III. Formal Features of Jónas Hallgrímsson's Poetry and the Present Verse Translations

1. Strophic Forms

About half of Jónas's poetry is composed in strophic forms without end-rhyme. These forms are the descendants of the ancient oral poetry practiced by all pre-literate Germanic peoples and mentioned by the Roman historian Tacitus as early as ca. 100 A.D. In earliest Germanic times, this traditional oral poetry seems to have been stichic, not strophic, i.e., it was through-composed like modern blank verse, not broken up into strophic or stanzaic units. Beowulf (from eighth-century Anglo-Saxon England) and the Hildebrandslied (from eighth-century Germany) are both stichic.

In Scandinavia, however, for reasons that have never been fully understood, this ancestral stichic poetry came to be articulated in strophes already in preliterate times.5 The transformation was more or less complete by the time written records begin and it was oral poetry of the strophic type that was carried from Norway to Iceland in the ninth and tenth centuries by the country's earliest Scandinavian settlers. In fact those settlers brought with them two distinct styles or types of strophic poetry.

One type is known today as eddic (or eddaic) poetry, after the so-called Poetic Edda (a collective name for poetry of this kind). Eddic poetry was composed in simple types of strophes, was always anonymous, and generally consisted of narrative or dialogue poems about the pagan gods of Scandinavia (Óðinn, Þórr, Freyr, etc.) or the great legendary heroes of the Germanic Migration Period (Sigurður, Brynhildur, etc.).6

The other type is known as skaldic (or scaldic) poetry after its chief practitioners the skalds, many of whom were semi-professional poets at the courts of the Norwegian kings. Skaldic poetry was composed in extremely elaborate types of strophes, is almost always attributed to named poets, and can be regarded as a species of occasional poetry that was produced to celebrate memorable events or to be recited on important contemporary occasions.7 Snorri Sturluson's Prose Edda was compiled in the early thirteenth century as a handbook for poets who wanted to keep this poetry alive and continue to work in its traditions.8

A. Eddic Strophes

There are two basic types of eddic strophe, fornyrðislag ("the manner of the old utterances") and ljóðaháttur ("the style of songs").

1. Fornyrðislag

Fornyrðislag (pronounced FORT-near-this-lahg) is the more ancient of the two types. It is the immediate descendant of the older stichic poetry and has enjoyed uninterrupted currency in Iceland for a thousand years, from the days of the Scandinavian settlement in the ninth century until the late nineteenth century, when its practice more or less lapsed. Jónas was intimately familiar with it, not only through the work of his immediate predecessors and contemporaries but through the medieval classics of the Poetic Edda (which he had studied with great care and which his own poems frequently cite).9 He makes use of this strophic form more frequently than any other.

strophe with metrical markup

(In this example, and those that follow, diagonal strokes above lines indicate stressed syllables.)10

In the present translations, as in Jónas's Icelandic originals, a fornyrðislag strophe (or vísa) normally consists of eight lines (ljóðlínur, braglínur, or vísuorð). In exceptional cases, following eddic precedent (reinforced by the influence of Jón Þorláksson's Icelandic translation of Milton's Paradise Lost), Jónas composed "expanded" strophes of ten lines or "abbreviated" strophes of six.

Syntactically, each line of a fornyrðislag strophe tends to contain a complete phrase, word-group, or breath-group. Enjambment is frequent, however, and this effectively counters any tendency toward choppiness or sing-song. The two halves of a fornyrðislag strophe are usually syntactically independent. Metrically, each line is a self-contained unit consisting of two stressed syllables and a variable number of unstressed syllables.11 The latter range in number from one (rare) through two, three, and four (normal) up to five and six (rare).12 These stressed and unstressed syllables can be arranged (with respect to one another) in a number of different rhythmic patterns, but there are almost never more than two stressed syllables per line13 and these stand like piers or pylons in the stream of sound, with the unstressed syllables swirling and eddying around them (see Example 1).

The present translations contain lines of six different rhythmic patterns or types, the first five of which are based on Jónas's usage (and early Germanic usage in general). These types may be classified as follows. (Important: the symbol x represents one or more unstressed syllables. The examples are all taken from "The Vastness of the Universe.")

