Mc Kinnell - The Making of Havamal

8/12/2019 Mc Kinnell - The Making of Havamal

1/41

Hvaml has 164 stanzas; the next longest eddic poems (Atlaml and Sigurarkvia in1

skamma) consist of 105 and 71 stanzas respectively.

For surveys of critical opinions up to the mid-1980s, see Evans (1986, 835), Harris (1985,210711 and 14748). Although there has been much disagreement about the origins ofHvaml,

even von See (1972a, 12) accepts that it was compiled using some pre-existing (oral) poeticmaterial, not all of which has been successfully integrated into the whole.

See the discussion in Jackson (1994, 35 n. 2); Dr Jackson notes in a letter to me that the3

capital beginning stanza 111 is smaller than the other two and suggests that it may not have been

THE MAKING OFHVAML

John McKinnell

The Traditional Sections of Hvaml

Hvamlis easily the longest surviving eddic text, besides being one of the1

most disparate, and as early as 1891, Karl Mllenhoff suggested that itshould be regarded as consisting largely of a sequence of six poems whose

only common feature is that inn is taken to be the speaker in all of them. Manyother scholars have followed that view since his time. Mllenhoffs six poems were2

I. Stanzas 179 (approximately): The Gnomic Poem;II. Stanzas 95 (or earlier) 102: inns adventure withBillings mr;III. Stanzas 103 (or 104) 110: inns adventure with Gunnlo;IV. Stanzas 111 (or 112) 137: Loddffnisml, the poem addressed to (anotherwise unknown man called) Loddffnir;V. Stanzas 13845:Rnatal, the list of secrets;VI. Stanzas 14663:Ljatal, the list of magic songs.

The Codex Regius has large initials (such as it usually uses only at the beginningsof poems) at the beginnings of stanzas 1, 111, and 138, and although the3


2/41

John McKinnell76

intended to be significant; but as Richard North has pointed out (1991, 126), the size of the capitalitself is less significant than that of the indentation left for it in the main text, and this is three linesdeep, as at the openings of stanzas 1 and 138.

See Neckel and Kuhn (1962, 34, 40); all quotations from eddic poems are taken from this4

edition. For a list of the paper manuscripts, see Boer (1922, Einleitung IX).

See Dronke (1969, 21417) for its probable use in the skaldic verse of Torf-Einarr (ca890).5

SkjBI, 23.6

seventeenth-century paper manuscripts have no textual authority, their scribesseem also to have recognized divisions of the text at these points, where they have

the headingsLoddffnisml(stanza 111) and Rnatalsttr insorRnattrins(stanza 138). But there are no manuscript indications of major divisions at4

the beginnings of Mllenhoffs poems II, III, or VI, and his theory does notaccount for all the stanzas in the text. At least some of those between stanza 80 andstanza 94 have little to do either with what precedes them (a catalogue ofmiscellaneous pieces of pragmatic practical advice) or with what follows (innsunsuccessful attempt to seduceBillings mr). This article will consider a numberof other views of the origins of the poem and will finally suggest one that seems to

me to explain its diversity.

Hvaml, Hugsvinnsml, and the Disticha Catonis

Klaus von See (see especially 1972b, 1999) regards the whole of Hvaml as aproduct of the Christian Middle Ages and argues that its gnomic sections havebeen influenced byHugsvinnsml, an Old Norse verse adaptation of the Vulgate

version of the Disticha Catonis , a popular work of moral advice in Latin verse. If

this is correct,Hvaml is essentially the work of a single poet and must be one ofthe latest poems in thePoetic Edda.

Some details of von Sees argument fail to withstand scrutiny. He maintainsthat abstract lexical terms such as hyggiandi intelligence (Hvaml 6/1) andmetnarpride (Hvaml79/4) in stanzas where there is a correspondence with

Hugsvinnsmlmust be derived from Christian thought and are therefore proof oflate date. But as Larrington points out (1993, 99100), this sense of hyggjandi alsoappears in Hamisml 27/2, a poem which probably dates from the late ninth

century, while the participial adjectivehyggjandi sensible also appears in orbjo

rn5

hornklofisHaraldskvi 11/4, ca900. Metnardoes not appear elsewhere in Old6

Norse verse before the twelfth century, but other similarly formed abstract nouns


3/41

THE MAKING OFHVAML 77

See Skj BI, 121, 119.7

See Dronke (1969, 4245);Atlakvias use of valhollas a common noun meaning exotic hall8

(sts 2/3, 14/11) suggests a date earlier than the adoption of the word as the name of themythological hall of inn, a sense that is first found in Eirksml 1/3 (ca954, SkjBI, 164).

Further on both sides of this argument, see Evans (1989), Hermann Plsson (1990), Khne9

(1983), Larrington (1991 and 1993), and von See (1975, 1978, 1987, and 1989).

do, for example, hagnar advantage (Einarr sklaglamm, Vellekla25/6, from ca986),snnar windfall (Vellekla 14/2), vornu warning (Atlakvia 8/4, date un-7

certain but possibly around 900). Larrington also demonstrates that the strand of8

Christian-inspired terminology that runs throughHugsvinnsml is notably absentfromHvaml. But the fact that these elements of vocabulary inHvaml fail tosupport von Sees theory does not amount to positive evidence againsthis view,and a proper assessment of it requires more detailed comparison between the texts.

In comparing texts containing traditional and proverbial wisdom, it is difficultto be sure of influences from one work to another, because similar fragments of

proverbial wisdom are likely to arise independently in different cultures. Pieces of

advice such as dont get too drunk, know when to keep your mouth shut, or bea reliable friend would be useful in any society (for these, seeHvaml 12, 27, and43 respectively). For this reason, when two comparable texts are in the same lan-guage, we must insist on similarities of image, phrase, or grammatical structure as

well as of thought before accepting that one has influenced the other.9

There are some striking verbal resemblances betweenHugsvinnsml and morethan one section of Hvaml, but if von Sees theory were correct, we shouldexpect these to fall into two groups:

1. Instances where theDisticha Catonis,Hugsvinnsml, andHvaml all sharethe same meaning.2. Instances where Hugsvinnsml shows a development of thought from

Disticha Catoniswhich is shared byHvaml.

In fact, we find three kinds of correspondence which suggest a quite differentrelationship between the texts. They are

1. commonplace sentiments shared by all three poems but expressed in differentways (A15 below);2. more distinctive resemblances betweenHugsvinnsml andHvamlwhichare not shared byDisticha Catonis(B16 below); and3. instances whereHugsvinnsml shares an idea withDisticha Catonisbut ex-

presses it in phrasing that is used inHvaml for a different idea (C14 below).


4/41

John McKinnell78

In the quotations below, correspondences betweenHvaml andHugsvinnsmlare indicated in bold type, while agreements between Disticha Catonis and

Hugsvinnsml are underlined.

A. Contexts Where All Three Poems Contain Similar Statements

A1.Gttir allar In all directionsr gangi fram before one goes forward

um scoaz scyli, one ought to look round,um scygnaz scyli; one ought to peer round;

vat vst er at vita because you cannot know with certaintyhvar vinir where enemies

sitia fleti fyrir. may sit in the hall in front of you. (Hvaml 1)

Um at litask A man needs to look roundarf mar alla vega in all directions

ok vi vti varask. and beware of disaster.Glo ggekkin Clear of judgementskyli gumna hverr, each man must be,

frr ok forsjll vesa. wise and fore-seeing. (Hugsvinnsml 80)

Quod sequitur specta quodque imminet ante, videto;illum imitare deum, partem qui spectat utramque.

(Have an eye to what is going on, and look in advance at what is coming; imitate that god( Janus) who looks in both directions.) (Disticha Catonis II, 27)

Gnomic poems in any tradition can be expected to advocate vigilance, since thewhole point of gnomic poetry is to put people on their guard. TheDisticha Catoniscouplet here urges us to look forward and back in time, and it lacks the sense of

potential danger that is found in both Old Norse texts.Hugsvinnsmlpicks up theidea of foresight from Disticha Catonis, but prefaces it with the idea of lookinground in allspatialdirections for imminent danger. This is also found inHvaml,but the idea of looking backward and forward in time does not appear; instead, wehave the vivid image of enemies sitting in the hall that a man is about to enter. Itseems probable that the Janus image in Disticha Catonis reminded the poet of

Hugsvinnsmlof a poetic phrase about looking in all directions that already existedin Old Norse poetry, and that he adopted this either fromHvaml or from some

other traditional source. There is no particular resemblance here betweenHvamlandDisticha Catonis.


5/41


A2. er horscr oco gull When a sensible, taciturn mankmr heimisgara til, comes onto the premises

sialdan verr vti vorom; theres seldom a penalty for the wary;vat brigra vin for a more reliable friendfr mar aldregi a man never gets

enn mannvit mikit. than great common sense. (Hvaml 6/49)

mlugrskal He ought to be taciturnok stilla orum vel and arrange his words well

s er vill gus st geta. who wants to get G ods love.ra krapt a loftier powerfr mar aldregi a man never gets

en vera tungu trr. than to be true in his tongue. (Hugsvinnsml 19)

Virtutem primam esse puta compescere linguam;proximus ille deo est, qui scit ratione tacere.

(Think it the foremost power to be able to control your tongue; he is nearest to God, whoknows by reason (when) to be silent.) (Disticha CatonisI, 3)

The value of knowing when to keep quiet is another commonplace of gnomicpoetry, but hereHvaml is quite different in spirit from the other two sources.

Hugsvinnsml 19/13 provides quite a close translation of the sentiment of theLatin, and ra krapt(19/4) is probably derived from Virtutem primam at thebeginning of the Latin distich.Hugsvinnsml 19/46 andHvaml 6/79 showthe same structure:

x (comparative adj. + accusative noun)fr mar aldregi en y (accusative topic phrase).

