas extracted from extended poems, establishing what poems might have existed, assigning stanzas to individual poems, and ordering stanzas within poems. The task of reconstruction is hampered by the fact that, in the absence of completely preserved poems pre-dating Einarr Skúlason’s Geisli ‘Light beam’ in the mid-twelfth century (except perhaps for Egill Skallagrímsson’s Hǫfuðlausn ‘Head-ransom’), we cannot be certain what the normal range of possible structures is. Nevertheless, much does remain and in so highly self-conscious an art-form the skalds have left many clues about the structures of their poems, which are outlined in §4.5 below.
The uncertainty inherent in many aspects of reconstruction is reflected in the often wide differences between previous attempts to arrange all or much of the skaldic corpus, of which the principal are:
(ii) Gudbrand Vigfusson (Guðbrandur Vigfússon) and F. York Powell in CPB (1883; selection).
(iii) Finnur Jónsson in Skj (1912-15); E. A. Kock in Skald (1946-9) retains Finnur’s ordering.
(iv) Bjarne Fidjestøl, Det norrøne fyrstediktet (1982, 169-77 et passim), covering only the royal panegyrics and reconstructing them by reference to stanza numbers in Skj.
(v) Discussions of individual poems in editions, articles and monographs, or published or unpublished theses.
The present edition is indebted to all these sources, and agrees substantially with the attributions and reconstructions that have become familiar especially through Finnur Jónsson’s monumental Skj. However, the radical rethinking that has been involved has also produced significant disagreement (in some cases because traditional assumptions are found not to have any firm basis), resulting in reconfiguration, renaming and renumbering, and all possible measures have been taken to assist readers wishing to work across the old and new systems by cross-referencing all stanzas to their locations in Skj A and B and other major editions. The principles on which reconstruction has been undertaken in this edition are discussed below, but the fundamental one has been to give priority to the explicit witness of medieval manuscripts, unless there is strong reason to doubt it, since while there may be reservations about that evidence, still greater uncertainties may be involved in departing from it. Specific issues of reconstruction, especially where problematic, are discussed fully in the Introductions to individual poems.
Questions of authorship involve a two-pronged approach: assigning extant stanzas to skalds, and attempting to establish what the oeuvre of individual skalds was, i.e. what poetry existed. Occasionally skalds identify themselves within their stanzas, but the principal evidence for authorship is external, and typically takes the form either of introductions to fragmentary citations in prose works (e.g. explicitly Svá segir Þjóðólfr ‘As Þjóðólfr says’ or implicitly Ok enn ‘And further’) or headings to complete poetic texts (e.g. that of Geisli in Flateyjarbók; cf. SkP VII, 6). This evidence is usually (though not always) consistent across different sources and is generally respected in this edition, except in extreme cases such as attributions to troll-women, which are classified as anonymous. Light on the oeuvre of skalds is very occasionally thrown by cross-references to their predecessors’ work: Hallar-Steinn’s Rekstefja (HSt Rst 34I) places itself in a tradition of encomia for Óláfr Tryggvason by referring to those of Hallfreðr and one ‘Bjarni’. But again the main source for this kind of information is external, contained in Skáldatal ‘Enumeration of Skalds’ found in the Uppsala (U) text of Snorra Edda and in AM 761 a 4°ˣ, which list Scandinavian rulers together with the poets who composed for them (see SnE 1848-87, III, 205-86 for both versions and a conflated text). The information can be useful in reconstructing the oeuvre of a skald, though the two versions do not wholly agree and only cover court poetry, and some of the detail has met with scholarly scepticism. Akin to this, though far less systematic, are the occasional references to poetic compositions in prose works, as when Hallfreðr is credited in his saga (Hallfr, ÍF 8, 178) with an Uppreistardrápa ‘Drápa concerning Creation/Regeneration/Reparation’ or when Þjóðólfr Arnórsson is accused in a short tale of composing Sóptrogsvísur or Sorptrogsvísur ‘Dustbin vísur’ (ÍF 23, 277; the ms. has ‘sop-’). Discussions of authorship may also include observations on content, style or metre, but given the conservatism of much skaldic poetry, the difficulty of dating it, and the incompleteness of the surviving corpus, such internal evidence can only be used with caution and in an ancillary role.