Type A/ x / x(for example, the phrase "hushed in distance" or "venturesome poet")
Type Bx / x /("on wings of light" or "behind me I hear")
Type Cx / / x("to cast anchor" or "through dim greyness")
Type D/ / x x("dark cataracts")
Type E/ x x /("swifter than wind")
Type Fx x / /("as I dart on")14

Short lines in which one of the unstressed syllables is omitted sometimes occur (e.g., "spark of light" [Type A short]). Lines of Type A, D, and E (all of which begin with a stressed syllable) can be preceded by one or more extra unstressed syllables (this is known as anacrusis), e.g., "with steady courage" (Type A with anacrusis). In the translations, polysyllabic words (including noun compounds) participate in these six patterns with their normal English pronunciation, exhibiting no appreciable secondary stress (i.e., words like cataracts or star-clusters are pronounced

word with metrical markup


word with metrical markup

Thus "star-clusters shine" is a possible Type E line, in these translations, whereas "the star-clusters" is not a possible Type C line).15

To exemplify all this in concrete form:

strophe with metrical markup

In Jónas's originals, as in the Poetic Edda, lines of Type A predominate, thanks to the generally trochaic rhythm of the Icelandic language.16 In these English translations there is a more even distribution of the types, with (not surprisingly, in view of the "iambic" character of English) an increased proportion of lines of Type B.

Each pair of lines in a fornyrðislag strophe is linked by alliteration, i.e., the occurrence of the same sound at the beginning of several syllables. This device has the effect of highlighting the words — often nouns and adjectives — that bear the heaviest load of meaning. Since alliteration is obligatory in fornyrðislag and the position of the syllables in which it occurs is strictly regulated, it has a structuring effect similar to that of rhyme in other poetic traditions and can usefully be referred to as structural alliteration.17

The structural alliteration in fornyrðislag involves sometimes two, sometimes three of the four stressed syllables in a pair of lines. The third stressed syllable of the pair regularly participates in the alliteration and is therefore known as the "head-stave" (höfuðstafur). Either — or both — of the first two stressed syllables of the pair may alliterate with the head-stave; they are known as "props" or "supports" (stuðlar).18 The fourth stressed syllable of the pair almost never participates in the alliteration.19 In practice this means that the odd lines in a fornyrðislag strophe contain one or two alliterants, the even lines one. Thus (with head-stave and props indicated in boldface type):

strophe with metrical markup

It is of fundamental importance to remember that alliteration is a matter of pronunciation (i.e., it involves the linking of identical sounds) and not a matter of spelling. Thus the initial sounds in the English words "sigh," "cycle," and "psyche" alliterate (though they are spelled differently), whereas the initial sounds in the words "thigh" and "thy" do not alliterate (though they are spelled the same way).

In the present translations, a given English sound alliterates only with itself, e.g. [g] with [g], [m] with [m], [ʃ] with [ʃ] etc.20 (The pronunciation of the translator's Midwestern United States dialect is adopted as normative.) There is, however, a group of three important exceptions to this general rule.21 The first two exceptions accord with ancient Germanic (and to some extent with modern Icelandic) usage; the third is an ad hoc rule governing usage in the present English translations:

In medieval and modern Icelandic poetry (as in Old English poetry like Beowulf), stressed syllables usually occur at the beginning of words, with the consequence that alliteration is usually word-initial (as it is everywhere in Example 1).26 Modern English is a very different kind of language, however, with different structures and different resources. Stressed syllables often occur in the middle or at the end of words, with the consequence that in such cases alliteration — if there is to be alliteration! — must be word-internal (as in lines 3 and 5 of Example 3, or lines 3 and 7 of Example 5).27

strophe with metrical markup

Fornyrðislag seems a simple enough form, but its simplicity is deceptive. Firm control is needed if the lines are not to become prolix and nerveless. In 1848 the poet Gísli Brynjúlfsson put his finger squarely on the problem:

I can't write in fornyrðislag and I think it's a very difficult thing to do: such enormous imaginative force is necessary if the form is not going to degenerate into mere prose and lose all its poetic dignity. . . . Even Jónas himself hasn't fully satisfied its demands, except in his translations, because he lacks the necessary fertility of imagination. Among recent writers, Bjarni [Thorarensen] is the only one who has known how to make successful use of this form. (DH81)

Note that the alliteration pattern of a fornyrðislag strophe (which consists of eight lines alliterating in pairs) may be represented by the following notation: 2222 (i.e., 2 + 2 + 2 + 2).

