FAULKES (1997)- Poetical Inspiration

POETICAL INSPIRATIONIN OLD NORSE AND

OLD ENGLISH POETRY

By

ANTHONY FAULKES

PROFESSOR OF OLD ICELANDIC

IN THE UNIVERSITY OF BIRMINGHAM

The Dorothea Coke Memorial Lecturein Northern Studies

delivered at University College London28 November 1997

PUBLISHED FOR THE COLLEGE BY THEVIKING SOCIETY FOR NORTHERN RESEARCH

LONDON

UNIVERSITY COLLEGE LONDON 1997

PRINTED IN THE UNIVERSITY OF BIRMINGHAM

ISBN: 0 903521 32 6

POETICAL INSPIRATION IN OLD NORSEAND OLD ENGLISH POETRY

IN THE VENERABLE BEDES ACCOUNT of the earliest known Anglo-Saxon poet Caedmon, we are told that Caedmon was an un-lettered herdsman who was unable to take part in the eveningentertainment of his fellows because he could not sing. One nightafter he had left a party early because of his embarrassment anangel came to him in a dream and told him to sing of creation,and the result is Caedmons hymn, a short poem about Godscreation of the world for men, the most remarkable thing aboutwhich is its complicated verbal and metrical structure. There-after Caedmon was found to have a gift for poetry; and it isinteresting to note that his later poems were evidently producedby the learned monks of the monastery retelling to him storiesfrom the Bible which he is said to have chewed over like a cowchewing the cud and to have reproduced in poetical languagewith great sweetness (Hist. eccl. IV 24).

From Bedes account it is clear that the divine gift Caedmonreceived was one of expression, the ability to tell existing storiesin verse form, what oft was thought but neer so well expressd.If he was inspired, it was not the content, but the language thathe was inspired with, as is argued by C. L. Wrenn in The Poetryof Caedmon (1946, 286), who suggests that the real miracle ofCaedmon was that an unlettered peasant should become able touse the style and manner of learned court poetry. There is a closeanalogue to the story of Caedmon in the Old Icelandic story ofHallbjo rn hali in Flateyjarbk (F IX 2279). His problem wasthat he wanted to compose an elegy for an earlier poet but couldnot find the words. He fell asleep on the poets grave mound andthe dead man came out to him in his dream and gave him theability to compose poetry, instructing him to compose in ascomplex a metrical and verbal style as possible. One interesting

3

POETICAL INSPIRATION

4

similarity between the stories of Caedmon and Hallbjo rn is thatwhile both are said to have gone on to compose much other poetry,no further verse by either of them has been recorded.

Stories about individuals receiving inspiration in composingpoetry from supernatural sources are widespread, but what I wantto compare with Bedes account of Caedmons inspiration is whatSnorri Sturluson wrote about poetry in Iceland in the thirteenthcentury. Snorri is one of the few writers of medieval Scandinaviawho express clear critical attitudes about the nature of vernacularpoetry, and although he was writing towards the end of the periodwhen oral skaldic poetry was cultivated in the North, it is likelythat his views about poetry would have been shared by manypoets and their audiences in his time and earlier. He has a mythabout the origin of poetry in his Edda, according to which poeticalinspiration is given to gods and men by inn in the form of analcoholic drink brewed from the blood of a person of immensewisdom created from the spittle produced by two races of godsto seal their truce after their war together. This drink had to berecovered by inn after it had been in the possession first ofdwarfs and then of giants, and when he escaped with it back tothe other gods he carried it in his stomach while flying in theform of an eagle; he was chased by another eagle and could notprevent some of the liquid from being expelled backwards; thatpart of it was left for rhymesters or poetasters. The rest which hewas able to bring up into containers in sgarr was for the gods,and for human poets and scholars.

There are many analogues to this story too in the mythologyof various countries, some of them well outside the area ofGermanic culture. Skaldic poets frequently alluded to the storyin their poetry. It seems at first sight that this would support theidea of poetry having being seen in medieval Scandinavia as aninspirational activity of which the symptoms were similar todrunkenness. Thus the seventeenth-century Icelandic poet Magnslafsson in his essay on Norse poetry published in 1636 speaksof a poet reciting appearing to be vino madens and uses the termskldvingl or poetica vertigo:

Further, our poetry also has this peculiarity: whereas in ordinarylanguages anyone can put together poems in accordance with the

IN OLD NORSE AND OLD ENGLISH POETRY

5

fashion of his country, force the words into some sort of rhythmicalpattern, and by practice achieve some fluency; in ours no one becomesa poet, or can put together even the simplest kind of poem withoutgreat labour, however much he wants to, unless he is especiallygifted with poetic inspiration. And this inspiration, like other naturalpassions, affects some more violently and others more gently. Someproduce good poetry after long working out, others pour out allkinds of poems extempore with a more violent kind of impulse, sothat whatever they try to express turns out to be poetry, as thatmost ingenious Roman poet once claimed of his own poetic fluency,and they find verse just as easy as prose. Further, this sort of qualityreveals itself by clear signs straightway at a very early age. Normust it be forgotten that this activity of the spirit is hottest at thetime of the new moon, and you would say that an outstanding poetexplaining poetical matters to others, or occupied in delivering poetry,was under the influence of drink, was afflicted with a rather severeattack of melancholy, or was seized by some madness; and oftenthis quality can be detected even in strangers from a certain particularmannerism which we call skldvingl or poetical delirium.

Magns lafsson, however, was influenced by Renaissanceattitudes to poetry (and the Roman poet he refers to is Ovid,Tristia IV x 25), though he may not actually have readShakespeares words about the poets eye in a fine frenzy rollingor the lunatic, the lover and the poet, and I know of no medievalIcelandic stories of individual poets being inspired either literallyor metaphorically to compose particular poems by draughts ofthe mead of poetry, and although inn is sometimes referredto by poets (most notably by Egill Skalla-Grmsson) as the giverof the gift of poetic expression, he is never invoked by poets inScandinavia in the way that the Muses were invoked by classicalpoets. What inns mead gives is not the inspiration to composea poem, but the ability and skill to express oneself in verse (andthe ability to be a scholar). Poets speak of composition andperformance as pouring out inns mead, not as drinking it;Skldskaparml verses 4 and 17 by Skld-Refr are ambiguous,but verse 18 by Einarr sklaglamm is clear. Refr said this:

(4) Often the kind man brought me to the raven-gods [inns]holy drink [instructed me in poetry].(17) To you we owe Falrs cups [the mead of poetry], nobleSlaughter-Gautr [inn].


6

Einarr sklaglamm said this:(18) I shall succeed in bailing the draught of Host-Tyrs [inns]wine-vessel [the mead of poetry] before the ship-impeller [seaman,i. e. Earl Hkon]I need no urging to that.

When skaldic poets refer to the receiving of inns mead, it isgenerally their hearers who do this, not the poets (Skldskaparmlverses 27, 28, 29). The role of inn is made more complex byreferences to the god Bragi as also being a patron of poetry, thoughthere are no stories to illustrate in what way he relates to humanpoets. It may be added that there is not a close connection betweenpoetry and religious ritual expressed in Old Norse poetry, exceptperhaps in some eddic verse.

