III: BRAGIR

Rimur metres consist of four, three or two trochaic lines connected with alliterations and rhymes. The length of the foot is fixed; therefore a metrical syllable may occasionally be split into two text syllables without affecting the length of the line. The four-line metres are a combination of two couplets with four stressed syllables in the first line of each, and two such syllables (first and third, second and third, or third and fourth) alliterate with the first stressed syllable of the second line. The following couplet can serve as an illustration:

Hlægi snauir t ára t afli
t eflist fléttist gróurinn.

Gunnar Alexandersson made this; he is a hagyrdingur (skilled versifier) like many other kvædamenn. It is possible to reduce the countless rímur metres to variants of relatively few according to the number of lines and feet. The basic metres are simple schemes, but their variants and subvariants created by different combinations of rhymes can be difficult and complex (d´yr ) to a varying degree. Gunnar's couplet above is d´yr to a high degree, for if the first letter of each word is dropped the result is a meaningful second couplet, and the stanza is a variant of the metre langhent. (In rímur metres a vowel may alliterate with any other vowel, as in the second couplet.):

Hlægi snauir tára tafli
teflist fléttist gróurinn,
lægi nauir ára afli
eflist léttist róurinn.

When the last syllable of the first line of each couplet (the definite article -i ) is dropped, the metre becomes gagaraljód :

Hlægi snauir tára tafl
teflist fléttist gróurinn,
lægi nauir ára afl
eflist léttist rourinn.

And when the last syllable of the second line (the definite article -inn ) is also dropped, a variant of the most common metre, ferskeytt , results:

Hlægi snauir tára tafl
teflist fléttist gróur,
lægi nauir ára afl
eflist léttist róur.

This variant is considered the most difficult one, for the stanza may be read backwards as well as forwards:

Róur léttist eflist afl
ára nauir lægi,
gróur fléttist teflist tafl
tára snauir hlægi.

The term bragur (pl. bragir ) is used to mean verse metre, but in Vestfirir and Breiafjörur it has additionally the same meaning as stemma or kvæa- melody. The close relationship between verse metre and stemma suggested by this term is certainly evident in kvæaskapur, but the relation is rather distinctive.When Gunnar performs the variations of his couplet, he does not use the same stemma for langhent as he uses for gagaraljód and ferskeytt, even though the difference between langhent and gagaraljó does not seem to be any greater than between gagaraljó and ferskeytt .
Each stemma can only be used with specific basic metres and their variants, not with others. Thus all the metres performed by the informants of Breiafjör and Vestfirir fall into two groups (I and II) with the exception of n´yhenda (III) introduced by the rímur poet Sigurur Breifjör (1798-1846) early in the last century. The division into basic metres is tabulated in the foldout at the end of the book.1
There are certain stemmur which Gunnar finds difficult to kvæa, and he calls them söngstemmur because they must really be sung. Without doubt he is primarily referring to tunes based on triadic harmonies borrowed from foreign songs and adapted to rímur metres. Such tunes are not prominent in Breiafjörur (either in the old or new style) even if several are known. The stemmur most used are entirely different and cannot be simply classified as variants of a tune or tune family. Instead, most of them can be roughly represented in terms of one or more of the melodic outlines or contours shown in Figure 1. Their compass is not fixed, and therefore the levels -1 to 2 are relative, but 0 at the end of a stanza is an exception, for it represents the fixed pitch of the seimur, stable for each performance.
In the following discussion the narrowest ranges will be referred to as "narrow," and the widest as "wide," and the intermediate without any specification. The contours are either short, corresponding to a couplet, or long, corresponding to a stanza. A short contour is either repeated or combined with another one in order to extend through a whole stanza.2 A performer who performs metres of all three groups uses at least three stemmur, one for each group. Karl Gumundsson certainly knows more than three basic stemmur (Nos. 29, 30, and, 31b), as two more have been recorded, but they appear only exceptionally. On the other hand Pétur Ólafsson, not only knows but constantly uses seven stemmur with group I and four with group II. Those of the others who have been recorded sufficiently know several stemmur with I, the most frequently performed group, but use mainly one, and with II the tendency to use only one stemmais still stronger. Since III is performed infrequently, the repertoire is more difficult to define, but Pétur, and Thórur, the most recorded performers, have only one stemma each, and of the others only Kristján uses two.
Stemmur based on contours A, B, C, D, and E (Fig.1) are used generally with group I and F and G are only found with group II. A few III stemmur are based on I contours but most can be classified as fixed diatonic tunes.