In a late poem modelled on the "Höfuðlausn" ("Head Ransom") attributed to the 10th-century skald Egill Skallagrímsson, Jónas makes use of runhenda, a rhyming variant of fornyrðislag that was first employed extensively in the poem in question. In runhenda, lines rhyme in pairs (couplets) with masculine and feminine rhymes alternating ad libitum. Thus:

strophe with metrical markup strophe with metrical markup

In runhenda we find for the first time the collocation of structural alliteration and end-rhyme that characterizes almost all modern Icelandic verse and sets it apart from the verse of any other modern European poetic tradition.28

2. Ljóðaháttur

A ljóðaháttur [lee-OH-the-how-turr] strophe consists of six lines. In Jónas's poetry, lines 1 and 2, and 4 and 5, form normal fornyrðislag pairs, each line containing 3-7 syllables. Lines 3 and 6, however, are longer, 5-9 syllables, and each can contain (and almost always does contain) not two, but three stressed syllables. Two of these stressed syllables — any two — alliterate with each other. There is usually a syntactic break in the middle of the strophe.

strophe with metrical markup

Ljóðaháttur was a popular form in medieval times — roughly a quarter of the poems in the Poetic Edda are written in ljóðaháttur — but it seems to have fallen out of fashion between the 14th and 18th centuries, when it was revived, just in time for Jónas to take advantage of it. He uses it most often in poems written in Iceland before his departure for Copenhagen in 1832, though it also makes a few distinguished appearances in his later work.

Note that the alliteration pattern of a ljóðaháttur strophe (which consists of an alliterating pair of lines + a single line with internal alliteration + another alliterating pair + another single line) may be represented by the following notation: 2121 (i.e., 2 + 1 + 2 + 1).

B. Skaldic Strophes

The rules governing skaldic strophes are much more complex than those governing eddic strophes and Jónas makes use of such strophes a good deal less frequently.

The basic form, on which (in the course of time) there came to be many variations, is called dróttkvætt (literally, "for recitation at court"). Its origins in early medieval times may have been influenced by contemporaneous Irish poetry.

The dróttkvætt line has an intricate metrical structure that encourages subtle rhythmic counterpointing. The dróttkvætt strophe consists of eight lines, like a fornyrðislag strophe, with a syntactic break in the middle. But there are three important differences:

(1) Each line now contains not two, but three stressed syllables. The even lines almost invariably begin with a stressed syllable and every line ends with a trochee (/ ⏑). In the verse of the skalds, the portion of the line that precedes the final trochee conforms to one of the five fornyrðislag types;29 in Jónas, it almost always conforms to Type A (/ x / x).

(2) The number of syllables per line is much more predictable; there are usually six, more rarely five, occasionally seven or more.

(3) In odd lines, double alliteration is now obligatory;30 it can occur at the beginning of any two of the three stressed syllables. (In even lines, the headstave continues to occupy the first stressed position.)

All three of these features can be observed in Examples 9 and 10. (The symbols | and ‖ will be explained in a moment.)

strophe with metrical markup strophe with metrical markup31

The symbol ‖ represents the termination of an internal rhyme, the symbol | the termination of an internal consonance. In Icelandic these rhymes and consonances are known as hendingar ("hitches").

In every even line of a dróttkvætt strophe, two of the three stressed segments32 form a full rhyme (e.g., hat‖ ~ bat‖); in Icelandic this is called an aðalhending ("perfect hitch"). In every odd line, two of the three stressed segments form a consonance (e.g., hat| ~ bet|; in Icelandic this is called a skothending ("half hitch").33 Rhyming and consonating segments may be word final (ditch‖ ~ witch‖) or word-internal (torching ~ orchard).

The rules governing the construction of internal rhymes and consonances in these translations are relatively straightforward. They are based on Jónas's rules for himself, but simplified and regularized. (Bear in mind that rhyme and consonance are — like alliteration — a matter of sound and not spelling.)34

(1) In even lines, segments closing with a vowel sound (e.g., hoe [hou:] and row [rou:]) and segments closing with a consonant sound (e.g., home [hou:m] and roam [rou:m]) can occur and rhyme; in odd lines, only segments closing with a consonant sound can occur and consonate.35

(2) When segments close with a consonant sound (e.g., home [hou:m]), or a series of consonant sounds (e.g., homespun [ˈhouːmspʌn]), only the first consonant sound following the vowel must participate in the rhyme or consonance (for most people's ears, this is enough to establish a sound-congruence). A second and third consonant, if present, may participate, but it is not obligatory that they do so. Thus pass [pas] "rhymes" not only with mass [mas], but with mast [mast] and mastless [ˈmastləs], and luck [lʌk] "consonates" not only with whack [wak], but with wax [waks] and waxless [ˈwaksləs].