Muchthough not allskaldic poetry is characterised bycomplexity of diction and metre. There are several anecdotesillustrating the difficulties of understanding it (as well as someindicating the difficulty of composing it; Snorri refers to this inHttatal in the introduction to stanza 17), and some that indicateattitudes to its complexity. These anecdotes are mostly from thethirteenth century, and so do not necessarily tell us about attitudesto poetry in the Viking Age, but they are probably contemporarywith Snorri and reinforce the impression gained from his writings.The puzzle element in Icelandic poetry is dominant; the sameword, ra, is used of interpreting poetry as of interpreting runes,riddles and dreams. Snorri in several places in Skldskaparmlrefers to poetry in terms of concealment. He makes gir describea kenning for poetry as vel flgit rnum, concealed well inrunes; Bragi had just referred to the use of this kenning as vrfelum rnum ea skldskap sv, we conceal [it] in runes orin poetry thus (the basic meaning of the word rn is secret).gir refers to other kennings as myrkt, obscure (Edda SnorraSturlusonar 1931, 813). Snorri refers to the purpose ofSkldskaparml as to help young poets at kunna skilja at erhulit er kveit, to be able to understand what is composed so asto be concealed (Edda Snorra Sturlusonar 1931, 86). He alsospeaks ironically of the use in poetry of gera ofljst at vant er atskilja, making too clear so that it is difficult to understand. Hisaccount of ofljst gives a good idea of how important he thoughtthe puzzle element in poetry (Edda Snorra Sturlusonar 1931, 193):


7

Lti means two things. Noise is called lti, disposition (i) iscalled lti, and i also means fury. Reii also has two meanings.It is called reii [wrath] when a man is in a bad temper, the gear(fargervi) of a ship or horse is called reii. Far also has two meanings.Fr is anger, far is a ship. People frequently use such vocabularyso as to compose with concealed meaning, and this is usually calledword-play (ofljst [obvious]). People call it li [joint] on a personwhere the bones meet, li is a word for ship, li [troop] is a wordfor people. It is also called li [help] when someone gives anotherassistance (lisinni). L is a word for ale. There is what is calleda hli [gateway] in an enclosure, and hli is what people call anox, and hl is a slope. These distinctions can be made use of inpoetry so as to create word-play which is difficult to understand,if it is a different distinction of meaning that has to be taken thanthe previous lines seemed before to indicate. Similarly there arealso many other such words where the same term applies to severalthings.

Middle English poetry is claimed to have a large element of thedesire for concealment in its purpose and style by A. C. Spearing(Readings in Medieval Poetry 1989, 97). In Gsla saga (ch. 18),the hero (who is a poet) is depicted as reciting a poem which heshould have kept to himself. It is overheard by his sister, whogot the verse by heart from the one hearing, and goes home,and by then she has worked out its meaning. In the verse Gslihas revealed his guilt for the secret killing of her husband (F VI589). There is a comparable riddling confession of guilt in Grettissaga verse 11 (ch. 16; F VII 47). In Sneglu-Halla ttr (F IX26395), which is even more obviously fictional than Gsla saga,but which nevertheless must express attitudes and values of thethirteenth century, the hero is depicted as not only a poet but ajoker; he comes to the court of Harold Godwineson and recitesa poem supposedly in his honour. The king and his court areunable to understand the poem immediately, and when the poetasks for his reward the king orders silver to be poured over hishead so that he can keep as a reward an amount comparable towhat the audience retained of his poem. Halli smears tar overhis hair so that a lot of the silver sticks, and then makes good hisescape, because he knows that eventually the king will realisethat the poem was nonsense. This is clearly a satire on the


8

incomprehensibility of Viking poetry; the point does not seemto be the linguistic barrier between Icelanders and Englishmenin the eleventh century, since saga-writers do not seem toacknowledge that there was one. In Gunnlaugs saga it is claimedthat then the same language was spoken in England as in Norwayand Denmark (F III 70).

Two verses supposed to have been spoken by Haraldr harribefore the Battle of Stamford Bridge in 1066 have been taken todemonstrate that kings preference for elaborate and obscure verse.The story is told in Snorri Sturlusons Heimskringla (F XXVIII,1878). He is said to have left his armour in his ship and is goingto have to fight without it. He states this in a simple, straightforwardand quite expressive verse in the simple metre of eddic poetry.Then he says, etta er illa kveit ok mun vera at gera ara vsubetri, that is badly composed and I shall now make another betterverse, and utters a rather complex verse in very involved skaldicstyle which is really rather difficult to interpret, though there areplenty that are worse.

I now want to look at some examples of skaldic verse and thetechniques used in them, and what Snorri says about them, tosee what impression we can get about what was actually valuedin this poetry when it was composed. Snorri discusses two aspectsof poetry at length in his Edda, vocabulary and metre; he hasvery little to say about other aspects of it such as content. In hisdiscussion of vocabulary he spends most time on nominal groupswhere instead of referring to a person or thing by its normalname, a poet replaces (or conceals) that name with another, whichmay be a poetical term like steed for horse, or a kenning,which has at least two elements, such as when gold is calledfire of the sea or a king distributor of gold. Many of theseterms are not metaphorical and few can be described as images;generally they do not focus on the characteristics of individualpersons or things, but convey the general concept of gold objector ruler and so on without expressing any feature of them eitheras individuals or as being in a particular situation. That is, a rulerwill be a distributor of gold even when he is fighting a battleand gold will be called the fire of the sea even when it is in theform of a mans arm-ring on his arm. If the man wearing the


9

gold ring is fighting a battle on land the mention of the sea willhave no relevance to his situation at all and does not contributeto the picture of the action being described.

Among the practical difficulties of understanding and translatingsuch verse is the problem of working out the syntactical relationsof words and phrases and knowing how involved and complexthey are supposed to be. In an inflected language like Icelandic,understanding the functions of words depends less on their positionin the sentence than on their grammatical form, so that it is oftenpossible in verse to depart from the normal word-order withoutmaking the functions of words ambiguous, and skaldic poets usethis facility extensively, so that not only the order of elements inthe sentence is varied, but words belonging together as parts ofa single phrase are often separated or interwoven with parts ofother phrases. In kennings, one element is generally in the genitivecase, and in most words, though indeed not in all, the genitiveis easily recognised and distinct from other forms of the word.But when there are two kennings juxtaposed, each of whichcontains a genitive, it can sometimes be difficult to see whichgenitive belongs with which base-word.

I am afraid I am now going to get rather technical, and I amsorry if this induces any feeling of poetical vertigo in you, butit is necessary to go into technicalities to understand the poetry,and this is believed to be oral poetry which was understood orallyby its original audience. There is a difficulty in knowing how topresent skaldic verse for a modern audience, though I am con-vinced that it is possible to make it meaningful even to thosewith little knowledge of Old Norse. It is common in moderneditions of skaldic verses to present the text in prose word-order,i. e. to print the verse first in its manuscript form and then to re-arrange the words so that those that belong to the same phrasecome together. Snorri himself uses this method of explication(i. e. re-ordering the words as prose) with two verses in Httatal(which is a poem he himself composed), verses 1718. This ismainly because in these verses he is illustrating refhvo rf, wherewords of opposite meanings are juxtaposed though they do notalways have those opposite meanings in the verse as a whole.His verse 17 and explanation (which is not entirely a re-ordering


10

of the words, but also in part a prose paraphrase) are as follows.The prose word-ordering is printed in bold type.