However, this has no counterpart in Disticha Catonis, and the sentiments inHugsvinnsml and Hvaml are quite different. In Hugsvinnsml the highestpower is integrity of speech, whereas theHvaml stanza is concerned purely with

the practical advantages of being cautious and sensible. Again, it seems likely thattheHugsvinnsmlpoet remembered the formula from a familiar stanza that began

with advice to speak little, but that he wanted to produce something more moralthan the hard-headed aphorism of theHvamlpoet. The opposite possibility that the Hvaml poet deliberately removed the moral element seems anunlikely thing for any twelfth- or thirteenth-century poet to do.

A3.Ef tt annan, If you have another (friend)

annz illa trir, whom you dont trust well,vildu af hnum gott geta, but you want to get some good from him,

fagrt scaltu vi ann m la, you must speak him fair


6/41

John McKinnell80

en fltt hyggia but be false in your thoughtoc gialda lausung vi lygi. and repay falsehood for a lie. (Hvaml 45)

Flrs manns orum, In a deceitful mans wordstt hann fagrt m li though he may speak fairarptu eigi eim at tra. you do not need to trust.

Glyslig or With specious wordslttu gegn koma; you should reply,

gjaltu sv lku lkt. thus pay him like for like. (Hugsvinnsml 42)

Qui simulat verbis nec cordis est fidus amicus,tu cui fac simile: sic ars deluditur arte.

(To a man who pretends in words but is not a true friend in his heart, you behave to him

likewise: thus guile is deceived by guile.) (Disticha CatonisI, 26)

All three poems share the general theme of repaying false friendship in kind, butinHvaml this forms a natural antithesis to the preceding stanza, on how to culti-

vate a real friendship (see C2 below), and this is not shared by eitherHugsvinnsmlorDisticha Catonis.HvamlandHugsvinnsmlshare the fair/false antithesis andthe image of repayment, neither of which is in Disticha Catonis. Again, it seems

probable that theHugsvinnsmlpoet has used a familiar Norse expression, derivedfromHvaml or from popular proverbial wisdom, and thatDisticha Catonis has

not influencedHvaml at this point.

A4.Mealsnotr Moderately wisescyli manna hverr, each man should be,

va til snotr s; never let him be too wise;rlo g sn his own fate

viti engi fyrir, let none know beforehand eim er sorgalausastr sefi. his mind will be most free from sorrow.

(Hvaml 56)

rlo g sn His own fateviti engi fyrir, let none know beforehand,

n um at o nn ali. or nurse anxiety about that.Flestir at vita, Most people know at mun flrvo rum that for those wary of sin

daui ok lf duga. both death and life will bring benefit.(Hugsvinnsml 125)

Multum venturi ne cures tempora fati:

non metuit mortem, qui scit contemnere vitam.(Do not pay much attention to the time of your coming fate; he does not fear death, whoknows how to despise life.) (Disticha CatonisIV, 22)


7/41


The idea that one should make the most of life without worrying about fate ordeath is another commonplace: Eat, drink and be merry, for tomorrow we die, as

the English proverb puts it, adapting Ecclesiastes 8. 15. But the rationale for thisoutlook varies.Disticha Catonis advises against worrying about when one will diebecause this life should be despised.Hvaml 56, by contrast, suggests that this lifeis happiest when one does not know ones own fated future (that is, all unfortunateaspects of it, not merely when one will die). Hugsvinnsml125/12 is identical

withHvaml56/45 in phrasing, but after conveying the first half of theDistichaCatonissense, theHugsvinnsmlpoet gives a different rationalization: there is noneed for the virtuous man to worry, because both life and death will bring him

good things. Again, it seems likely that theHugsvinnsmlpoet adapted and moral-ized a poetic phrase fromHvamlor some similar source, rather than that theHvaml poet suppressed a Christian moral in favour of an outlook which anyChristian hearer would have regarded as irresponsible.

A5.At hrom ul At a hoary sagehlu aldregi! never laugh!Opter gott Often its good,at er gamlir kvea. what old people say.

Optr sco rpom belg Often from a shrivelled bag of skinscilin or koma. perceptive words come. (Hvaml 134/510)

Eigi skaltu hlja, You must not laughef vill horskr vera if you want to be wise

at o ldruum afa. at an aged grandfather.Optat ellibjgr man, Often someone bent with age rememberssem ungr veit ekki til: what a young person does not know about:

kennir gott gumum. he teaches men what is good. (Hugsvinnsml 121)

Cum sapias animo, noli ridere senectam:

nam quicunque senet, puerilis sensus in illo est.(If you would be wise in mind, do not laugh at old age: for whoever grows old has the senseof a child in him.) (Disticha Catonis IV, 18)

The sentiment that one should not laugh at old people is commonplace in gnomicwritings see for example Ecclesiasticus 8. 6: Despise no man for being old. TheDisticha Catonis couplet personifies Old Age rather than focussing on an individ-ual old man, as the two Norse texts do, and its rationale is that the old should beexcused because they cannot help being childish. This is the direct opposite of the

reasoning inHvaml andHugsvinnsml, both of which follow the instruction notto laugh at an old man with a construction beginning with opt that expresses the

view that the old often have wisdom to teach. The poet ofHugsvinnsml might


8/41

John McKinnell82

have reversed the sense of his Latin original through sheer independence of mind,but it seems more likely that he was influenced by existing wisdom poetry in Old

Norse, perhaps byHvaml itself.None of these correspondences looks convincing as evidence for indirectinfluence onHvamlfromDisticha Catonis.

B. Resemblances between HvamlandHugsvinnsml that Are Not SharedbyDisticha Catonis

B1.

At hyggiandi sinni About his own intelligencescylit mar hrsinn vera a man should not be boastfulheldr gtinn at gei. but rather guarded in mind. (Hvaml 6/13)

Af hyggjandi sinni About his own intelligenceskyldi mar hrsinn vera a man should be without boastingnema grisk arfir ess. unless the need for this arises. (Hugsvinnsml71/13)

Insipiens esto, cum tempus postulat, aut res;stultitiam simulare loco, prudentia summa est.

(Be foolish when the time or circumstances demand it; to feign foolishness in the (right)place is the height of wisdom.) (Disticha CatonisII, 18)

The noun hyggjandi intelligence cannot be regarded as a late and Christian lexicalitem (see above). The texts ofHugsvinnsmlandHvamlare obviously relatedhere, but their warning against boasting about ones own wisdom is quite differentfrom the advice inDisticha Catonisto show wisdom by feigning foolishness.Hugs-vinnsml does keep the notion of there being an appropriate occasion, but it is forshowing off ones wisdom rather than for pretending to be a fool. In this case, it isdifficult to see why the poet ofHugsvinnsml should have deviated from his source

so decisively if he was not influenced by existing Old Norse poetry.

B2.svir mar The foolish manvakir um allar ntr stays awake for nights on end

oc hyggr at hvvetna; and worries about everything; er m r, then he is tireder at morni kmr, when morning comes,

alt er vl, sem var. all that was wrong still is. (Hvaml 23)

Hugsjkr mar The anxious mankvir hvetvetna, worries about everything,aldri honum dagr um dugir. day never does him any good.

(Hugsvinnsml 142/46)


9/41


Suspectus caveris, ne sis miser omnibus horisnam timidis et suspectis aptissima mors est

(Beware of suspicion, lest you be miserable all the time, for death is most suitable for thosewho are timid and suspicious.) (Disticha Catonis IV, 43)

HereHvaml andHugsvinnsml share some phrasing, as well as the idea that theanxious man is foolish to stay awake worrying because he will be too tired to makegood use of the next day (although inHugsvinnsml this appears in a truncated andallusive form which suggests that the sentiment may already have been a familiarone in Old Norse). TheDisticha Catonis sentiment, on the other hand, is aimedat those who are made miserable by suspicion rather than by anxiety and suggests,

rather aggressively, that they deserve to die. The idea of staying awake at night doesnot appear inDisticha Catonis.

B3.Vin snum Towards his friendscal mar vinr vera a man should be a friend

oc gialda gio f vi gio f. and repay gift with gift. (Hvaml42/13)

Hatri hafna; Do not promote hatred;Hl at o ngum; Laugh at no one;

Gjalt gio f vi gio f. pay gift in exchange for gift. (Hugsvinnsml 9/13)

Irasci ob re noli, neminem irriseris. [. . .]mutuum da.

Do not get angry about any matter, laugh at no one. [. . .] Give a loan. (Disticha Catonis,Breves sententiae 30, 31, 16)

HereHugsvinnsml derives its first twosententiaefrom adjacent statements in thebreves sententiae(brief opinions) that begin Disticha Catonis; it follows the first

loosely and the second exactly. Its third sententia, however, agrees exactly withHvaml and is quite different from the non-adjacent brevis sententia 16 inDisti-cha Catonis, which is suggested as its source by Hermann Plsson (1985, 2930).It is possible that the translator of Hugsvinnsml misunderstood this sententia,associating mutuum a loan with mutualis mutual and thus assuming the idea oftwo-way giving. But this mistake would be all the more likely if the proverbialadvice found inHvamlwas already traditional in Old Norse verse, since it wouldlead the Hugsvinnsml poet to assume that this piece of proverbial advice was

similar to one that he already knew. More probably, theHugsvinnsmlpoet maybe repeating a Norse aphorism that had no source inDisticha Catonis.


10/41

John McKinnell84

For example Fronte capillata: Possumus autem per capillas divitias intellegere (With hair10

in front: we can, though, understand hair to mean wealth); see Ruggerini (1988, 264).