(3) Word-internal vowel rhyme is only possible when the vowel sound that rhymes is followed by another vowel sound (e.g., greyer ~ inveighing); it cannot occur when a vowel is followed by a consonant (i.e., the first part of phony does not rhyme with doe).36

(4) When an English word of more than one syllable appears in the translations of skaldic-style verse in this collection, only the syllable bearing the heaviest stress can participate in a rhyme or consonance.37

As a general rule, one of the two hendingar in a line of dróttkvætt must occupy the third stressed position in the line (and examples 9 and 10 observe this rule throughout). On rare occasions, however, in the oldest skaldic poetry, both hendingar are found in the two earlier stressed positions. Snorri Sturluson calls this license Fleinsháttur; the present translations occasionally avail themselves of it.

Note, finally, that the two stressed segments containing the consonance, in any given odd line, need not coincide with the two syllables that participate in the alliteration.

In addition to basic dróttkvætt strophes like the one in Example 9, Jónas wrote poems in a number of forms related to basic dróttkvætt. Among the translations in this collection, for instance, there is one example of hrynhenda ("the tumbling or falling meter"). This contains four stressed syllables per line instead of three and — unlike dróttkvætt in the hands of medieval skalds — tends to be stubbornly trochaic (which is what gives it its name):38

strophe with metrical markup

The great popularity of hrynhenda, over the centuries, has sometimes been held responsible — at least in part — for the dogged trochaicness of much modern Icelandic verse.

Another variation on the dróttkvætt strophe is known as draughenda. In classic draughenda, as described by Snorri Sturluson in the Prose Edda, an extra unstressed syllable was added to the dróttkvætt line either before or after its first stressed syllable. Jónas evolves a more supple form of this in which the added syllable can appear after the second stress as well as the first, or even be omitted entirely:

strophe with metrical markup

"Solstice" (the poem cited here), "Above the Ford," and the long poem "The Pipit" are composed in this modified form of draughenda.


5 See Winfred P. Lehmann, The Development of Germanic Verse Form (Austin: University of Texas Press and Linguistic Society of America, 1956), passim.

6 For an English translation of the major poems of the Poetic Edda, see Poems of the Elder Edda, tr. Patricia Terry, Revised edition (Philadelphia: University of Pennsylvania Press, 1990).

7 For a general introduction to the subject, see Lee M. Hollander, The Skalds (Princeton, N.J.: Princeton University Press, 1945). For a more sophisticated approach, as well as guidance in reading the poetry, see E. O. G. Turville-Petre, Scaldic Poetry (Oxford: At the Clarendon Press, 1976) and Roberta Frank, Old Norse Court Poetry: The Dróttkvætt Stanza, Islandica XLII (Ithaca: Cornell University Press, 1978).

8 For a complete English translation, see Snorri Sturluson, Edda, tr. Anthony Faulkes (London: J. M. Dent & Sons Ltd, 1987).

9 Some of the rules and procedures of medieval Icelandic poetry were not fully understood by Jónas and his contemporaries, nor did they appreciate the effects on the language of the "quantity shift" (hljóðdvalarbreyting) of the 15th century, which changed Icelandic from a language in which syllable quantity played as important a role as syllable stress, into a language in which stress alone was significant (as it is in Modern English).

10 In Germany and the English-speaking world, it is customary to print the structural units of early Germanic verse as half-lines separated by a typographical caesura. Example 1, treated in this fashion, would read as follows:

I am the speeding      spark of light
flung by God      from the forge of Chaos.
I soar on wings      swifter than wind
above the paths      of the pulsing stars.

Students of Beowulf or the Poetic Edda in English or German editions and translations will be familiar with this layout. But it is not the way this type of poetry — whether medieval or modern — is conventionally printed in Iceland.

Since, in both Old English and Old Icelandic manuscripts, poetic texts are written out continuously as if they were prose, some sort of editorial rearrangement is necessary if an editor or translator wishes to highlight the metrical structure of a work.