Sks glar verr skirsltt skar hafi jarar;hlfgranda rekr hendirheit ko ld loga o ldu;fljtt vlkat skilr fylkirfrilro uls svarrnsi rsir sto varreirglar fro mum meium.

. . . Hr eru snd essi vsu sextn orto k sundrgreinilig, ok eruflest ofljs til rtts mals at fra, ok skal sv upp taka: sks gl,at er gull; skir gulls, at er mar; hann verr skar jarar hafisltt, at eru Firir, sv heitir fylki i Nregi; hlfgrandi, at ervpn; hendir loga o ldu er mar, er rekr ko ld heit sverinu, at erat hegna siu; fljtt vlkat m at kalla er skjtt rit er, at skilrhann af friinum; konungr heitir fylkir, rnsi rsir stovar svarro uls fro mum meium.. . . Here are demonstrated in this stanza sixteen phrases of contrarymeanings, and most of them have to be turned to their proper meaningby means of word-play, and this is how it is to be understood: ditch-glede, i. e. gold; attacker of gold, that is a man; he defends thecleft of land smoothed by the sea, i. e. Firir, this is the name ofa district in Norway; with shield-harmer, i. e. a weapon; throwerof wave-flame is a man, who drives away cold threats with thesword, that is to punish wickedness; hastily-weighed may be saidof what is unpremeditated, he perceives this from the hostility; aking is called fylkir [leader]; the ruler puts a stop to the plunderinghabits of bold sea-sun-trees.

Verse 18 is similar. This sort of effect, where words have to beunderstood in two meanings at once, must have been very difficultto convey in the pre-literary period. It is doubtful, however, whetherthis method of re-ordering the words of a verse into prose word-order could have been used in the pre-literary period (it is onlywith these two verses of Httatal, and in a less exhaustive way,verses 46 of the same poem, that Snorri himself uses it), thougha somewhat similar second explanatory text was apparentlypreserved orally alongside the real text in the case of Vedicpoems, which are probably much deeper in the oral period of


11

their culture than any Scandinavian verse is in its (Macdonell1917, xiixiii). But it is difficult to see how people could haveunderstood complex skaldic verse without some such assistance,and in particular how the audience could have been made to realisewhich words belonged together in a verse when there is a genuinepossibility of taking them in two ways.

In my explanations I prefer not to re-order the words, since Ithink that the word-order not only embodies the structure of theverse, but also to a large extent the meaning, and I think it unlikelythat this procedure would have been used in pre-literary times inScandinavia. So I use the resources of the computer to identifywhich words belong together and how they can be representedin English; this method is in fact rather like that of E. A. Kockin his edition in Den Norsk-Islndska Skaldediktningen (19469).The asterisks mark words that have been unavoidably emended.Translation on its own cannot hope to reproduce much of themeaning or effect of the original. The numbers are the numbersof the verses in Finnur Jnssons edition of Edda SnorraSturlusonar of 1931.

An example of how words can be taken to belong to differentphrases is found in this verse of Einarr Sklason describing anornamented axe:

(193) Bleisu liggr bibjargs tveim megin geimasjs ek skkva strisnr ok eldrat mra.

There lies on each side of the blood-embers head (the head of thegold-adorned axe) both the purses snow and oceans fire; I mustpraise the punisher of Vikings.

There are two kennings for the decoration on the axe: sjs snrand geima eldror it might be sjs eldr and geima snr. Arethey differentiated as gold and silver, or are they, contrary to ourintuition, just general kennings for precious metal, as is impliedby Snorris commentary to verse 194 and the short anonymoustreatise on kennings that Finnur Jnsson called Den lille Sklda(Edda Snorra Sturlusonar 1931, 256/22: gull m kenna til snsok ss, gold can be referred to as snow and ice)? The kenning


12

skkva stri is also ambiguous: skk can mean gold, and skkvastri could mean destroyer, i. e. dispenser of gold, generous prince.

My next example is anonymous:

(317) *Orgildi var ek (Eldis) [MS O rgildis]ls Fjo rgynjar (mla)dyggr; s heir ok hreggi(hrynbes) r steja.

I was loyal to the liberal payer of Fjo rgyns eels [serpents] tinklingbed, and honour be to the storm of the river-anvil-[rock-]Eldirs[giants] speech.

The two complicated kennings in this verse (referring to the samegenerous man) can be analysed in two quite different ways withoutthe meaning being much affected, since the basewords (o rgildi,hreggi; both dative, with ek var dyggr and s ok heir) are in thecontext synonymous (liberal payer; storm, destroyer) and boththe strings of determinants mean gold: o rgildi hrynbesFjo rgynjar ls and hreggi r steja Eldis mla, or o rgildi rsteja Eldis mla and hreggi hrynbes Fjo rgynjar ls.

The next example is from Einarr sklaglamms Vellekla:

(223) Ne sigbjarka serkirsmmijungum rmuHrs vi Ho gna skrirhlut fast of sir.

The three nouns serkir, (sm)mijungum, skrir are base-wordsin kennings for, respectively, mail-coats, warriors, missiles: Nehlut fast of sir serkir mijungum vi skrir, the firmly-sewnmail-coats did not defend warriors from missiles. The wordswhich may be determinants in these kennings are sigbjarka, sm-,rmu, Hrs, Ho gna. There seem to be too many of them, and itis difficult to see which base-word each belongs with. PresumablyHogna goes with skrir, sigbjarka (battle birches, i. e. warriors:it is unusual to refer to men by feminine tree-names) with serkir,and smmijungum, bow-giants is another kenning for warriors.What do we do with rmu Hrs? Take it as a dative phrase, ininns tumult, i. e. in battle? But rma means battle on itsown. Attaching Hrs or rmu to any of the other kennings resultsin redundant elements to the kennings, though it is not certain


13

that such redundancy was avoided by skaldic poets; there is quitea lot of it in their work.

In many verses, two parallel statements are juxtaposed orinterwoven; obviously if each has a nominative subject and/oraccusative object, only the meaning of the verb can determinewhich subject and object go with which verb, as in this verse inan account of a battle by ttarr svarti:

(340) O rn drekkr undarn, Eagle drinks breakfastylgr fr at hrm sylg, she-wolf gets from corpses drinkopt rr lfr kpt, often reddens wolf its jawsari getr ver ar. eagle gets food there.

The question is, which of the two possible interpretations shouldwe accept and prefer, and why, taking undarn as object of drekkr(drinks breakfast) and sylg as object of fr (gets drink) orvice versa (drinks liquid, gets breakfast)? (There are not thesame possibilities in lines 3 and 4.) The first, which is syntacticallysimpler, is preferred by E. A. Kock, the second, which issemantically more logical, by Finnur Jnsson. However they aretaken, both phrases refer to the devouring of dead warriors bybirds of prey, but sylg would fit better than undarn as the objectof drekkr, giving interchange of the elements of the two clauses.Kocks interpretation makes the quatrain an example of Snorristtmlt, where each line is a separate integral statement. Tryingto understand which interpretation the poet meant raises basicquestions of the meaning of utterances of the kinds that thephilosopher Wittgenstein puzzled over. What does it mean tosay that when I say something I mean this rather than that whenwhat I have said could mean either? His arguments would supportthe idea that skaldic verse too was a game, a social ritual, wherethe function of the activity was more important than the meaningor content of the utterances.