B4.Fullar grindr I have seen full sheep-penss ekfyr Fitiungs sonom, belonging to Fitiungrs sons

n bera eir vnar vo l; now they carry a beggars staff;sv er aur (possession of) wealth issem augabrag, like the twinkling of an eye

hann er valtastr vina. it is the most fickle of friends. (Hvaml 78)

Hrsan mann I have seen a long-haired mans ek ho la lii; in a band of warriors,

var honum skalli skapar. yet a bald head was fated for him.Sv er mar thats what a man is likesem margt fjr who has lots of money

ok verr um sir snaur. and becomes needy at last. (Hugsvinnsml79)

Rem tibi quam scieris aptam dimittere noli;fronte capillata, post haec occasio calva.

(Do not let something slip that you know is suitable for you; opportunity has hair in front,but is bald behind.) (Disticha CatonisII, 26)

Disticha CatonisII, 26 is about opportunity rather than wealth: one must seize anopportunity before it is too late.Hugsvinnsmlpicks up the image of baldness but

changes the basic idea. Here the comment is on wealth and its mutability, and thisis also the subject ofHvaml78. The two Norse poems also share a formula whichusess ekin the first half-stanza andsv erfollowed by a sentiment about wealthin the second. The Norse sources are also more fatalistic: wealth is simply seen asmutable, and there is no sense that one can alter this by diligently seizing onesopportunities. The easiest way of explaining this change would be to suggest thatthe poet ofHugsvinnsmlknew that theDisticha Catoniscouplet was sometimesexplained in terms of material wealth, but was also familiar with an Old Norse10

sentiment like the one inHvamlabout the transitory nature of wealth.

B5.gan mann attract a good personteygo r at gamanrnom into a pleasant relationship with you

oc nem lcnargaldr, mean lifir. and while you live, learn the magic of being well liked.(Hvaml 120/57)

Kostum safna. Collect virtuesKynn ik vi ga menn. acquaint yourself with good people,

Vinn eigi lo st n lygi. do not practise sin or lying. (Hugsvinnsml 13/46)


11/41


Virtute utere. Practise virtue.Nihil mentire. Lie about nothing. (Disticha Catonis,Breves sententiae 35, 44)

Lines 4 and 6 in this stanza ofHugsvinnsmlare derived from separatedsententiaein the opening section ofDisticha Catonis, but there is no clear parallel there to line5, which somewhat resembles lines 56 in theHvaml stanza. Ruggerini (1990,284) suggests that line 5 inHugsvinnsml may have been intended to parallel brevis

sententia42,Existimationem retine(keep your reputation), but it is more like thesentiment inHvaml, and a number of other breves sententiaeare also omittedfromHugsvinnsml (nos 5, 33, 3638, and probably 16), so there is no necessityto insist on a source in Disticha Catonis. The rest of this stanza of Hvaml is

without parallel in eitherHugsvinnsml orDisticha Catonis. It seems probable thatthe Hugsvinnsml poet, in describing the characteristics of moral integrity, hasechoed one of the pieces of advice in Hvaml of which it was possible for aChristian to approve.

B6.Illo feginn Pleased by evil

vertu aldregi, never be,en lt r at go geti. but rather rejoice in good. (Hvaml 128/57)

Almanna lof Everyones praiseef eignask vill if you want to gainok heita gr me gumum, and be called good among men,

annars gfu at anothers misfortunefagna aldregi. never rejoice.

Ger at gu gaman. Be pleased by what is good. (Hugsvinnsml 120)

Si famam servare cupis, dum vivis, honestam,fac fugias animo, quae sunt mala gaudia vitae.

(If you want to keep a good reputation as long as you live, make sure to drive out of your

mind the joys of life that are evil.) (Disticha Catonis IV, 17)

Disticha Catonisputs forward a worldly reason for rejecting vicious worldly desires(i.e. concern for ones reputation), and this is reflected inHugsvinnsml 120/13but has no counterpart inHvaml.Hugsvinnsml 120/46 shares some phrasing

withHvaml128, and both contain a warning againstschadenfreude: one shouldrejoice over good rather than at the misfortune of others (or at evil more generallyinHvaml). The commentaries onDisticha Catonisshow that the Latin text wasalso sometimes understood in this way, but its basic sense is that one should avoid

pleasures that corrupt.


12/41

John McKinnell86

C. Contexts whereHugsvinnsmlDerives its Thought fromDisticha Catonisbut Uses the Phrasing of a Different Thought inHvaml

C1.Fir sns, Concerning ones own property er fengit hefr which one has gained

scylit maro rf ola. one should not put up with being needy.Opt sparir leiom, Often one saves for an enemy atz hefir lifom hugat, what was intended for a dear one,

mart gengr verr, enn varir. much goes worse than expected. (Hvaml 40)

Fengins fjr Of the property you have gainedneyttu framliga. make use boldly.

Vertu ns mildr matar. Be generous with your food.Aura njta Allow your goldlttu auma fira, to be used for the wretched

ef gerir arfir ess. if the need for this should arise. (Hugsvinnsml 60)

Fac sumptum propere, cum res desiderat ipsa:dandum etenim est aliquid, cum tempus postulat aut res.

(Make suitable payment when the case makes it desirable to do so: for something ought tobe given, when the time or situation demand it.) (Disticha Catonis II, 5)

The verbal resemblance betweenHvaml andHugsvinnsml is limited to the firstcouplet of each stanza, including the idea that the resources spent are those whichone has previously gained (which is not inDisticha Catonis). Otherwise, the threetexts express different ideas, though all are about the desirability of spendingmoney. InHvamlthe thought is that one should not hoard ones resources tothe extent of going without what one needs, because one cannot know who willinherit ones wealth.Disticha Catonissuggests that there are times and situationsin which it is proper to pay out money, but does not say what they are. TheHugs-vinnsmlpoet begins with an alliterating phrase about the money one has gainedthat is reminiscent ofHvaml, but then develops the thought ofDisticha Catonis

with the assumption that the proper occasion for giving is when alms for the poorare needed. This is what one might expect of a moral Christian poet who had both

Disticha CatonisandHvamlin mind, but wanted to Christianize the moralityof his Latin source and counter or rebuke the selfish attitude evident inHvaml.

C2.Veiztu, ef vin tt, You know, if you have a friend

ann er vel trir whom you trust well

ok vill af honum gott geta: and you want to get good from him:gei scaltu vi hann blanda you must share thoughts with himoc gio fo scipta, and exchange gifts,

fara at finna opt. go to visit him often. (Hvaml 44)


13/41


Ef vin tt If you have a friendann rvildr s, who is dear to you

fs hann gott at gera. encourage him to do goodOra inna Even if for your words hann kunni o ngva o kk, he is not grateful,

skaltu hann vi vammi vara. you must still warn him against sin.(Hugsvinnsml 25)

Cum moneas aliquem nec se velit ille moneri,si tibi sit carus, noli desistere coeptis.

(When you warn someone and he does not want to be warned, if he is dear to you do notgive up what you have begun.) (Disticha CatonisI, 9)

Here, the theme of theHugsvinnsml stanza is the same as that of the DistichaCatoniscouplet: one should continue to admonish a dear friend for his own goodeven when he does not want to listen.Hvaml 44 is far more practical and cynical:this is how you should cultivate your friend for what you can get out of him.

Hvaml 44/12 andHugsvinnsml 25/12 share the formula ef vin tt, ann. . . (if you have a friend, who . . .), and their third lines also sound similar, but

Hvamls self-centred af honum gott geta (to get something good out of him)may have been deliberately adapted to the more Christian message inHugsvinns-

ml, where the focus is on the moral good the friend is encouraged to do ratherthan on what the person being advised can gain from him.

C3.snotr mar, The foolish man,ef eignaz getr if he manages to get

f ea flis munu, money or a ladys (physical) love,metnar hnom raz, pride grows within him.enn mannvit aldregi, but never common sense

fram gengr hann drigt dul. he goes straight on in folly. (Hvaml 79)

Fgirni rangri From wrongful avariceskaltu firra ik. you must distance yourself.

Ljt er lkams mun. Bodily lust is ugly.Orstr betra A better reputationgetr mar aldregi a man never gets

en vi syndum sj. than by bewaring of sins. (Hugsvinnsml 72)

Luxuriam fugito, simul etvitaremementocrimen avariti; nam sunt contraria fam.

(Flee from lust, and at the same time remember to avoid the crime of avarice; for they are(both) damaging to reputation.) (Disticha Catonis II, 19)


14/41

John McKinnell88

Here, the verbal resemblance between Hvaml and Hugsvinnsml is ratherremote, no more than the use in both of the words munu andaldregi at the ends

of lines 3 and 5 of the stanza. The idea in Hugsvinnsml andDisticha Catonisisthat lust and avarice are harmful to ones reputation for being a moral person. InHvamlthe viewpoint is quite different: here, money and a womans love are notseen as temptations to evil, but as good things that the fool never has the commonsense to make the best use of. Once again, it looks as if the translator ofHugsvinns-ml has combined the sense of his Latin original with a phrasal echo from a moreamoral wisdom verse that already existed in Old Norse.

C4.

athi n hltri With scorn or laughterhafu aldregi never treat,

gest n ganganda! a guest or traveller! (Hvaml 132/57)

go fga menn Unlucky peopleok o lmusur and paupers

skaltu eigi at hltri hafa. you must not hold up to laughter.(Hugsvinnsml 14/13)

Minorem ne contempseris. Do not spurn a lesser person.M iserum noli inridere. D o not mock a wretched man.