Leaving aside the question of whether the Anglo-German layout is the best way of handling Old English or Old High German poetic texts, one can certainly entertain serious doubts about its appropriateness for medieval Icelandic texts. Finnur Jónsson wrote in 1892:

It has become habitual among the Germans [Finnur might have added: "and the English and Americans"] to talk about "long lines" (Langzeilen), especially of eddic poetry. They collapse two lines together into one, printing them the way they print early German verse. This is absolutely wrong. "Long lines" have no place in ancient Scandinavian and Icelandic verse. This is shown clearly by the commentaries on the Háttatal (for example, the commentary on the 97th strophe). And in a short account of Einar Skúlason (12th century) it is explicitly stated that a dróttkvætt strophe contains eight lines. (Síb13)

See further Einar Ól. Sveinsson, Íslenzkar bókmenntir í fornöld, I ([Reykjavík]: Almenna Bókafélagið, 1962), p. 108.

11 For a clear account of the relationship between stressed and unstressed syllables in modern Icelandic verse, see Síb68-72 (where a careful distinction is made between word-stress and sentence-stress).

12 As Jónas's practice matures over time, the number of unstressed syllables per line tends to increase.

13 Only on rare occasions in Jónas's work does one come across a line (in a fornyrðislag context) that contains three unambiguous stresses, e.g.,

line with metrical markup


line with metrical markup

14 This pattern (the so-called iambic a minore pattern) does not occur in ancient Germanic verse. Examples of it occur sporadically in the fornyrðislag of modern Icelandic poets (for example, in the first line of the following pair by Jón Þorláksson:

lines with metrical markup[K100]);

and when Jónas composes a line like

line with metrical markup

he is certainly very close to it.

The appearance in the present translations of lines of this type is legitimated by its long and distinguished history in English verse (for example, in Matthew Arnold's "Sohrab and Rustum," where it is used in a masterful fashion).

15 Here the practice of the translations departs significantly from Jónas's practice in his originals. In Old Icelandic, trisyllabic words had (or were capable of having) strong secondary stress (e.g.,

word with metrical markup


word with metrical markup

and in medieval Icelandic verse, syllables with secondary stress could occupy full-stress positions. Jónas imitates this procedure, regularly deploying trisyllabic words of this sort as if they still had strong secondary stress — though in his day they no longer did (see Síb69, Bbk18-9). It is possible that in reading his poetry aloud he artificially heightened the stress on such syllables.

16 As Jónas's technique develops over time, in fact, lines of Type A become more and more frequent in his fornyrðislag strophes, giving them a smoother and more regular rhythmical contour.

17 As opposed to the nonstructural or sporadic alliteration characteristic of English verse since the Middle Ages, for example in Shakespeare's lines:

Let me not to the marriage of true minds
Admit impediments. Love is not love
Which alters when it alteration finds
Or bends with the remover to remove.

Structural alliteration is characteristic of all early Germanic verse. And indeed it seems only natural, after the Germanic Stress Shift had fixed the floating accent of Indo-European on the root syllables of words, that a poetic system would evolve that took advantage of this characteristic new feature by linking these syllables in recurring sound-patterns.

The fundamental role played in medieval Icelandic verse by structural alliteration (stuðlasetning, stuðlun) has been recognized since early times. Snorri Sturluson's nephew Ólafur Þórðarson hvítaskáld ("the White Poet," d. 1259) asserted in his Rhetoric (Málskrúðsfræði) that alliteration is

the basis of the metrical framework that holds northern poetry together, just as nails hold together a vessel built by a shipwright, which would otherwise consist of nothing but disarticulated planks. Similarly this device [i.e., alliteration] holds together the meter in poetry by means of the letters [he means the sounds] that are called "props" [stuðlar] and "head-staves" [höfuðstafir]. (See Den tredje og fjærde grammatiske Afhandling i Snorres Edda, udgivne. . .af Björn Magnússon Ólsen [København: Fr. G. Knudtzons Bogtrykkeri, 1884], p. 96 f.)

Structural alliteration remains a characteristic feature of Icelandic verse down to the present day (though its use is no longer obligatory, as it was in Jónas's time).

18 Since a line of fornyrðislag is so extremely short, containing only two stressed syllables, double alliteration in odd lines was not insisted upon. It was felt to be an ornament, if it could be managed, but never became obligatory.

19 If it repeats the alliterating sound of the third stressed syllable (the head-stave), the line is guilty of ofstuðlun ("over-alliteration").