There is a similar kind of syntactical ambiguity in this verseabout rr fighting a giant from jlfr of Hvinirs Haustlo ng :

(67) yrmit Baldrs of barmi Baldrs brother spared notbergslgnum ar dlgi (rocks) the greedy enemy therehristusk, bjo rg ok brustu, (shook, crags also shattered;brann upphiminnmanna. the sky above burned) of men.


14

The interwoven statement here can be taken as berg hristusk,bjo rg ok brustu, or hristusk bjo rg ok berg brustu. Both the verbsand both the nouns are near synonyms, and various orders arepossible, though Kock prefers to take berg with slgnum so thatthe parenthesis has only one subject. Manna could conceivablygo with upphiminn rather than with dlgi. With these twosuggestions adopted, lines 3 and 4 become integral statementsand there is no parenthesis at all. It is doubtful whether that iswhat a medieval audience would have preferred, and sky of menis a less normal expression than enemy of men (= giant). Arenormal expressions what we should expect in skaldic verse? Islogic a valid criterion of meaning and authorial intention? Whatare the underlying aesthetic preferences, for naturalness and simp-licity or for complexity and puzzlement? There are also multiplepossibilities of arranging the kennings in lines 58 of this stanza.

There is another example in the description of the death of thegiant jazi in Haustlo ng:

(104) Hfu skjtt en skfu Began quickly and shavedsko pt ginnregin brinna shafts mighty powers to burn

Here there are two plural verbs, skfu and hfu brinna, and twoplural nouns, sko pt and ginnregin, which could each be eithernominative or accusative. The two statements can be read shaftsquickly began to burn, and the mighty powers shaved them, orthey quickly began to burn and the mighty powers shaved shafts,or sko pt could be read both as the subject of the one verb and asthe object of the other, shafts quickly began to burn, and themighty powers shaved shafts. Skjtt could also, of course betaken with either clause or with both (the phenomenon of oneword belonging to two separate clauses is discussed below, p. 18).

If there is more than one adverbial phrase in a verse containingmore than one clause, it can be difficult to see which phrase goesin which clause. For instance, Bragi the Old describes rrsfishing for the Midgard serpent like this:

(42) Var l Viris arfavilgi slakr er rakisk, Eynfis o ndri,Jo rmungandr at sandi.


15

In this verse there are two clauses, the main clause Var Virisarfa l vilgi slakr, Virirs sons (i. e. rrs) fishing-line layby no means slack and the subordinate clause er rakisk Jo rmun-gandr, when Jo rmungandr (the Midgard serpent) uncoiled. Thetwo adverbial phrases Eynfis o ndri, in Eynfirs ski (on theboat) and at sandi, on the sand can each be taken with eitherverb, though as we picture the scene of rrs fishing for theMidgard serpent the first seems to go better with l and the secondwith rakisk.

(2) Now [there] is for steed-logs of the sea [i. e. for sea-warriors]N er jdraugum gis

eagles flightand rings, [i. e. birds of prey are gathering,arnar flaugok bauga, a battle is taking place]

I think that invitation they will receivehygg ek at heimbo iggi

of god of the hangedover the field. [i. e. of inn inHangagosof vangi. Valhalla]

In this verse attributed to Hvarr halti the prose word-orderis obtained by exchanging the last two words in the first coupletwith the last two in the second. This is required not by the syntax,for it is grammatically possible to read the words in the originalorder, but by the sense, for bauga will not do as a parallel toeither arnar or flaug, and the adverbial phrase of vangi will notdo with iggi heimbo Hangagos. The battle is taking place onthe field, and the warriors will receive inns hospitality andplunder. And although the order is required by the verse-form(to provide the hendings), it is difficult to believe that it is thatthat has determined the order, for there are plenty of alternativewords that could have been used to express the meaning. Thepoet must have wanted the unnatural word-order.

Eysteinn Valdason describes rrs fight with the Midgard sepentas follows:

(47) So it happened/reacted planksSv br vir at sjurCOALFISH made run forward broadSEIR rendi fram breiar


16

EARTHS; out on the gunwaleJARAR; t at bori

Ullrs stepfathers (rrs) fists bangedUlls mg[s] hnefar skullu.

It is the syntax rather than the kenning that is obscure in thisverse. Br vi could be impersonal, it happened, or seir jarar(earths coalfish, i. e. the Midgard serpent) could be the subject,the serpent reacted; rendi is causative to make run and usuallytakes a dative object, but can also take an accusative object (thengenerally with the object being a liquid, to pour), or be intransitive.So it happened that the earths coalfish made the broad planksrun forward or so the earths coalfish reacted that the broadplanks ran forward. The meaning of the two interpretations isnot very different, and there seems no way to decide betweenthem. While there is little support for the use of rendi with anaccusative object, if sjur is the subject the verb needs to beemended to rendu. Finnur Jnsson achieves further complexityby taking the last line and a half as the at-clause, and breiarsjur rendu fram as the second main clause: the earths coalfishreacted so that out on the gunwale Ullrs stepfathers (rrs)fists banged; the broad planks ran forward. This is the sort ofinterweaving of clauses that Kock condemns as an unnatural andover-academic interpretation, though it certainly can be arguedthat the natural meaning of the verse is that the serpents reactioncaused rrs fists to bang on the gunwale, rather than that itwas the cause of the boat moving forward; but Finnurs versiondoes involve assuming a most unnatural word-order.

Arnrr jarlaskld in orfinnsdrpa says:

(297) Bitu sveren ar *uru [MS urir]unngjo r fyrir Mo n sunnanRo gnvalds kindund randirramlig flkins gamla.

The most obvious interpretation of this would be to take Rognvaldsins gamla kind, Ro gnvaldr the Olds descendant = Earl orfinnras the object of unngjor sver bitu, thinly made swords pierced,but from the context of the verse we know that it is not aboutEarl orfinnr being wounded, but about his victory. Bitu must


17

therefore be without an object and kind must be dative of advantage.Thinly made swords pierced for Rognvaldr the Olds descendant;and mighty armies pressed forward under their shields there.Would the possibility of a misunderstanding have been understoodas a joke? Fyrir Mo n sunnan, to the south of the Isle of Mancould go with either clause or both, and should perhaps beunderstood with ar.

As well as the separation of words belonging to the same phrase,there are clear examples of the parts of a compound word beingseparated from each other (the classical term for this is tmesis),and even of the separation of the parts of two compound wordswithin the same half-verse. If the separated first elements areplaced before other words than those they belong with, it can bedifficult to be sure which words they are part of, and there areexamples of elements of a pair of compound words beinginterchanged.

One of the clearest and most undeniable examples of tmesis is:

(101/34) var I- me jo tnum Then was I- among the giants-ur nkomin sunnan. -ur newly come from the south.

Iur is a form of the name of Iunn, the keeper of the godsapples of eternal youth, who was abducted by the giant jazi.The two parts of her name, although they exist as separate words,make no sense separately in this sentence.