(Disticha Catonis,Breves sententiae 47, 52)

Hugsvinnsmlroughly follows the two non-adjacentsententiaeinDisticha Catonis,while the category of people at whom Hvamltells us not to laugh (guests andtravellers) is quite different. The phrasing shared byHvamlandHugsvinnsmlis limited to the expression at hltri hafa (to treat with laughter), which may betoo common an idiom to be worthy of notice.

HasDisticha CatonisInfluencedHvamlDirectly?

None of the distinctive similarities of wording or image between HvamlandHugsvinnsml can be shown to derive from Disticha Catonis, but von See alsoargues thatHvamlhas been influenced byDisticha Catonisdirectly at two points

where there are no close parallels between HvamlandHugsvinnsml (1972b,1012). One of these, betweenDisticha CatonisII, 5 andHvaml 40, is uncon-

vincing (see C1 above). The other is as follows:

Mikit eitt Not only big thingsscala manni gefa, should one give someone,

opt kaupir sr litlo lof; praise is often bought with little;


15/41


Ruggerini (1988) also suggests that the poet of Hvaml (or at least of the Gnomic Poem)11

might have been influenced byDisticha Catonis directly, though only by its first two books, whichhe may have known as a Latin school text.

Gylfaginning ch. 2 (Faulkes 1982, 8; Faulkes 1987a, 8).12

me hlfom hleif with half a loaf oc me ho llo keri and with a tilted bowl

fecc ec mr flaga. I gained myself a partner. (Hvaml 52)

Ne dubita, cum magna petes, impendere parva:his etenim rebus coniungit gratia caros.

(When you seek great things, dont hesitate to incur small expenses; for with these thingsfriendship joins those who are dear.) (Disticha Catonis I, 35)

These passages certainly show a resemblance, though perhaps rather a common-place one. But there is also a difference (apart from the image of sharing of foodand drink which is the most distinctive feature of theHvaml stanza). InDisticha

Catonis, the thought seems to be that one should onlymake gifts that are worth lessthan what one seeks to gain, whereasHvaml(for once) seems less self-centred the aim is to gain a friend, and small giftsas well asgreat ones are recommended.This correspondence hardly seems enough to demonstrate thatHvamlis directlydependent onDisticha Catonis, and the simplest explanation of the deviations of11

Hugsvinnsml from Disticha Catonis at most of the points where it resemblesHvaml is thatHugsvinnsml has usedHvamlas a minor source.

Transmission and Relative Dates

If von Sees theory were correct, we would also have to posit an improbably shortperiod of transmission between the writing of theDisticha Catonissource manu-script and the composition ofHvaml. Ruggerini has shown that the manuscriptofDisticha Catonisused by the poet ofHugsvinnsmlcan be assigned to the periodof transition from Vulgata recentiorto Vulgata recentissima, i.e. between the secondhalf of the twelfth and the beginning of the thirteenth century (1990, 286). Von

Sees theory would therefore require us to assume that this manuscript reached Ice-land almost immediately after the earliest date at which it could have been copied,thatHugsvinnsmlwas composed shortly after that, and that the composition of

Hvamlfollowed almost immediately. Even then, when Gylfaginningch. 2 (com-posed around 1225) quotes fromHvaml1 with the evident implication that itis a piece of traditional verse, we would have to regard this belief as mistaken,12


16/41

John McKinnell90

It has often been dated to some time in the thirteenth century (see Ruggerini 1988, 222 for13

a summary of critical opinions), but it could be later than that. For its detailed indebtedness toDisticha Catonis, see also Alexander (1931).

Diplomatarium IslandicumIV, 111; Hermann Plsson (1985, 16).14

See Tuvestrand (1977, 79); Hermann Plsson (1985, 16).15

For this view, see also Fidjestl (1992, 5), Khne (1983, 383). Hermann Plsson (1985, 23)16suggests that the relationship betweenHvamlandHugsvinnsmlis more complex, and that eachmay have influenced the other.

sinceHvaml1 could not then have been more than about half a century old atthe very most.

In fact, it seems more likely thatHugsvinnsml is a considerably later text. The13

earliest evidence we have for knowledge ofDisticha Catonis in Iceland is an entryin the mldagar (inventories of the movable property of churches and monasteries)for 1397, in which it is recorded that the monastery at Viey possessed a copy ofCato med Glosa. It seems most likely that this was a copy of the Latin text with14

an explanatory gloss in either Latin or Old Norse, though one cannot exclude thepossibility that it was a LatinDisticha Catonis followed by a copy ofHugsvinnsml.The earliest surviving manuscripts ofHugsvinnsml date from the fifteenth cen-

tury, and the work itself need not be much older than that. In this case, it would

15

seem that its poet has skilfully mingled his main source with echoes from the nativetradition of proverbial wisdom poetry that he derived from two different parts of

Hvaml, which he may have known in its present form, possibly even from theCodex Regius itself. This possibility seems to be strengthened by the fact that, like16

the present text of Hvaml but unlike Disticha Catonis, Hugsvinnsml namesitself in its last stanza and compliments its audience:

Hugsvinns ml A wise mans wordslt ek fyrir ho lum kvein I have had recited before heroes

ok kennda ek rekkum r. and I taught warriors good advice.Hyggins manns A thoughtful mans wisdomlsta ek ho lum speki. I have made known to noble men.

Hr er n ljum lokit. At this point the song is now finished.(Hugsvinnsml 148)

N ero Hva ml qvein Now the High Ones words have been recitedHva ho llo , in the High Ones hall,

allo rf ta sonom, very necessary to sons of men,o rf io tna sonom; useless to sons of giants;

heill, s er qva, good luck to him who recited,heill, s er kann! good luck to him who understands!


17/41


See also Ruggerini (1988, 23744).17

Although I disagree with her opinion, I would like to express my thanks to Dr Jackson for18

the warm generosity with which she has shared her ideas with me.

niti, s er nam, let him who has learned it, use it,heilir, eirs hlddo! good luck to those who listened! (Hvaml 164)

Alternatively, both Norse poems may draw independently on a common stock ofproverbial sayings in verse that were not always preserved together or in the sameform.

It thus seems unlikely thatHvaml has been influenced byHugsvinnsml, butprobable thatHugsvinnsml has borrowed fromHvamlor something like it.17

There is little to support von Sees argument that Hvaml is a single, very latepoem, though of course the rejection of his theory does not provide any positiveevidence that any part of the text is early.

Loddffnisml and Ljatal

Another view of the origins of the second half ofHvamlhas been put forwardby Elizabeth Jackson, who has argued that the whole ofHvaml 11164 is a single

poem, whose structure she characterizes as follows:

Frame identifying the main speaker as inn (st. 111)a) Section 1: gnomic advice (Loddffnisml, sts 11236)Transition stanza (st. 137)b) Section 2: on runes and sacrifice (Rnatal, sts 13844)Transition stanza (st. 145)c) Section 3: magic spells (Ljatal, sts 14661, 162/13, 163)Frame, again identifying the poem as inns words, spoken in his own hall and

preceded by a reminder of the presence of Loddffnir (sts 162/49 and 164).18

This analysis explains the metrical irregularity of the frame and transition

stanzas (though not of sts 14244) and the combination of gnomic and mytho-logical blocks of subject matter by suggesting that these were features of this genreof poetry, adducing Sigrdrfuml as another example of that genre. Jackson doesnot discuss the difficulty of reconciling the two apparent speakers and locations instanza 111 (see below), or the disruption of the metre and sequence of meaning inthe allusion to Loddffnir in stanza 162, but she does provide a persuasive


18/41

John McKinnell92

Compare Volsunga saga ch. 21 (Guni Jnsson and Bjarni Vilhjlmsson, 194344, I, 45;19

Byock 1990, 7172).

argument that stanzas 111 and 164 function as an introduction and conclusion tothe whole of the second half ofHvamlin its present form (1994, 3638).

The two stanzas that Jackson regards as transitional are also metrically odd,and neither has much to do with what follows it. The Loddffnir opening ofstanza 137 (in the variant of ljahttr known asgaldralag) repeats the refrain ofthe preceding gnomic section, but it is here followed by a sequence of medicalremedies expressed in a loose form of mlahttr. This is unrelated either to themoral advice that precedes it or to the account of inns sacrifice that follows.

Jackson suggests that the allusion to runes in stanza 137/14 may be intended toform a link with the magically effective runes of stanzas 14244 (1994, 39), but

since that passage is also in conflict with its surroundings, this again probably tellsus only about the work of some reviser subsequent to the original poet.Stanza 145 is also probably composite: its first half, which consists of five lines

of an unfinished ljahttr stanza, expresses the idea that it is better for humanbeings not to worship at all than to sacrifice too much, and seems, as Jacksonsuggests, to elaborate the theme of stanza 144, albeit in a different metre; it hasnothing to do with inns eighteen magic songs, which follow. The second halfof stanza 145 is a quatrain infornyrislag announcing what undr (inn) reistcarved when he rose up again (presumably after his journey to the world of thedead). Although this alludes to the sacrifice myth narrated in stanzas 13841, ittoo is in conflict with what follows, since it evidently sees inns newly acquired

wisdom as expressed in written runic inscriptions rather than in chanted spells.Jacksons second main argument is that the structure ofHvaml 11164 is

similar to that of Sigrdrfuml, which she analyses as follows:

Opening Frame: the narrative context of the encounter between Sigurr andthe valkyrie Sigrdrfa (sts 15)

a) Section 1: types of runic magic (sts 613, compare Ljatal)b) Section 2: the origins of runes (sts 1419, compare Rnatal)Transition: Sigurr commits himself to Sigrdrfa (sts 2021)c) Section 3: gnomic advice (sts 2237, compare Loddffnisml).