Occasionally, in the Poetic Edda, the fourth stressed syllable in a pair of lines — and not the third — will contain the head-stave (see Einar Ól. Sveinsson, op. cit., p. 109). This never occurs in Jónas Hallgrímsson's poetry or in the present translations.

20 Here, as elsewhere in this chapter, letters in square brackets indicate actual pronunciation in the phonetic spelling of the International Phonetic Association (IPA).

21 There are also a few minor exceptions, e.g., Hr- at the beginning of Icelandic proper names alliterates with r- in English (and not with h- as in Icelandic).

22 Some scholars have thought that what "alliterated," in the case of vowels and diphthongs, was not the vowel but the absence of a consonant. Others have argued that at the time when the rules for alliteration were originally developed in Primitive Germanic (the ancestor language of both Old English and Old Icelandic), vowel sounds in orally-delivered poetry were preceded by the catch or click known as a "glottal stop" and it was this — rather than the following vowel — that alliterated. (See Hermann M. Flasdieck, "The Phonetic Aspect of Old Germanic Alliteration," Anglia 69 [1950], 266-87). Recently Kristján Árnason has speculated that perhaps at one time, in Primitive Germanic, "only identical vowels alliterated, but that the umlaut developments leading to large scale phonemic splits caused different phonemes to form equivalence classes, and this complex situation led to a generalisation that all vowels came to form the same equivalence class" (RD108). Whatever the explanation, the modern Icelandic poet's task is unquestionably simplified by the fact that any vowel sound alliterates with any other vowel sound.

One sometimes wonders whether the phenomenon of promiscuous vowel alliteration, as well as the phenomenon described in the next note, did not have its origin in the practical experience of poets in Primitive Germanic times, i.e., the experimental realization by the founders of this tradition that their language was relatively poor in words beginning with vowels — so poor, in fact, that in order to create enough alliterative possibilities, all vowel sounds would have to be treated as co-alliterant; whereas it was very rich in words beginning with the sound [s] — so rich, in fact, that these words could advantageously be divided into alliterative subcategories.

23 It is not, perhaps, entirely fanciful to speculate that even in Primitive Germanic times the exceptional resonance of these combinations — especially noticeable when they occur in alliterating sequences — caused them to be treated in this special way. Other explanations of the phenomenon have been offered. According to Flasdieck (see previous note), the k, p, and t in these combinations were more prominent sounds than the s that preceded them and constituted the actual "alliterative element of the groups" (p. 275). Another scholar, wondering whether things were not in fact the other way around, has suggested that in the groups sk-, sp-, and st- "the s probably had a special emphatic quality; in Italian, s in such combinations is known as 'impure s', and in German initial sp and st are pronounced as shp and sht" (Alan Bliss, An Introduction to Old English Metre [Oxford: Basil Blackwell, 1962], p. 11).

24In Modern Icelandic poetry the combinations sk-, sl-, sm-, sn-, sp-, and st- are known as gnýstuðlar ("clanging props") and none of them is supposed to alliterate with any of the others (or with simple s-). In Jónas's own practice, however, sl- and sn- (but not sm-) alliterate with each other and with st-, and sm- alliterates with sp-.

25 This rule has been devised because the sounds [w] and [wh] have virtually fallen together in the speech of most Americans (including that of the present translator), who no longer — unless they are striving to be hypercorrect — make a distinction in pronunciation between "weather" and "whether," "weal" and "wheel," "woe" and "whoa!"

26 An exception to this rule is that both elements of a noun compound can participate in the alliteration, e.g., "munarmyndum" in the fifth line of Example 6.

27 In determining where a word-internal stressed syllable begins, actual pronunciation — never etymology or spelling — is the guide and the "maximal onset principle" is used to determine syllabication (segmentation). Thus "inane" [ɪˈneiːn] alliterates with "nothing" (not with "able") and "cascade" [kæˈsceiːd] with "scathing" (not with "cave"). The pronunciation indicated in Webster's Third International Dictionary is accepted as the arbiter in tricky cases.