Another is:

(93/4) - fyr -sko mmu not ago a short [time], i. e. long ago.

- corresponds to the English prefix un- , and like it is not normallyseparated from the word it negates. Both these examples are fromHaustlo ng, and it is interesting that some of the most confusingof these syntactical complexities are found in poetry on mytho-logical topics, though I cannot see that there can be any religiousreason for the poet to be obscure in these particular verses. Ifthere is a connection the poet might be considered more likelyto be parodying priestly language than just using it.

Not all kennings consist of base words with genitives; someconsist of compound words where the first half of the compoundreplaces a genitive. Tmesis dividing such compounds is evenmore difficult to understand when the elements of a word are


18

not only separated by more than a line, but are in reverse order.This example is from Eilfr Gurnarsons rsdrpa:(78) There into the forest against the forests

ar mo rk fyrir markarnoisily chattering wind [= current] they setmlhvettan byr settu(the slippery wheel-knobs [stones] did not)(ne hvlvo lur hlar)fishing-net shooting adders [spears] (sleep)hf- skotnara (svfu).

There they pushed shooting-snakes [spears] in the fishing-net-forest[river] against the talkative [noisy] fishing-net-forest wind [current].The slippery wheel-knobs [stones] did not lie asleep.

Here, hf- is taken as the first element in the compound hfmo rkand at the same time as the first element in the compoundhfmarkar (in both cases making a kenning for river, fishing-net-forest). Words that are to be taken as part of two separatephrases or clauses in Old English poetry are discussed by BruceMitchell in An Invitation to Old English 1995, 70. Bruce Mitchellspoils his argument for the existence of such constructions byaccusing those who do not accept them of being insensitive, whichis like accusing those who cannot see fairies of being blind; butaccording to Roberta Frank in Old Norse Court Poetry 1978,112, a similar feature occurs in Japanese poetry. There is a furtherpossible example in Skldskaparml in verse 104, quoted above.

In this verse from lfr Uggasons Hsdrpa the parts of thekenning-compound are again separated by two lines:(64) Helpful in counsel acts gods

Rgegninn bregr ragnastrip of land at Singasteinnrein- at Singasteinirenowned against mighty slyfrgr vi firna *slgjan [MS slgjum]Frbautis son defenderFrbauta *mo g -vri. [MS mo gr]

Here rein goes with vri to make a compound: frgr rgegninn


19

ragna reinvri, the renowned helpful in counsel defender ofthe gods strip of land, i. e. Heimdallr, defender of Bifrost. Tmesiscan be avoided, as Kock points out, by replacing rein with thegenitive reinar (Notationes norrn 1952), though this makesthe metre less normal.(260) Fjarlinna fannir

fast vetrlii rastar;hljp of *hna -gnpurhvals *rann- ugtanni.

The winter-survivor [bear] of the current [i. e. the ship] waded fastthe drifts of fjord-serpents [waves]; the greedy-tooth [bear] of themast-head [i. e. the ship] ran over the whales house-tops.

This is attributed to Marks Skeggjason and describes a sea-voyage. The kennings in the second couplet are hna [MS hvta]ugtanni and hvals ranngnpur. Emending rann [MS ann] toranns would get rid of the tmesis, but the two kennings remainwith their determinants effectively exchanged. Cf. Frank 1978,467. Another striking example of double tmesis is found in anotherverse of rsdrpa:(90) . . . salvani- -Synjar

sigr hlaut arin- *-bauti. [MS arinbrauti]The subject of sigr hlaut, gained victory is a kenning for rr,arin-Synjar salvanibauti, beater of the frequenter of hearth-stone-Syns dwelling, though Kock (Notationes norrn 467)makes it into one for giantesses, acc. with the preposition of.Here again the tmesis (but not the exchange of determinants)could be avoided by making both salvani and arin genitive.

There is an instructive example of how the possibility of seeingmaximum complexity in a verse was dealt with in the thirteenthcentury in Snorris commentary to the next verse, which isattributed to Vga-Glmr; this commentary at the same timedemonstrates Snorris acknowledgement that the elements ofcompound kennings can be interchanged:(255) Rudda ek sem jarlar I fought my way like earls

or *lk vforum I had a reputation for this formerlyme versto fum Viris with weather-staves of Virirsvandar mr til *landa. wand to win lands.


20

The MS has lr in line 2 and handa in line 4. It is difficult toknow which clause forum, formerly goes with, but moresignificant is that Snorris commentary analyses the kenning forwarriors as stafir vandar Viris vers, staves of the wand ofVirirs weather instead of the more obvious stafir vers Virisvandar, staves of the weather of Virirs wand. Virir is a namefor inn, whose weather is battle and whose wand is the sword.Snorris interpretation involves exchanging ver and vandar. Suchinterchanging of the elements of a complex kenning is certainlyfound, however, as in the next verse, which is attributed to EgillSkalla-Grmsson, who seems not to have been averse to word-play. This is one of the best-known examples of tmesis:

(140) Upp skulum rum sverum, Aloft shall we make our swordslfs tannlitur, glitra; O stainer of wolves teeth, shine;eigum d at drgja we have daring deeds to do dalmiskunn fiska. in the valley-mercy of fish.

The stainer of wolves teeth is the warrior addressed in the verse,who gives wolves blood to drink from his dead enemies. Thelast line has to be read miskunn dalfiska, in the mercy of valley-fish; valley-fish are snakes, their mercy or grace is the summerwhen they come out of hibernation, and this can either beinterpreted as tmesis (dal- separated from -fiska) or as transferenceof one of the elements of a kenning from a determinant to thebase-word. This does give a more regular rhythm to the line, butcan hardly be said to be determined by metrical considerations.

An example of a particularly complex verse is this one, whichis another quatrain from Einarr sklaglamms Vellekla:

(34) Rushes wave before princeEisar *vgr fyrir vsa, [MS vargr]works of inn profit meverk Ro gnis mr *hagna, [MS ho gna]pounds mead-containers swelltr reris aldaalways seas against skerry of songs [i. e. teeth]aldr hafs vi fles galdra.

There seem to be three kennings for poetry here, each of which


21

is the subject of a sentence. Vgr (wave), verk (works), alda (wave)look like base-words, the first and last in reference to the meadof poetry. Ro gnis (inns), reris (the mead-containers) andhafs (the seas) look like determinants. Some commentators haveby emendation produced alternatives to some of these, but theydo not reduce the problems, which arise from the fact that thegenitives do not seem to be closest to the nominatives they gowith. Verk Ro gnis (inns work) would make a recognisablekenning for poetry, and so would reris alda (wave of the mead-container), but that leaves vgr hafs (or aldrhafs), wave of theancient (?) sea which would not. Unless one is prepared to guessat some radical emendation, it is difficult to avoid the conclusionthat fyrir governs not vsa but mr, and the kennings are vgrRo gnis (inns wave), verk vsa (princes work) and alda hafsreris (wave of the sea of the mead-container).

It is interesting that such complexity so frequently occurs inverses that are actually about poetry and composition, and thatthe actual content of this ingenious verse amounts to little morethan I am performing rewarding poetry before the prince. It isa performative utterance rather than an informative one. It is theritual that is important.