A lacuna in the Codex Regius begins after stanza 37, so a Closing Frame may havebeen lost at the end. To judge from the prose summary of the same material inVolsunga saga, made before this quire of the Codex Regius was lost, this probablyconsisted of an interchange of vows of betrothal between Sigurr and Sigrdrfa.19


19/41


20/41

John McKinnell94

For a brief summary in English of the arguments of Mundal and Myhren, see McKinnell23(2005a, 211). Further on the story-pattern, see McKinnell (2005a, 20017).

Fidjestl (1999, 215); the latest poet whose use of it is higher than the average is ttarr svarti,24one of the poets of Kntr inn rki (King of England 101635 and of Denmark 101435). I am

here ignoring the problem that some stanzas have probably been spuriously attributed to a muchearlier period than their actual dates. The relative consistency of Fidjestls results suggests thatthese are probably a small enough minority to be statistically insignificant except in a few specialcases.

waking of Sigrdrfa in Sigrdrfuml 1 contains two phrases (folvar nauir andHrafns hrlundir) that suggest that Sigurr originally woke her, not from sleep, but

from the dead. This would link her to a story-pattern whose usual form makes aliving child raise his or her opposite-sex parent from the dead in order to gainsomething from them, usually a sacred number of protective charms. The replace-23

ment of the mother figure with a sexual partner is quite common infornaldarsgurepisodes (McKinnell 2005a, 18695) and is probably another feature that suggestsrecent date; and the waking of Sigrdrfa from sleep rather than from the dead mayresult from Christian unease about sexual relationships between the living and thedead (McKinnell 2005a, 22831).

However, the most concrete reason for thinking that the poet of SigrdrfumlknewHvaml is the occurrence in both poems of the wordgamanrnar, whichappears in Old Norse verse only inHvaml 120/6 and 130/6 and Sigrdrfuml5/8. In Hvaml its meaning is figurative, describing the beginning of a lovingrelationship with a good woman (Hvaml 130) or the maintenance of a cordialrelationship with a friend (Hvaml 120). But Sigrdrfuml takes it literally, torefer to a spell which Sigrdrfa carves to make Sigurr love her. The easiest way ofexplaining this conflict of meaning is to suggest that gamanrnar was a rarefigurative expression which the Sigrdrfumlpoet knew only fromHvaml andmisunderstood as referring to a literal piece of runic magic.

The one apparent difficulty about the view that the poet of SigrdrfumlimitatesHvaml is to be found in the incidence of the expletive particle umor of.Bjarne Fidjestl (1999) has shown that in skaldic verse this becomes steadily rareras time goes on. A few early skaldic poets use it rarely (e.g. orbjorn hornklofi), butthere is no late poet who uses it often, and no poet later than the early eleventhcentury in whose work its incidence is greater than the average for the wholecorpus. Fidjestl suggests that the same test may also be applied to the dating of24

eddic verse, although only to give a general suggestion of relative date (1999,21828). A high incidence of the expletive particle might suggest early date, but


21/41


a low incidence would not necessary show that a poem is late. We should also re-member that, since the metres of eddic poetry are looser and more easily modified

than those of skaldic poetry, the texts we have may often have been modified byperformers long after their original composition.If Jackson is correct in takingHvaml 11164 as a single poem, it is necessary

to combine the figures for the expletive particle in what Fidjestl callsHvaml II(stanzas 11137) andHvaml III (stanzas 13864) before comparing them withthose for Sigrdrfuml. The results of doing this are as follows:

Poem Lines Average Actual Deviation Deviationincidence incidence per 100 lines

Hvaml 328 7.8 9 +1.2 +0.39(stanzas 11164)Sigrdrfuml 255 6 10 +4.0 +1.65

This would suggest that Sigrdrfuml might be older than Hvaml 11164.Admittedly, two of Fidjestls instances in Sigrdrfuml seem doubtful: in stanza12/45,r um vindr, r um vefr((one) winds them round, (one) weaves themtogether), the two instances of the particle are probably meaningful, so notexpletive. But even if these are discounted, the incidence in Sigrdrfumlis 8, the

deviation +1.3 and the deviation per 100 lines +0.56.However, when one looks at exactly where the expletive particle appears in

Sigrdrfuml, it emerges that most instances are concentrated in a very few stanzas.Six (including the two I have questioned) are in the middle sections of the length-ened stanzas 1213, the last two stanzas of the list of types of runic magic. In bothstanzas these middle sections create grammatical or stylistic problems: a suddenchange to an unexplained third-person subject in stanza 12, and a sudden changeof tense with a new, delayed subject in stanza 13. These obscurities lead Sijmons

and Gering (192431, III.2, 213) to suggest that stanza 12/46 and stanza13/410 probably come from an older poem that is otherwise lost, and this maybe correct. A smaller concentration, of two instances, appears in stanza 20, wherethe giver of gnomic advice is telling the listener to choose whether or not he wantsto hear what she has to say. Again, there is an incongruity, since Sigurrs reply instanza 21/13 Munca ec flia, | tt mic feigan vitir, | emca ec me bleyiborinn (I shall not draw back | even if you know I am doomed, | I was born

without any cowardice) suggests that he is expecting a prophecy about his own

future rather than some gnomic advice. Again, it looks as if these two stanzas mayhave originated in a different context from that in which they appear now. Theremaining two instances are both in the gnomic advice section, at 26/6 and 37/6.


22/41

John McKinnell96

A similar practice can occasionally be seen in skaldic poetry, e.g. in Arnrr jarlasklds or-25

finnsdrpa 22, where W haley (1998, 26566) points out clear and probably deliberate echoes ofVolusp41 and 57, and in Hallfrer vandrasklds erfidrpa for lfr Tryggvason (for which seeSkjBI, 15063).

The latter is the last surviving half-line of the poem, and probably marked the endof the gnomic advice in it, since the prose paraphrase of this part of Sigrdrfuml

in Vo

lsunga saga ch. 22 includes, in order, all the surviving pieces of advice foundin Sigrdrfuml but no others. It looks, therefore, as if the poet of Sigrdrfuml mayhave made a habit of quoting short snatches of older verse at the ends of sectionsof his poem. If these decorations are discounted, the background usage of the25

expletive particle is reduced to one or two instances well below the expectedaverage for a poem of this length.

The parallel with Sigrdrfuml is interesting, therefore, and it could even sug-gest that the combination of gnomic wisdom with practical magic was perceived

as generic by the (rather late) poet of Sigrdrfuml, but it does not prove thatLodd-ffnisml andLjatal are parts of the same original poem. The striking dichotomybetween the Christian-influenced advice in Loddffnisml and the decisivelyheathen-looking myth ofLjatal rather seems to suggest that they are not.

Whole Text Readings

It is perhaps not surprising, in view of the lack of critical consensus about the

origins ofHvaml, that two important studies (Clunies Ross 1990 and Quinn1992) have chosen to ignore the problem of origins and to attempt readings of thetext simply as we have it in the Codex Regius (though both are admittedly con-cerned primarily with other issues). This is self-evidently a valid approach, since thecompiler of the present form of the poem, at least, must have envisaged laterthirteenth-century readers reading it in this way.

In the course of a more general study of voice in eddic poetry, Clunies Rossuses narratological analysis to point out that the poem fairly consistently implies

two characters: a narrator who assumes the role of an archetypal wise wandererwith supernatural powers, and a narratee, sometimes identified as Loddffnir,who is assumed to have the role of human pupil and who is placed in a successionof imaginary roles, such as the wary guest (st. 7), the greedy man (st. 20), the earlyriser (st. 59), or the student of runology (sts 80, 14243). She also argues that thenarrative sections ofHvaml (which are of course first-person narrative) serve


23/41


to identify the narrator as the god inn, to explain how he gained his wisdom,and to mark a fundamental difference of status between him and the narratee.

This analysis is perceptive and useful as far as it goes, but it does not alwayswork. It probably does not matter that some of the roles envisaged for the narrateeare inconsistent with each other (for example, the sensible, taciturn man of st. 6 canhardly be the same person as the drunken, garrulous fool of sts 2127). It is, afterall, possible to regard the narratee as an Everyman figure, who may be seen as wiseon some occasions and foolish on others, or to imagine an audience consisting ofa number of narratees, among whom different individuals may be picked out bythe narrator at different moments. But we ought at least to expect, if the complete

poem had been composed as a consistent whole, that the narrator should remainself-consistent throughout. Clunies Ross discusses one stanza (st. 111) in whichthis appears not to be the case and concludes, as a sub-divided ego, he (ie, inn)experiences himself as both subject and object, narrator and actor (1990, 22829).But this does not really get round the fact that the narrator here is evidently a silentrecipient of the wise speech of others, not a speaker himself; and there is another

problem, too, namely that the strophe appears to be simultaneously located in twodifferent places, at the seat of theulr by the spring of Urr (lines 23) and insideinns hall (lines 910). Another narratological problem arises in stanza 143,

where a list of originators of runes begins with inn me som (inn amonggods) and ends with ec reist silfr sumar (I carved some myself). Whoever thespeaker is here, he must be someone other than inn.