But consistency in this matter is not pursued beyond word boundaries. In actual phonetic fact a phrase like "an eel" is pronounced identically with the word "anneal" [əˈniːl] and therefore its stressed segment alliterates with "nothing" and "inane." By the same token, the stressed segment in a phrase like "this cape" [ðɪˈskeiːp] alliterates with "scathing" and "cascade." But in cases like this the present translations ignore the maximal onset principle, and the word-boundaries of standard English spelling are allowed to override the segment-boundaries of actual speech. Thus, in a line like "a red and wrinkled face," r- and wr- are fictively assumed to alliterate, whereas in order to have true acoustic alliteration the line would have to be rewritten "a drawn and wrinkled face" [əˈdrɔːnanˈdrɪŋkəldˈfeiːs].

28 Structural alliteration ceases to be a feature of Norwegian poetry in the 13th century and of English in the 15th, killed off by the influx of rhymed verse from the south.

29 Different analyses are possible; see RD95.

30 Sometimes Jónas nods; for example, line 27 of his very late "Leiðarljóð til herra Jóns Sigurðssonar alþingismanns" contains only one alliterant.

31 Note that the lines of the English translation, while imitating the rhythms of the the Icelandic original quite slavishly, take up a little more space on the page. The chief reason for this is that English spelling (unlike Icelandic) is not phonetic.

32 The word segments, instead of syllables, is used here because — in both Jónas's practice and in the present translations — the sequences of sounds that constitute skaldic rhymes and consonances often extend beyond the syllable-boundaries of actual speech. In Jónas's pronunciation, as in Modern Icelandic (and Modern English) generally — at least according to the so-called "maximal onset principle" — the first syllable of a disyllabic word ends with its vowel sound, and any consonant sound(s) following the vowel are articulated as the onset of the next syllable, e.g.: ta•la (Icelandic example), fa•tal (English example). In skaldic poetry, however, the boundaries of the sound segments involved in rhyme or consonance are non-coincident with these syllable boundaries, e.g.: tal|a, fat|al. On this whole question, which is naturally much more complicated than indicated here, see RD96-110.

33 It is permissible for consonances to be replaced by rhymes (e.g., in lines 5 and 7 of Example 9 or line 5 of Example 10), but not vice versa. Having three (instead of two) consonances in an odd line (e.g., harr|i hárr|ar kerr|u) is an optional ornament.

34 As they were for Jónas himself in a line like "að hann fremd‖arheit efnd‖i" (pronounced [aːð hanː frɛmˑd̥‖arheiːtʰ ɛmˑd̥‖ɪ]). The few phonetic transcriptions of Icelandic words in this work follow the system in Stefán Einarsson's Icelandic (Baltimore: The Johns Hopkins Press, 1949); it is a slightly modified version of the IPA alphabet.

35 There are no examples in Jónas's verse of skothendingar involving unlike vowels (whether followed by a consonant or not) of the type occasionally found in skaldic poetry, e.g.,

þrá| muna oss um æ|vi

(see RD97-103 for additional examples and discussion). But the whole subject of Jónas's usage and its relation to the tradition is in need of further study.

36 The only exceptions to this rule are consonant sounds standing at the beginning of the second element of compounds (dough‖ ~ glowworm) or suffixes like -ly (through‖ ~ unruly).

Jónas's rule for himself is slightly different. In certain circumstances, a vowel sound can rhyme with the same vowel sound followed by a consonant (e.g., in lines like

nó‖g er siglt, sem fyrr var ró‖ið


há‖m Almannagjá‖ar),

or even when both vowel sounds are followed by unlike consonants (e.g., in a line like

sæ‖kind þæ‖gum vindi).

37 In medieval skaldic poetry, syllables with secondary stress (see note 15) could participate in rhymes and consonances. Jónas frequently writes lines like

line with metrical markup

on the medieval model. Since it seems unlikely that, in everyday speech, Jónas would have given the segment -prúð- enough stress to chime effectively with búð-, he probably read such segments (when they occurred in his skaldic-style verse) with artificially heightened stress.

Adopting Rule 4 for the present English translations unfortunately means giving up the highly effective techniques of rhythmic counterpointing that are typical of skaldic verse (including Jónas's imitation of it), but it has the virtue of doing no violence to the stress patterns of standard English.

38 This meter developed in the early 11th century, perhaps under the influence of Latin hymnody and rhythmical prayer on the model of Thomas of Celano's "Dies irae dies illa." It is also known as liljulag ("the Lilja-meter") because of its use in a famous 14th-century poem of that name.

Copyright © 1996-8 Dick Ringler. All rights reserved.

Jonas' MS flourish for the end of a poem For technical assistance:
Library Technology Group
University of Wisconsin-Madison
General Library System