Word-play of various kinds is used by skaldic poets. There isan example of word-play using redundancy or deliberate tautologyin a kenning in a verse of Einarr Sklason:

(335) Hugins fermu bregr harmi The troubler of Huginns foodharmr. ends his trouble.

The troubler of Huginns food (carrion) is the one who eats it,i. e. Huginn himself, the raven; the trouble that he ends by eatingit is his hunger.

One device that Snorri mentions several times is called ofljst,literally (and ironically) too clear. Ofljst is a kind of word-playparticularly used for personal names, where the name, or an elementof it, is replaced by a synonym of the word as a common noun,e. g. Foglhildr for Svanhildr. Eilfr Gurnarson uses this device ina form reminiscent of crossword-puzzle clues for the name Hkon:

(36) You will have to, since of words*Veri *r, alls ora [MS verr ei]


22

for us grows about noble kinsman (Earl Hkon)oss grr of kon *mran [MS mrar]on mind-land of mead-container [breast, heart, sefreinu Snar where poetry is stored]seed, decide upon friendly giftss, vingjo fum ra.

There are two kennings in this verse, ora s (seed of words =poetry) and sefreinu Snar (the poets breast), though some peopleprefer to take them as s Snar (seed of the mead-container)and sefreinu ora (sedge-land of words = breast or tongue). Inaddition kon mran seems to mean Hkon (h- = high, noble,mrr = famous, glorious, noble; kon = kinsman, son, as well asbeing an element in the name Hkon). This kind of ofljst iswhere a word is effectively replaced by a synonym of a homonym.The same name is split into two parts by tmesis in a verse attributedto Queen Gunnhildr in Fagrskinna (F XXIX 75).

The name of inns consort Jo r (whose name means earth)is frequently replaced by a kenning and instead of talking aboutland or country, inns wife can be referred to, for example inthis verse from Hallfrers Hkonardrpa about Earl Hkonwinning land:

(10) Sannyrum spenr svera*snarr iggjandi viggjar [MS varr]*barrhaddaa byrjar [MS bjarr haddaa]*bikvn *und sik rija. [MS bifkvn of]

The keen wind-steed-[ship-]taker [sea-farer, Earl Hkon] lures underhimself [wins] with the true language of swords [battle] the pine-haired deserted wife of Third [inn; his deserted wife is Jo r,earth, i. e. the land of Norway].

The latent sexual imagery here and in other poems on the sametheme is likely to have been particularly relished by the earlsfollowers; it is unlikely to have any connection with myths aboutthe ruler wedding his realm. The gaining of the earls land isdescribed as rape, not marriage. The number of emendationsrequired in this and other verses in Snorris Edda implies thatscribes did not find them easy to understand.

Refhvarf is given particular attention in Snorris Httatal,


23

though it is not common in earlier poets, and some examples ofhis use of this device were mentioned earlier. This is where twowords are juxtaposed which have antithetical meanings, thoughin many of his examples the words are used with other meaningsthat are not antithetical, e. g. in verse 20:

Hlir hlr at stli,hafit fellr, en svfr elli(fer dvo l firrisk) harafram mt lagar glammi.Vindr rttr vir bendir;vefr rekr haf snekkjur;ver yrr; vsa ijur(varar fsir skip) lsa.

The bow [/warms] freezes at the prow, the sea [/lifted] falls but thetimber glides hard forwards against [/back] the waters uproar; thecrew [/movement] is deprived of rest. The direct [/straightens] wind[/twists] bends the sails; the cloth [/folds] drives [/unfolds] thewarships over the sea; the weather [/paces] whistles [rushes]; theship is eager for [/exhorts] harbour [/warns]; the labours reflectglory on the ruler.

Thus in this kind of verse words have to be understood in onesense in order for the reader or listener to understand the meaning,but in another in order to appreciate the effect of the antithesis.It is a series of puns.

Substitution of a word by one that is a near synonym occurs inngervingar. Snorris example is in kennings for gold of the typefire of the sea, which according to ch. 33 of Skldskaparmloriginated with the story of how gir, whose name means sea,used shining gold to illuminate his hall:

So this is the story of the origin of gold being called fire or lightor brightness of gir, Rn or girs daughters, and from suchkennings the practice has now developed of calling gold fire of thesea and of all terms for it, since gir and Rns names are alsoterms for sea, and hence gold is now called fire of lakes or riversand of all river-names.

It is in fact now difficult to know whether the kenning type fireof the water = gold was originally used with reference to thesea or to a river or, as Snorri maintains, with reference to gir.


24

Elsewhere Snorri uses the term ngervingar to mean extendedmetaphor, and this is effectively used by many poets. Hiscomplicated exemplification in Httatal verse 6 gives rise toanother of his detailed interpretations:

Svir ltr sknar naraslrbraut jo furr skra;tt ferr rgs r rttumramsnkr fetilhamsi;linnr kn svera sennusveita bekks at leita;ormr yrr vals *at varmri [MS r]vggjo ll sefa stgu.

The wise prince makes the adders of battle [swords] creep thescabbard-path [be drawn].

The mighty war-snake goes swiftly from the straight strap-slough[scabbard].

The sword-quarrel serpent can seek the stream of blood.The worm of the slain rushes along the minds paths [mens breasts]

to the warm war-river [flowing blood].This is extended metaphor to call a sword a worm and use anappropriate determinant, and call the scabbard its paths and thestraps and fittings its slough. It is in accordance with a worms naturethat it glides out of its slough and then often glides to water. Herethe metaphor is so constructed that the worm goes to find the streamof blood where it glides along the paths of thought, i. e. mensbreasts. Metaphor is held to be well composed if the idea that istaken up is maintained throughout the stanza. But if a sword iscalled a worm, and then a fish or a wand or varied in some otherway, this is called a monstrosity (nykrat) and it is considered a defect.

Nykrat (literally made monstrous) is thus a kind of mixedmetaphor. The effect of this is very similar to that of refhvarf, inthat it involves conceptual contradictions. Though Snorri condemnsit as unnatural (ok ykkir at spilla, and it is thought to be ablemish, Httatal 6/16) and his nephew lfr hvtaskld in theThird Grammatical Treatise 80 calls it a lo str, fault, a type ofcacemphaton, improper expression, many skaldic poets makeeffective use of the device (cf. Frank 1978, 52). Egill Skalla-Grmsson in one verse of his Ho fulausn refers to a sword assaddle of the whetstone, sun of battle, digger of wounds, blood


25

strip, sword-strap-ice. In a lausavsa which is one of the bestdescriptions of a sea-voyage in all skaldic verse, he calls the shipstem-bull and a sea-kings swan and the storm the giant of the mastand the wolf of the willow (F II 172; Turville-Petre 1976, 23):

The opposing (literally rowing in the opposite direction) mast-giant heavily strikes out a file with the chisel of storms before theprow out on the level stem-bulls path, and the cold-bringing willow-wolf grinds with it mercilessly in gusts Gestills swan round thestem in front of the prow.

The example in the Fourth Grammatical Treatise 131 is a versefrom Jmsvkinga saga in which a cudgel is referred to by fourdifferent kennings, though they do not actually conflict. Thereare other good examples in Einarr sklaglamms Vellekla(Skldskaparml verses 18, 278; see Foote and Wilson 1970,3656) and jlfrs Haustlo ng (Skldskaparml verse 92; seeMarold 1983, 191; 1993, 2917; Frank 1978, 468, 1578).