Similarly, the passages of mythological narrative, while they certainly identifythe narrator as inn, do not always serve to explain the origins of his supernatural

wisdom. The first allusion to his theft of the mead of poetry from Gunnlo (sts1314) has nothing to do with occult wisdom but is used with what is clearlydeliberate bathos to point the rather obvious lesson that ale is not as good for the

sons of men as they say it is (st. 12), so that the best drinking parties are thosewhere one leaves while still in possession of ones senses (st. 14/46) but it needsno god come from Valho ll to tell us this. The story of inns failed attempt atseducing Billingrs maid (sts 96102) seems, rather oddly, to illustrate the immedi-ately preceding statement that no sickness is worse than to be satisfied withnothing (st. 95/46), but is actually more probably an illustration of the sexualunreliability of women in general, which has been stated in stanza 84. The secondaccount of the Gunnlo story (sts 10410) seems to form a pair with this, and to

illustrate the statement of the sexual treachery of men towards women in stanza 91;again, there is nothing supernatural about these pieces of worldly wise cynicism.The narrative of inns sacrifice of himself on Yggdrasill (sts 13841) does


24/41

John McKinnell98

For Snorris discussion of eddic metres, seeHttatal 95102 (Faulkes 1991, 3739; Faulkes26

1987a, 21720). The earlier collection of examples of different metres in Ro gnvaldr KalisHttaly-kill (Key to Metres, mid-twelfth century) seems to begin with examples of ljahttr,fornyrislag,and mlahttr, though the text is somewhat damaged, see Kock (194649, I, 239).

explain the origin of his eighteen magic songs (sts 14663), but has little to do withthe intervening stanzas 14245. Where the narrative sections do relate to parts of

the gnomic material, therefore, they are either clearly self-mocking or else are notadjacent to the gnomic material that they verify. One is left with the feeling thatnarratorial analysis can help us to make the best of a text which still remains lessthan satisfactory at a number of points.

In an important study of the use of changes in eddic metre to indicate changesof literary register, Judy Quinn argues convincingly that certain metres weretraditionally associated with particular kinds of material, and could therefore beused as signals to the audience to indicate what kind of verse they were hearing at

any given moment. The traditional metre known asfornyrislag (the metre ofancient deeds), which is common to all the early Germanic languages from whichverse survives, is used for impersonal narrative (as inrymskvia) or prophecy (asin Volusp); Quinn shows that the lengthened variant of it known as mlahttr(the metre of formal speeches) cannot in practice be consistently distinguished asa separate metre except inAtlaml(1992, 102). All the same, SnorrisHttatal26

distinguishes between the two, apparently regarding mlahttr as indicating ahigher level of formality, and also gives two further variants offornyrislag: whereas

fornyrislag itself has only one alliterating stress in each odd-numbered line, Starka-arlaghas two, whileBlkarlaghas two and consistently places the headstave (themain alliterating stress of the couplet) on the first syllable of each even-numberedline. The other main eddic metre, ljahttr(the metre of (magic) songs) has a pairof short lines followed by a long line that alliterates only within itself and istypically found in stanzas of six lines (that is, two groups of three). It is usually usedfor gnomic material (sometimes with a specific example in the first half-stanzaleading to the statement of a general principal in the second half, as inHvaml 14,or vice versa, as inHvaml 52), and also in dialogues where at least one partici-

pant is a supernatural figure (such as the wisdom contests in Vafrnisml andAlvssml and thosesennurthat involve deities, as inLokasenna, or giantesses, as inHelgakvia Hjorvarssonar1230) (Quinn 1992, 10406).Httatalalso includesa variant of it, calledgaldralag(the metre of magic spells) in which the last odd lineof the stanza is replaced by a couplet (see e.g.Hvaml 61), and we should perhapsadd another, in which there is an extra odd line either in the middle or at the end


25/41


For this view, see also Evans (1986, 8, 34).27

of the stanza, which I shall call variantgaldralag (see e.g.Hvaml149). There arein addition some ljahttrstanzas which are lengthened to nine lines (e.g.Hva-

ml 138) or shortened to three (e.g.Hvaml 147), and these devices can also beused as deliberate metrical ornament.However, Quinn then proceeds to another kind of variation for which Snorri

provides no guide:

Yet there is a related kind of modulation between verseforms that does not seem to implyany change in the speakers stance, relying instead on the discursive signification alreadyestablished in the preceding stanzas. Such is the case with thefornyrislag sections withinljahttr catechisms which usually occur at a point where an aggregate of items is pre-sented (for example, Hvaml 813, 857, 137, and 144; Grmnisml 28, 478; and

Sigrdrfuml 1517). (1992, 111)

One should perhaps addHvaml 8990 to Quinns list. The function of all thesepassages is certainly to give lists, but not all of them are really infornyrislag:Hvaml 137 begins with a half-stanza of variant galdralagwhich contains theLoddffnisml refrain, but follows it with fivefornyrislag couplets and an odd line,as if an attempt were being made to produce a monstrous version of a ljahttrstanza. Grmnisml 28 (which contains part of a list of river-names) consists ofthree lines of ljahttr, fourfornyrislag couplets and an odd ljahttr line. Thesecond half of Grmnisml 47 and the first half of Grmnisml 48 (which contain

part of the list of inns names) are best regarded as regulargaldralag. But it mayalso be significant that these stanzas are often adjacent to others that fulfil the sameor similar listing functions and are even more metrically irregular (seeHvaml 80,14243, 145; Grmnisml 27; Sigrdrfuml 14, 19). The refined metrical con-sciousness exemplified in Snorris work can, as Quinn shows, be used to highlightchanges of register between one stanza and the next, but it does not explain wildirregularity within individual stanzas. In a poetic culture that made fine distinc-

tions such as those betweenfornyrislag, Starkaarlag, andBlkarlag, there mustbe a suspicion that at least some of these oddities are the result of interpolation ortextual corruption rather than artful changes of register. Some changes of metre27

without changes of function may perhaps be due to the same cause, as may passageswhere a block of lines in a changed metre interrupts a clear logical sequence in acommon metre before and after it. The analysis of changes of metre certainly helpsus to see how thirteenth-century readers or hearers might have been able to gainadded appreciation from some of the metrical shifts inHvaml, but neither it nor


26/41

John McKinnell10 0

the study of modulations of voice can hide the flaws that remain evident in someparts of the text as we now have it.

Strata in Hvaml

We are thus driven back to Mllenhoffs basic idea that the text is an amalgam ofmore than one original poem, and its sheer length also suggests that this is likely.Perhaps the most cautious and sensible of the composite theories of the poem isthat put forward by Bjarne Fidjestl (1999, 218), who divides it according to thethree points at which the scribe has placed large capital letters of the sort that he

generally uses at the beginnings of poems:

I. Stanzas 1110 (Hvaml I)II. Stanzas 11137 (Hvaml II)III. Stanzas 13864 (Hvaml III)

This division has the virtue of proceeding only from a piece of objective evidence,though of course it cannot be certain that the scribe of the Codex Regius knew

what the original sections of the text had been. But if he had some idea that his

material consisted of more than one poem, it may be useful to consider how thesepoems came to be blended together. I would like to suggest that the compilationof the text may have gone through three main periods, whose remains can be seenin different archaeological strata within it. Just as a field archaeologist identifiesdifferent periods of occupation of a site by identifying sudden changes in soilcolour, physical features, or artefacts, a philologist may be alerted by abrupt andunexplained changes in metre, dislocations and contradictions of subject matter,or the interruption of a coherent sequence of thought or narrative which is then

resumed. Suspicions may harden into a tenable theory when more than one ofthese symptoms occurs at the same point. And just as an archaeologist worksbackwards in time, uncovering the most recent levels first, it seems possible torecognize three levels in the development of the text ofHvaml.

1. Some lines and stanzas seem designed to impose an apparent unity on thewhole collection. These emphasize inn as the speaker and try to add a contextfor his recitation of the whole collection, but they sometimes disrupt the meaningand symmetry of the passages into which they are interpolated. This stage may have

contributed stanza 80, stanza 111/13 and 910, stanza 162/49, and stanza 164.2. Before this, various scraps of verse seem to have been composed or inter-

polated (rather like running footnotes) at points where they seemed relevant to


27/41


Interpolated expansions of lists of poetic names may also account for the metrical irregu-28

larities in Grmnisml and Sigrdrfuml (see above), but poetic lists can sometimes be part of apoem s design, as in the lists of names of offspring inRgsula 1213, 2425, and 41; these couldhave been expanded to some degree, but the absence of any serious disruption to their argumentor metre means that there is no reason to suppose this.

topics already under discussion. These stanzas show a common encyclopaedictendency to add mundane detail without much regard for metrical consistency or

for the artistic form of the poems to which they have been appended. This impulsetowards encyclopaedic collection is a characteristic feature of twelfth- and thir-teenth-century learning throughout Europe, and in Old Norse it often manifestsitself in the form of poetic lists, like theulur at the end of Skldskaparmlor thecomposite list of names of dwarves in Volusp 1116. Some of these interpola-28

tions may have been made in more than one stage. I would suggest that theencyclopaedic stanzas are stanzas 8183, 8590, 137, and 14245.

3. When stages 1 and 2 are removed, a number of more or less complete poems

remains; these were presumably grouped together, possibly in oral tradition butprobably in a manuscript now lost. Their only common feature was that innwas imagined to be the speaker in most or all of them. Each probably began witha large capital letter in the lost manuscript, and three of these capitals (at thebeginnings of sts 1, 111, 138) have been copied into the Codex Regius.

A. The Work of the Editor

1. Stanza 80 seems to be made up out of fragments borrowed from elsewhere inHvaml. Its first five lines are related to stanza 142, of which they may be an oralvariant:

at er reynt, er at rnom spyrr, Rnar munt finna oc rna stafiinom reginkunnom, mio c stra stafi, mio c stinna stafi,eim er goro ginregin er fi fimbululroc fi fimbululr, (st. 80/15) og goro ginregin

oc reist hroptr ro gna. (st. 142)

(It proves true when you ask about runes, (You must find runes and meaningful signsthe ones known to the gods, very strong signs, very firm signswhich the magic powers made which the mighty sage paintedand the mighty sage painted,) and the magic powers made

and Hroptr of the gods carved.)