A fundamental distinction is to be made between skaldic stanzas belonging to extended poems (whether now complete or fragmentary) and lausavísur (freestanding, occasional stanzas), and scholars attempting to distinguish between the two have assigned varying weight to external and internal evidence. These two types of evidence are compatible in most cases, but where there seems to be disparity, the present edition is predisposed towards the external evidence (cf. also Fidjestøl 1982). The following discussion moves from general principles to caveats, and focuses especially on court poetry as the most prolific and most problematic genre of extended poem, though long poems exist elsewhere, for instance the Christian poetry in SkP VII or the fragmentary Grámagaflím of Bjǫrn Hítdœlakappi in Bjarnar saga (Bjhit GrámV) or poems of Egill Skallagrímsson (Egill ArkvV, Egill HflV, Egill StV) in some manuscripts of Egils saga.
Stanzas from extended poems may be explicitly identified as such in their prose contexts, e.g. Svá segir í Velleklu ‘As it says in Vellekla’ or Þess getr Þórðr Kolbeinsson í Eiríksdrápu ‘Þórðr Kolbeinsson mentions this in Eiríksdrápa’. Where a further stanza follows on immediately, or else is introduced with a phrase such as Ok enn kvað hann ‘And he spoke further’ or Ok enn ‘And further’, this most likely indicates that it belongs to the same poet and poem as the preceding one, though this cannot be assumed automatically. Most stanzas are not explicitly cited from named poems but only affiliated to their poets using formulas such as Sem Sigvatr segir or Svá segir Sigvatr, both ‘As Sigvatr says’, or Þess getr Sigvatr ‘Sigvatr mentions this’. Nonetheless, the collocation of such formulas (which we can collectively term svá tags) with the titles of poems (e.g. Vellekla and Eiríksdrápa above), and the fact that the stanzas they introduce tend to share certain characteristics of content and style, suggests that (despite the caveats below) these formulas are reasonably reliable indicators of stanzas from extended poems.
Turning now to the internal evidence of content and style, extended court poems are typically characterised by retrospective narrative or description, often quite detailed and leisurely, of events such as battles and voyages, and they are often punctuated by grandiose general praise, apostrophes to living rulers, prayers for the souls of the deceased and other ‘formal’ features (cf. §4.5). The skald may refer to himself as poet (his composition or performance of the poem), or as actor (his involvement in events or his relationship with the ruler), though the extent to which skald-as-actor figures in extended poems is contested, as discussed below. The absence of other features normally associated with lausavísur may also be an indicator that a stanza originally belonged to an extended poem.
In addition to the internal characteristics of individual stanzas and the manner of their citation in prose works, many sets of stanzas attributed to the same skald share narrative or descriptive content, or have a concatenation of imagery or metrical features which suggest that they belonged together in the same poem, in some cases clearly forming sequences.
Stanzas suspected to be lausavísur often have a sense of immediacy effected through features such as present-tense or periphrastic future verbs, deictic adverbs referring to the ‘here and now’ not merely of the performance situation but of the events narrated, or certain lexical items such as verbs or nouns of hoping or expecting. The skald is likely to figure in the action, resulting in first-person references, and there may be second-person addresses to (often non-royal) persons. The style may also be more straightforward than in formal encomiastic poems, with simpler clause arrangement and less elaborate kennings, and the sense of the stanza will be complete in itself. Such characteristics again correlate strongly with a particular manner of citation. In this case there is ‘staging’, or narration of a scene in which the skald is present as one of the dramatis personæ and utters a stanza, usually prefaced by a þá tag such as Þá kvað Þormóðr ‘Then Þormóðr spoke’, as distinct from the svá tag associated with extracts from extended poems. Occasionally prose writers stipulate that a skald who was present spoke a vísa, which is the nearest thing to an overt statement of lausavísa status available to us.