In this lecture I have tried to give an impression of the variouskinds of complexity that are to be found in skaldic verse, and ofhow difficult this makes the verses to understand. It seems thatthis complexity was one of the most highly valued aspects ofskaldic art. If we were not sure for other reasons that most skaldicverse was composed and received orally in preliterary times, thestyle and manner of the verse would suggest that it was literaryand meant to be read, as it clearly was by the time Snorris Eddacame to be written, as his treatment and discussion of it shows.Snorri was a very literary poet and scholar. If skaldic poetry wasreally composed and performed orally before that time, Vikingpoets were able to call upon a very high order of verbal ingenuityand skill indeed. Of course it is often argued that in an oral society,such skill, and the power of verbal memory that goes with it,was more developed anyway than it is in literate societies, andthere may be something in this. There are nowadays people whocan play chess without a chessboard and probably people whocan solve crossword puzzles without using pen and paper.Composition and comprehension of a skaldic verse without havingit on the page in front of one is not a more difficult achievementthan these. But the picture this gives of the Viking is of a reallyrather intellectual type, far from the wild inspired figure evoked


26

by Carlyle (in his work On Heroes), and I do not of course wantto suggest that Vikings were incapable of high intellectual activity.Actually I think that the Viking was really rather an intellectualchap and thought of things that would astonish you. And Vikingpoets were probably more skilled in verbal expression, in spiteof being less well educated in a formal sense, than Anglo-Saxonpoets, and even than early Welsh and Irish poets, perhaps eventhan Provenal poets, while being less idiosyncratic and crazythan the writers of Hisperica famina.

Consider these examples of complex interweaving of sentences.The first is from Halldrr skvaldris tfarardrpa:

(379) You were able there theirr knttu ar eira(you were never [shields vart aldrigi (skjaldarfire thundered though homes] of victory*leygr aut of sjo t) sigribereft) treasure to divide.sviptrgrsimum skipta.

The pattern here is abcba, with nesting of three statements eachof 56 words. Shields fire is a kenning for sword. (This verse islacking in the Codex Regius; the Utrecht MS has laugr in line 3.)

From Kormakrs Sigurardrpa:

(292) Heyri sonr Srarsannreynis fentannaO RR greppa *ltk uppi [MS ltr]jast-Rn Haralds mna.

Let the son of Haraldrs true trier [friend] listen; I cause to be heardmy yeast-Rhine of the men of the Sr of fen-teeth.

The kenning for the poem (the teeth of the fen are rocks, the Sr ofrocks is a giantess, whose men are giants; their yeast-liquid is themead of poetry) is a complex one made more complex by theinterweaving of its elements with the rest of the sentence. Orr,generous, being nominative, could go with sonr or with ek; Kockprefers to make it the first half of the compound orgreppa. If thefirst possibility is adopted, the syntactical pattern becomes abababab.


27

Another example from Haustlo ng:

(66) Did all, but (of Ullr)Knttu o ll, EN Ullarfrom end to end (because of the stepfather)ENDILG fyrir mgiground was battered by hailGRUND VAR GRPI HRUNDIN,the hawks sanctuaries [the skies] burn*ginnunga v *brinna [MS ginnjunga v hrinna]

The word en seems to link the two clauses, though Ullar fyrirmgi, because of rr) seems most likely to belong to the first(all the hawks sanctuaries did burn), leaving en isolated. Endilgmost naturally goes with grund, giving a pattern of elementsabababa. Here too it might be possible to think of Ullar fyrirmgi as belonging with both clauses or with the whole statement,rather than choosing between them.

The twelfth-century Icelandic poet Einarr Sklason seems tohave particularly liked certain kinds of complexity in his verse,and in these two examples he is describing an ornate weapon;the ornateness of the article described seems to be mirroredthough not actually describedin the ornateness of the verse:

(147) Horns glorious child can I [Horn = Freyja; her child isHrrbarni kn ek Hornar Hnoss = treasure, the precious axe](we got a valuable treasure) possess,hlutum dran gripstra,oceans fire rests on damager [oceans fire = gold;brandr rymr gjlfrs grandi damager of shield = axe]gold-wrapped of shield;gullvfiu *hlfar; [MS hlar]seed (bears her mothers)-ssberr sinnar murBATTLE-SWANS granted me [battle-swans feeder = the ruler,SVANS unni mr GUNNAR who feeds ravens by fighting battles]FEEDER of Fris servants [seed of Fris servants is thefstr- GANDI Fra gold ground by Fenja and Menja]


28

(Freyrs niece eyelashes rain). [Freyrs niece = Hnoss, theFreys nipt br driptir. precious axe; her mothers eyelashes

rain is the golden tears shed by Freyja, i. e. the gold adornment on the axe]

Gullvfiu in line 4 could go with hrrbarni (it would fitsemantically with this word; it is babies you wrap up ratherthan axes) or grandi. In line 5 -ss is supposed to be the secondelement of the compound fstrss, though Kock seeks tosimplify the structure by reading fstrgandi as one word andtaking the kenning for gold as Fra ss. This verse seems tome to be one that is virtually impossible to understand withoutwriting it down and re-ordering the words. Freyjas daughterHnoss is twice referred to without being named in this verse; ineach case what is meant is hnoss as a common noun, meaninga treasure, referring to the axe. This is what is described aboveas ofljst.

Snorri himself uses constructions of great complexity in Httatal 98:

Veit ek verari er vell gefa,bro ndum beitaok ba snekkjurhra hrraren heimdregaunga jo fraen auspo ru.

I know of young princes that brandish swords and set up warshipswho are worthier of higher praise than a stay-at-home, ones whogive gold than one sparing of wealth.

Here line 6 is parallel to lines 34 and line 8 to line 2. The proseword-order would probably read the lines in the order 1, 7, 2,8; 3, 4, 5, 6. (Finnur Jnsson in Skjaldedigtning B II 87 takesthem in the order 1a, 7, 2a, 3, 4, 1b, 5, 6, 2b, 8.)

A device which seems very literary, and may be derived fromsome foreign model, is that of vers rapports. This verse isattributed to rr Sreksson, and Snorri quotes it simply toillustrate the kenning for Njo rr in line 6; he makes no commenton its form:


29

(59) Var sjlf sonar Became herself of her sonnama snotr una Did not come to loveKjalarr of tami Kjallarr (inn) tamedkvut Hami It is said that Hamir did notGorn bani Gurn the slayergobrr Vani the bride of the gods the Vanrheldr vel mara rather well horsesho rleik spara. bow-warfare hold back.

Here it is the meaning of the words and our knowledge of thestories referred to, rather than the grammar, that requires us totake the four statements as consisting of lines 1+5, 2+6, 3+7and 4+8. The pattern is thus abcdabcd.