28/41

John McKinnell10 2

This is then combined with a single line about the value of keeping quiet whichresembles the advice offered to the fool in stanza 27/13:

hefir hann bazt, ef hann egir. (st. 80/6) snotr,er me alda kmr

at er bazt, at hann egi; (st. 27/13)

(He will grasp them best if he keeps quiet.) (When the fool comes among people,It is best that he should keep quiet;)

The awkward switch of subject from you to he may be due to the different con-texts from which the two parts of this stanza have been taken. The editor mayhave felt a need to associate all the major wisdom sections of his collection with

the esoteric knowledge connected with runes, and the advice to keep quiet, whichin stanza 27 is given to the fool, may have been imported here in order to linkstanza 80 with stanza 79, which also begins with a formula about the foolish man.

2. If the Odinic references were removed, stanza 111 would become a regularljahttr stanza in which an unidentified human speaker tells how he listened toadvice.

S ec oc agac, I saw and kept silent,s ec oc hugac, I saw and I thought,

hldda ec manna ml; I listened to peoples speeches;of rnar heyra ec dma, I heard secrets discussed,n um rom o go, they did not keep silent about advice;

[. . .] heyra ec segia sv: [. . .] what I heard said was as follows:(Hvaml 111/48, 11)

This would make a good, well-shaped opening for an advice poem (though notnecessarily one in which inn was the speaker) and may have been the originalbeginning ofLoddffnisml (Mllenhoffs poem IV). The editor seems to haveadded five lines to this, of which 13 associate what follows with inn by givingthe sage a seat by the well of Fate:

Ml er at ylia It is time to chantular stli , in the seat of the sage

Urar brunni at at the well of Fate

Theulr may be intended to be inn, as Clunies Ross has argued, but he is clearlynot the speaker of the stanza, and it is idle to speculate about how he comes to havea seat beside the well. The allusion to Urarbrunnrwas probably suggested by thereference to Yggdrasill, the Tree of Fate, in stanza 138. Lines 1011 Hva ho lloat, | Hva ho llo (at the High Ones hall, | inside the High Ones hall) seem tobe elaborated fromHva hollo (stanza 109/4). They introduce a further problemin the form of a second location (inside inns hall) which is inconsistent with


29/41


Further on the problems connected with this stanza, see Evans (1986, 2627).29

the first one; the repetition seems designed to associate the following advice withinn and to link some of the separate parts of the text with each other. 29

3. The text of stanza 162 also looks disturbed. The first three lines contain theseventeenth of inns eighteen magic spells; they make good sense as a stanza ontheir own. There then follow a couplet and four odd lines which seem to be elabo-rated from the refrain ofLoddffnisml. The easiest way of explaining these is thatthe editor has added them in order to link this section to a preceding one, but thathe has mistakenly inserted them in the wrong place, before instead of after the lastmagic spell. If this is correct, the editor was probably already working with a manu-script in front of him, since such a mistake would not arise in oral performance but

might very well do so on parchment.4. Stanza 164/2 echoes the phraseHva hollo from stanza 109/4 and stanza111/10, and this stanza was most likely composed by the editor to round off the

whole text, though some lines of it may be derived from the originalLoddffnisml.In general, the editor seems to have been determined to associate the whole

text with inn, and we should not automatically follow him in doing so for everypart of it. The stanzas that I have associated with the editorial stage of the poemsdevelopment are all metrically irregular, and three of them also produce problemsof content (an awkward change of grammatical subject in st. 80; two locations andan inconsistent narrator in st. 111; comment in st. 163/49 on a list of spells thatis not yet complete). In each case, removing the interpolated material also removesthe metrical oddity and the logical difficulty (assuming that st. 163 was originallya regular half-stanza; as the second-last of inns eighteen magic spells, it seemsto form a symmetrical pair with the second spell, st. 147).

B. The Encyclopaedic Stratum

Any stanzas that are attributed to the second, encyclopaedic stage must beassumed to have been added for a reason, and this usually seems to be the desire todevelop further a theme which has been touched on in the preceding stanzas. I

would suggest that the encyclopaedic stanzas are the following.1. A list of things which one should approve only when they are complete, or

of which one should take immediate practical advantage because they are unlikelyto be available for long (sts 8183). Unlike the preceding ljahttrstanzas (179),this list is in a form of mlahttr (or possibly, as Quinn suggests, rather heavy


30/41

John McKinnell10 4

fornyrislag). It looks like an addition which has been hooked on to the theme ofmutability in some of the preceding stanzas, especially stanzas 7678. If we ignore

the editorial stanza 80, it may have been placed after instead of before stanza 79by mistake, in which case we must assume that this interpolation was also made ina manuscript rather than in oral performance.

2. A composite list of things not to be trusted (sts 8587 and 8990 are infornyrislag, while st. 88, in ljahttr, looks like a secondary addition in the samevein, since the change of metre here does not signal any change in content). Thistheme of unreliability has probably been added to expand on the statement of theunreliability of women in stanza 84. As Quinn points out, stanzas 8587 contain

no verb, and in their present position they seem to depend on the phrase scylimangi tra no one should trust (st. 84/2). However, this does not necessarilyimply that they belong to the same original poem as stanza 84; they could havebeen added here as a kind of discursive footnote, with the change in metre beingused by an interpolator to mark his digression. However, if we regard stanza 88 asa secondary interpolation, the whole mlahttr list up to the end of stanza 89 leadsup to the emphatic:

verit mar sv tryggr, let no one be so credulousat esso tri o llo. as to trust in all these. (Hvaml 89/78)

The postponement of the single main verb is an effective rhetorical device, and ifstanzas 8587 and 89 are read in this way, they might have existed as an indepen-dent fragment before they were appended to stanza 84. Stanza 90 begins a newrhetorical structure, in which a beautiful but deceitful woman is compared with alist of other unreliable things, and it might be another secondary interpolation, butas there is no change of metre at this point, this must be regarded as no more thana possibility.

The interpolation of stanzas 8590 has obscured a clear and elegant balancebetween stanza 84 (on the fickleness of women towards men) and stanza 91 (onthe deceitfulness of men towards women), and this theme is maintained untilstanza 110. It is clearly different from the gnomic material of stanzas 179,although the heavy interpolation of this part of the poem obscures the beginningof the sexual comedy poem; this may explain why the Codex Regius scribe doesnot begin stanza 84 with an enlarged initial.

3. A list of medical remedies (stanza 137), added to the end of the generallyethical advice in the ljahttr stanzas ofLoddffnisml, but mechanically reusingthe refrain of that poem (on its metre, see above). Since this begins with whatseems to be a remedy against accidentally poisoned ale (Evans 1986, 131), it maytake its cue from the warning in stanza 131/7 to beware of excessive drinking.


31/41


See Hodgson (1973, liilxix; especially lviii). Further, see McKinnell (2007).30

Alternatively, since it follows a series of stanzas mainly concerned with the treat-ment that should be given to needy visitors (sts 132, 13536), it may have been

prompted by the thought that one reason for such visitations may be the need formedical help. At all events, this stanza is markedly different in both metre andcontent from the moral advice that precedes it and the mythological narrative thatfollows.

4. A probably composite fragment of advice about runes and heathen sacrifice.Stanzas 14243 are in a metrical form which cannot really be identified: each con-sists of a couplet like the beginning of a ljahttr strophe, followed by an inflatednumber of odd lines (five in st. 142, three in st. 143). Stanza 144 is a repetitive,

incantatory series of questions infornyrislag. Stanza 145 begins with five lines ofa ljahttr stanza, but where we would expect a closing aphorism in the sixth linethere is a full stop, followed by two couplets offornyrislagwhose content is unre-lated to what precedes them.

There is also a dislocation of the narrating voice in these stanzas. There is nodoubt that the imagined speaker of stanzas 13841 is inn, who says he was

geiri undar wounded with a spearoc gefinn ni, and given to inn,

silfr silfom mr. myself to myself. (Hvaml 138/46)

But in stanza 143 both inn and ec [. . .] silfr I myself are listed among theoriginators of runes, so that here they must be distinct characters. inn is alsoreferred to in the third person (asundr) in stanza 145/6. Ottar Grnvik (1999,2000) has suggested that the speaker throughout is a shaman-priest of inn whouses his suffering on the tree to achieve a mystic union with the god he worships;but this religious idea relies on a tradition of Christian mysticism based on medita-tion on canonical scriptures (which Old Norse heathenism did not have), which

was not widely known in northern Europe until the thirteenth century, and whichdid not surface in English and German vernacular works such as The Cloud of Unknowinguntil the fourteenth.30

To these problems we must add a severe disruption of content. Whereas stanzas13841 and 14663 form a complete and coherent account of the mythologicalexperience of inn himself and the mysterious magic spells he learned in the

world of the dead, stanzas 14244 are concerned with the practicalities and originsof runes. They have probably been added here to expand on the statement nam ecupp rnar (st. 139/4); in this context, rnar is best translated secrets, since it is


32/41

John McKinnell10 6

clearly oral spells that inn learns immediately afterwards (st. 140), but the inter-polator has understood them as written inscriptions. Stanzas 144 and 145/15

turn to the equally practical, human concern of how to make heathen sacrifice,though they are in different metres and seem to reflect rather different outlooks:stanza 144 is concerned with how esoteric runic knowledge and heathen worshipshould be practised, while stanza 145/15 express the view that it is better not to

worship the gods at all than to do so too much. After this, there is a sudden switchintofornyrislagthird-person narrative of how undr (i.e. inn) carved inscrip-tions for the fates of peoples when he returned from the dead (st. 145/69), andthis seems to have little connection either with what precedes or with the list of

magic songs that follows. This seems not so much like change of register as confu-sion arising from the collection and insertion of a disparate jumble of fragments oforal folk poetry. On the other hand, when stanzas 14245 are removed, the wordsand deeds, each of which found out the next for inn (st. 141/ 47), lead directlyinto the first of the eighteen spells, which are the material to which they refer.