The general tendency for internal and external evidence to agree inspires a certain confidence in the techniques of reconstruction, and there is a high degree of consensus among the principal published attempts to reconstruct the skaldic corpus (listed above). However, there is also a significant amount of uncertainty. In some cases the external and internal evidence do not agree, and this reminds us that assumptions about the meaning of the external evidence of prose context are indeed only assumptions. It is reasonable to see the þá tags as probably signalling lausavísur and the svá tags as signalling extracts from extended poems, but except where poem titles are given or a single vísa stipulated it is not certain that they are intended as source references or literary-historical statements; the only certainty is that they differentiate kinds of narrative contexts. Hence authors using these tags may not necessarily mean ‘this was composed as a lausavísa’ but ‘I am treating this as a lausavísa, improvised by the skald in the midst of the action’. Indeed, scholarship has pointed to many narrative contexts surrounding lausavísur, especially in the sagas of Icelanders, which seem to have been created out of the stanza rather than representing ancient traditions, and some of the stanzas in question are believed to have come from extended poems. Hence extended poems may contain features traditionally associated with lausavísur such as present-tense narration and description and first-person references (Poole 1991, passim). Moreover, the range of choice is not merely between lausavísur on the one hand and stanzas from long formal poems on the other. There is a spectrum from drápa (extended poem with refrains) – flokkr (extended poem without refrains) – vísur (set of stanzas) – to lausavísa (freestanding occasional stanza). Added to this, some hybrid forms may have been possible: Fidjestøl (1982, 84-5) suggested that some stanzas that appear to have the characteristics of lausavísur may have been loosely attached to praise poems, for instance as a framework. In cases of doubt, for example where stanzas ‘staged’ as lausavísur are suspected to have originated in extended poems, the evidence of the medieval sources is followed in this edition but alternatives are discussed in the Introduction. By the same token, the absence of medieval evidence is taken seriously, and a particular area of difficulty is the numerous extracts apparently from court poems which are cited from named skalds but not named poems in Snorra Edda. Unless the internal evidence for ascribing these to known court poems is strong, they have been treated as fragments.
The deduction that a stanza not only has the general characteristics of an extended poem but belongs to a particular poem is then made on the basis of external indicators, especially an explicit title or embedding in a narrative about a particular ruler, or internal indicators such as the naming of an individual or a strong focus on him through narrative, praise or prayers, or through naming of persons, places, weapons or ships known to have been part of his biography. In terms of those biographies there is little choice but to rely on the Old Norse prose works themselves, but in some cases there is external support; for instance, the attack of Magnús góði on the Wends at Jómsborg (Wolin) in the 1040s is not only commemorated by more than one skald and several Old Norse prose writers but also recorded three decades later by Adam of Bremen in his Gesta Hammaburgensis ecclesiae pontificum (ed. Schmeidler 1917, 137).
Metre can rarely be invoked, but it does distinguish between poems in the rare cases where a skald composed on the same subject in more than one metre. More subtle metrical features, such as a predominance of certain metrical types, could only be an adjunct, cautiously used, to arguments for reconstruction, and the same applies to aspects of diction, imagery and style.
A recurrent problem in reconstructing court poetry is that of divining which poems are erfidrápur ‘memorial drápur’ (see, e.g., Fidjestøl 1982, 193-8). Verbs in present or past tenses, and the presence or absence of apostrophes, can help distinguish non-erfidrápur from erfidrápur, but the two are otherwise difficult to separate, and there may be alternation of tenses within same poem; still more widespread is alternation of second and third person pronouns and verbs.
Editorial ordering of stanzas within extended poems is informed partly by the known skaldic terminology and structures, described in §4.5 below. Beyond that, it appears reasonable to assume that the events were treated in a logical sequence: that a hero’s career, or the phases within a particular campaign or battle, will be covered in chronological order, or that a journey will make geographical sense. Some proof of this is afforded by poems such as Sigvatr Þórðarson’s Víkingarvísur (Sigv VíkvI) or Halldórr skvaldri’s Útfarardrápa (Hskv ÚtdrII) that offer an explicit numbered sequence. Similarly, the lausavísur and fragments attributed to given skalds are as far as possible ordered chronologically in this edition. Stanzas containing no chronological indicators and preserved independently of historical narratives (especially in Snorra Edda) present particular problems of ordering. Unless there is evidence for a particular ordering (which might include comparison with better-preserved poems), this edition follows the tradition of placing such stanzas after those whose sequence can be determined chronologically.
As with other aspects of reconstruction, the contexts in which stanzas are cited in narrative prose works provide valuable clues as to their ordering in their poems of origin, and some narratives bear all the signs of having been constructed from verse sequences. However, there is always a possibility that prose authors chose to depart from the poem’s order or knew the poem in an incomplete or differently ordered form, and the fact that different prose texts quite frequently disagree about the make-up of stanzas (i.e. which helmingar or lines form a single stanza and in what order), the ordering of stanzas, and what events they describe, suggests a need for caution in using this kind of evidence. This is especially true since such evidence can only rarely be confirmed by sources independent of the poetry itself and of prose compilers’ interpretations of it. Where serious discrepancies and difficulties exist these are noted in the Contexts and/or Notes to the stanzas in question and in the Introductions to poems.