Finnur Jnsson in his great comprehensive edition of skaldicverse offered interpretations that frequently choose the mostcomplicated ways of interpreting it even when simpler analysesare easy to see. E. A. Kock compiled a huge set of notes (Notationesnorrn) on the interpretations of skaldic verses, in which hefrequently took issue with Finnur Jnsson for not choosing simplersolutions, accusing his interpretations of being desk inter-pretations rather than ones that would be natural to oral poets,and a rather ill-tempered dispute then took place between thetwo scholars. Many of Kocks interpretations, however, in spiteof being syntactically simpler, actually involve doing much moreviolence to the natural meanings of words and to natural grammarthan is involved in Finnurs tortuous word-order, though neitherhe nor anyone else has been able to deduce any grammatical ormetrical rules to explain the apparently arbitrary structure ofinterwoven sentences in skaldic verse. Neither Hans KuhnsSentence Particle Law nor his theory about the caesura (1983,8997, 188206) help very much, though they go some waytowards it. And Finnur seems to me to have more feeling andintuition for the Old Norse language than Kock. From anexamination of the complex and unnatural effects that areundoubtedly sometimes really used by Viking poets, and fromSnorris comments on them, it looks as though these poets didprefer complex meanings and effects to simple natural speech.

Nevertheless, I am doubtful about whether Viking poets everintended verses to have more than one interpretation (as for


30

instance in verse 193 quoted earlier), and whether that kind ofambiguity was part of the acknowledged skill in the activity;ofljst uses words with double meanings, and the meaning isoften concealed, or turns out to be other than at first seemedimplied, but the verse as a whole only ever means one thing.There could always, of course, have been disagreements in theaudience about how a particular verse ought to be read, but itseems that both poets and audiences wanted complexity.

Snorri speaks of the need for kennings to be in accordancewith nature (Skldskaparml ch. 33: This is all consideredacceptable when it is in accordance with genuine similarity andthe nature of things [me lkindum ok eli]). But skaldic verseis in its essence unnatural, in diction, word-order, and grammar,and in ch. 33 Snorri is talking about word-substitution, notmetaphor or images. The parallel with Viking art, particularlythe interlace ornament that was popular throughout the VikingAge, has often been made. The Vikings visual art, like theirverbal art, contains representations of objects in the real world,but stylised and made complex to such an extent that it becomesvery difficult to interpret, and the real nature of what is depictedseems to be deliberately concealed rather than revealed by theart. The reason for this, as with the use of runes, seems to berelated not to the desire to be esoteric and hieroglyphic for religiousor social or any other reasons, but to an aesthetic preference forcomplexity and puzzlement. Skaldic poetry seems to have beena restricted code to an even less extent than Anglo-Saxon or OldFrench poetry, which was probably confined to the aristocraticor religious ranks in society, whereas poetry seems to have beenan almost universal activity among the Vikings, both as composersand audiences.

I have already mentioned Carlyles notorious but unfortunatelyinfluential lectures On heroes, hero-worship, and the heroic inhistory (1841). He valued Old Norse poetry for its visionaryqualities, and describes inn as the originator of Old Norsepoetry in this way in his first lecture:

Strong sons of Nature; and here was not only a wild Captain andFighter; discerning with his wild flashing eyes what to do, with hiswild lion-heart daring and doing it; but a Poet too, all that we mean


31

by a Poet, Prophet, great devout Thinker and Inventor,as the trulyGreat Man ever is. A Hero is a Hero at all points; in the soul andthought of him first of all. This Odin, in his rude semi-articulateway, had a word to speak. A great heart laid open to take in thisgreat Universe, and mans Life here, and utter a great word aboutit. A Hero, as I say, in his own rude manner; a wise, gifted, noble-hearted man. And now, if we still admire such a man beyond allothers, what must these wild Norse souls, first awakened into thinking,have made of him! To them, as yet without names for it, he wasnoble and noblest; Hero, Prophet, God; Wuotan, the greatest of all.Thought is Thought, however it speak or spell itself. Intrinsically,I conjecture, this Odin must have been of the same sort of stuff asthe greatest kind of men. A great thought in the wild deep heart ofhim! The rough words he articulated, are they not the rudimentalroots of those English words we still use? He worked so, in thatobscure element. But he was as a light kindled in it; a light of Intellect,rude Nobleness of heart, the only kind of lights we have yet; aHero, as I say: and he had to shine there, and make his obscureelement a little lighter,as is still the task of us all.

That is, he regarded the obscurity of Old Norse verse as a resultof its primitiveness. I regard the Vikings techniques of versecomposition as inspired, not what they say; and their obscurityas arising from sophisticated rhetorical complexity. If I wantedto find something to set against the Vikings reputation as raider,rapist and pillager to demonstrate how civilised and intelligenthe really was, I would choose first not his trading activities, noreven his navigational and organisational skills, nor yet his artistryand craftsmanship except perhaps in shipbuilding, but his verbalskills. I frequently find it necessary to remind myself, in the wordsof a better scholar than myself, that in our reading of medievalliterature we must always try to praise the right things for theright reasons. In accordance with that, I think we should praiseskaldic poets for their highly developed verbal skill rather thanfor their perceptions of their world or for their inspiration orimagination. And in any case I believe that verbal skill is muchmore to their credit than anything else they might be said to haveachieved. In the words of Bragi the Old (verse 300b): hagsmirbragarhvat er skld nema at? (skilful craftsman of versewhat is a poet but that?)


32

REFERENCES

Carlyle, T. On Heroes, Hero-Worship, and the Heroic in History. London1841. [Cited from the Centenary Edition, 1898]

Edda Snorra Sturlusonar. Ed. Finnur Jnsson. Copenhagen 1931.Finnur Jnsson (ed.). Den norsk-islandske Skjaldedigtning. Kbenhavn

191215.Foote, P. and Wilson, D. M. The Viking Achievement. London 1970.Frank, R. Old Norse Court Poetry: The Drttkvtt Stanza. Ithaca 1978.F = slenzk fornrit I ff. Reykjavk 1933 .Kock, E. A. (ed.). Den norsk-islndska skaldediktningen. Lund 19469.Kock, E. A. Notationes norrnae. Lund 192346.Kuhn, Hans. Das Drttkvtt. Heidelberg 1983.Macdonell, A. A. A Vedic Reader for Students. Madras 1917.Magns lafsson, De poesi nostra discursus. In Anthony Faulkes (ed.).

Two versions of Snorra Edda from the 17th Century I, 40815.Reykjavk 1979.

Marold, Edith. Kenningkunst: Ein Beitrag zu einer Poetik der Skalden-dichtung. Berlin 1983.

Marold, Edith. Ngerving und Nykrat. Twenty-eight papers presentedto Hans Bekker-Nielsen on the occasion of his sixtieth birthday 28April 1993. Odense 1993, 283302.

Mitchell, B. An Invitation to Old English and Anglo-Saxon England.Oxford 1995.

The Saga of Gisli. Trans. G. Johnston. London 1963.Snorri Sturluson. Edda: Httatal. Ed. Anthony Faulkes. Oxford 1991.Spearing. A. C. Readings in Medieval Poetry. Cambridge 1989.The Third and Fourth Grammatical Treatises. In Bjrn Magnsson lsen

(ed.). Den Tredje og Fjrde Grammatiske Afhandling i Snorres Edda.Kbenhavn 1884.

Turville-Petre, E. O. G. Scaldic Poetry. Oxford 1976.Wrenn, C. L. The Poetry of Caedmon. Proceedings of the British

Academy 1946, 27795.

Documents

FAULKES (1997)- Poetical Inspiration