C. The Main Poems

When we remove the material in the two later strata of the text, there remains theproblem of how many original poems there were. I would suggest that there wereperhaps four:

A. The Gnomic Poem (roughly sts 179);B. The Poem of Sexual Intrigue (sts 84, 91110);C.Loddffnisml Advice to Loddffnir (st. 111/48 and 11, sts 11236);D.Ljatal List of Magic Spells (sts 13841, 14661, 162/13, 163).

It seems probable that when these were first collected together in written form,each of them began with an enlarged capital in a manuscript that was the ancestorof the Codex Regius. Three of these enlarged capitals have been copied into thesurviving, interpolated manuscript; the fourth, at the beginning of Hvaml B,may have been lost because the presence of interpolated material all round it meantthat it was no longer obviously the beginning of a new section of the text.

The Gnomic PoemStanzas 179 are loosely structured, and there is no certainty that all these stanzas

are of the same age or origin, though some sequences clearly belong together. Thusstanzas 615 seem to be grouped mainly round the idea of what constitutessensible behaviour on the part of a guest, and these are contrasted with stanzas


33/41


Of the other first-person references, two could refer more specifically to inn, though both31are enigmatic: st. 49, where he says he gave his clothes to two trmenn, and sts 6667, where heportrays himself as often having been an unwelcome guest (possibly of giants?). Four relate toexemplary sights or experiences: st. 39, I have never met a man so generous that he wouldnt accepta gift; st. 52, with half a loaf and a tilted cup I got myself a companion; st. 70, Ive seen a fire

burning for a rich man when he was lying dead outside the doors; st. 78, Ive seen the pens ofFitjungrs sons full of animals, but now they are beggars (though Evans (1986, 113) suggests thatthis last instance may refer to a historical and political event). One is probably proverbial: st. 73, Iexpect a hand inside every cloak.

1617 and 1927, which concern various kinds of foolishness. Stanzas 4147 aremainly on the theme of friendship. Sometimes non-adjacent stanzas look from

their theme and expression as if they may have been associated with each other inoral performance (for example, st. 1, on the need to be prepared for enemies to bepresent when one enters a hall, with st. 38, on the need to have ones weaponsabout one at all times; and st. 79, on the reaction of the fool to good fortune, seemsto adapt the opening formula of sts 2327). Stanzas 76 and 77 may even be

variants of the same traditional stanza.However, in terms of voice and attitude, the Gnomic Poem seems reasonably

consistent. Although most stanzas present only a third-person description of some-

one in a typical situation, the narrator jocularly identifies himself as inn whenhe recalls how drunk he got when he visited Gunnlo (sts 1314). The other usesof the first-person pronoun do not obviously refer to inn and sometimes seemmore appropriate to an experienced human traveller, but none of them is actuallyincompatible with the god as speaker, though one seems rather unlike him:

Ungr var ec forom, Once I was young.fr ec einn saman, I travelled alone,

var ec villr vega; then I went astray in my ways;auigr ttomz, I thought myself rich

er ec annan fann, when I found someone else mar er mannz gaman. man is the joy of man. (Hvaml 47)

Unlike rr, inn is usually portrayed as one who travels alone, and who typicallyappears as an old man, while these words seem more appropriate to an experiencedhuman Everyman figure. The second-person pronoun produces fewer problems,31

since the person addressed can be consistently taken to be the less experiencedrecipient of the narrators advice (see sts 19 and 4446). Most importantly, how-ever, the viewpoint from which the advice is given remains consistently practical,

amoral, and unheroic; for example, the live man always gets the cow (st. 70/3) andno one has any use for a corpse (st. 71/6).


34/41

John McKinnell10 8

The Gnomic Poem is also quite consistent metrically. Galdralag is used in thefirst half of stanza 1, the second half of stanza 61, and as a half-length stanza by

itself in stanza 73 (though this could alternatively be regarded as an interpolatedquatrain offornyrislag) (Evans 1986, 109), and variantgaldralag in the first halfof stanza 74. Stanzas 6 and 27 have nine lines each, and stanza 65 only three. Butall of these fall within the boundaries of the metrical variations for deliberate effectthat I have discussed above, and in some cases there is an obvious literary point: instanza 1,galdralagis probably used to suggest the supernatural mystique of thespeaker; stanza 6 opens a new theme (how the responsible guest should behave);and stanza 27 is the end of the section about the fool. The unexpected shortening

of stanza 65 may be a joke; since its subject-matter is the potentially serious conse-quences of what one says to other people, the implication may be that the less thenarrator says to other people (that is, his audience), the safer he is likely to remain.

The date of the Gnomic Poem is hard to estimate. Adapting Fidjestls cal-culations for the incidence of the expletive particle to stanzas 179, one gets thefollowing figures:


Hvaml A 478 12.5 15 +2.5 +0.57This might suggest that the poem is rather early, and this could be correct. Thereseems to be an echo of stanza 76 or stanza 77 in Eyvindr skldaspillirsHkonar-ml, a funeral lay composed for King Hkon the Good ca96265:

Deyr f, Cattle die,deyja frndr, kinsmen die,

eyisk land ok l; land and people become desolate;(Hkonarml 21/13)

Deyr f, Cattle die,deyia frndr, kinsmen die,

deyr silfr it sama; the self dies in the same way;(Hvaml 76/13 and 77/13)

In a lay of praise composed for a funeral feast, it also seems probable that Eyvindrexpected his audience to bear in mind the second half of one of the same twostanzas ofHvaml (which may be oral variants of each other):

enn orztrr but reputation

deyr aldregi, never dieshveim er sr gan getr. for the man who gets a good one.

(Hvaml 76/46)


35/41


For a more detailed literary study of this poem, see McKinnell (2005b).32

ec veit einn, I know one thing at aldri deyr: that never dies:

dmr um dauan hvern. the fame of each dead man.(Hvaml 77/46)

This suggests that some version of stanzas 7677 probably already existed in thelater tenth century, but it is not possible to say whether it was already part of anoral text that resembled the Gnomic Poem or merely a floating poetic aphorism.

The Poem of Sexual IntrigueThis looks like a single, well-structured poem consisting of three parts that relateclearly to each other. After balanced opening statements of the unreliability of

women towards men (st. 84) and of men towards women (sts 9192), there is anequally balanced reflection on the folly and yet the irresistibility of love (sts 9394),followed by two exempla: that of Billingrs maid illustrating the unreliability of

women (sts 95102), and that of Gunnlo illustrating that of men (sts 10310).The narrator is clearly identified as inn by his first-person narrative of boththese episodes (except in sts 10910, where he mockingly adopts the viewpoint ofthe outwitted giant Suttungr), and the narratee effectively disappears, since thesecond-person pronoun is only used in the direct speech of Billingrs maid to

inn in stanza 98. Both mythological episodes look as if they have been adaptedin a jocular way. This is clear in the Gunnlo story, whose usual point the acqui-sition of the mead of poetry is ignored in favour of inns self-congratulationat having seduced the giantess and got away with breaking his oaths. It is harder tobe sure of in the story of Billingrs maid, which is not known from any othersource, but since that shows some resemblances to the story of inns ultimatelysuccessful seduction of Rindr, it may be that this story has also been adapted tomake a comic point. This comic approach may suggest a poet for whom the oldgods were no more real than, say, Venus and Adonis are in Shakespeares poem.32

The metre of the Poem of Sexual Intrigue is also quite regular, with only a fewvariations, which are probably all artful. Stanzas 102 and 103 each have nine linesand form a balanced fulcrum, since one concludes the episode of Billingrs maidand the other begins that of Gunnlo. The second half of stanza 105 is in variant

galdralag, probably to suggest that Gunnlos responsiveness to inn is magicallyinduced; and the first half of stanza 109 is ingaldralag, suggesting the magic whichmakes the frost giants fail to recognize, when they arrive at inns hall, that he isthe same individual as the fugitive they are looking for.


36/41

John McKinnell11 0

See Fidjestl (1999, 23145, especially 23841).33

See Kristjn Albertsson (1977); for a full discussion of this and other, less probable,34

suggestions, see McKinnell (2005b, 9597).See McKinnell (2005b, 97100). Similarly, von See (1999, 37475) points out that the35phrase hold oc hiarta (flesh and heart) in st. 96/45 can be closely parallelled in French and G ermancourtly love lyrics of the twelfth century.

The results derived for this poem from Fidjestls study of the criteria for datingeddic poetry are rather confusing. In stanza 106/12 Rata munn | ltomc rms

um f (The augers mouth | I caused to make room for me) there is probablyalliteration between originalr-in rumsand original vr-inRata, which may indicatea date of composition after the loss of the consonant combination vr- in WestNorse. However, this probably only indicates a date after ca900, and in any case,the etymology of ratihas been disputed. The figures for the use of the expletive33

particle seem at first sight to suggest an early date:


Hvaml B 140 3.7 9 +5.3 +3.79Onlyrymskvia,Alvssml, and Vafrnisml show higher incidences per onehundred lines than this, but much else suggests that this poem is rather late. Stanza84 probably includes a reference to the Wheel of Fortune, and stanzas 9195 also34

contain several probable echoes of OvidsArs Amatoria and one of VirgilsEclogueIX. These would all indicate a poet with an education in the secular Latin classics35

that can hardly have been available in Norway or Iceland before about 1100. How-ever, eight of the nine instances of the expletive particle in Hvaml B are in the

twelve narrative stanzas of the two mythological exempla; the ninth (brig bristum lagit deceit lodged in

Mc Kinnell - The Making of Havamal

Documents