Fig. 1

When the II metres are performed, the two lines within each couplet are never separated by lengthening of syllables except the second and third line of IIa, but the last foot of the second line is invariably long; this a stanza is performed as if it consisted of two long lines. The connection of lines and couplets of the I metres are much more varied. Ia and Ib stanzas are often performed without any break (Nos. 10 and 16 for example), but a lengthening of the last syllable in the third line or in the first and the third is even more common (Nos. 1, 13, and 20). In this way there is no break between the couplets, and frequently when the contour is long, no melodic division occurs either (Nos. 22a1, 23b, and 28b). This practice of separating the first line from the second and the third from the last and joining the second and third together runs completely contrary to the metrical structure and meaning of the stanza, as each couplet tends to be complete to a considerable degree. "Natural" performances of these verse metres with the last foot of the first couplet lengthened are rare. A few performers make a short pause after the second line no less than after the first and third, but perhaps this is only a breathing pause, especially if the tempo is slow. Ic has seven syllables in every line and is frequently performed without a break (Nos. 24c and 26), but the lengthening of the seventh syllable of one or more lines is just as common (Nos. 11 and 25). The last foot of the long line in Id is lengthened in relatively few instances or followed by a pause (Nos. 18 and 24b). Usually the first two lines are joined together and separated from the third, since the second line ends with a lengthened syllable in most cases (Nos. 3, 10b, and 20). The regular manner of adapting a I stemma to Id is to drop the second measure (compare No. 10a and b). But in some instances the only consideration seems to be not to alter the contour. In Nos. 7 and 18, for instance, the third measure is dropped rather than the second, and in No. 17b the first measure is contacted as a whole, the contour remaining essentially the same.
When a II stemma based on an F contour is adapted to any II metre, it is usually compressed or lengthened without altering the contour, except in the case of IIb, where the part corresponding to the third verse line (the second in IIa) is dropped. No. 39d shows how Kristján adapts his G contour stemma to the metres IIa, b, f, and g respectively. In IIa the second melodic line is an interpolation that is dropped in IIb, and then the second half of the contour is the same as in the remaining metres except that its length is reduced to half. Ending the first half of this contour a step higher in IIf (No. 39c) than in the other II metres he performs is not a personal stylistic trait, as Pétur Ólafsson ends the first half of the same contour in this way with IIe and IIf. So probably this was a common practice with the II metres ending in a monosyllable. Jón Ásmundsson has a rather unusual II stemma based at least partly on contour A (No. 38), and he treats it in a manner almost completely opposite to Kristján's G contour. Similar varieties of the contour are used with IIa and IIg (No. 38a and c), both halves ending on level 0, but in IIc (No. 38b) the first half ends as in IIc on level 1, so no distinction is made between ending on a monosyllable or a disyllables. Since the second half of IIa is identical to either half of IIg with regard to number of lines and syllables, this could be the reason why Jón treats them similarly.
On the whole the adaptions of contours to rímur metres are not necessarily the simplest possible, and the basic division into groups I and II is certainly not the most obvious one. For instance, there is only a small step from the four-line I metres to IIf, as may be seen from Gunnar's couplet, but only Gumundur Ólafsson uses his I stemma with it. The two three-line metres could also be expected to fall into the same group, and Thórur seems to see them as related, since he uses his II stemma with both, but as he is the only exception in the west perhaps the similarity is superficial. The first line of IIa and of IIb is of the same metrical and alliterative construction as a IIc couplet; therefore the long line could be regarded as two short lines, and IIa has actually been printed in that way in old editions of rímur.3 Id, on the other hand, is a genuine three-line metre, as the long line has only two alliterations (three are needed for two lines), and it is tempting to see it as a Ia couplet with the second half of the first line left out, especially since the usual procedure of adapting I stemmur to Id is to drop the second measure.
In the most general relations between the verse metres I and II as written and performed may be presented in a table:

VERSE METRES	Ia-c	Id-e	IIa	IIb	IId-g
As Written	four-line	three-line	three(four)-line	two(three)-line	four-line
As Performed	(3+1/1+2+1/ 2+2/1+1+1+1)	(2+1/1+1+1)	(1+1+1)	(1+1)	(1+1)
	short contours predominate		long contours predominant

I performances are characterized by short lines, variously connected, and short contours; whereas long lines, uniformly connected, and long contours are characteristic of II performances. As mentioned earlier, short contours are related or else combined with each other in various ways to cover a whole stanza. Different combinations sometimes occur in the same performance, and such performances are not unusual in the kvædaskapur of a few performers.
Thórur's I performances are the most varied in this respect.4 The short contour combinations A'A(' ), A''A, BA, CwCw (A(' ) means A' or A and Cw wide C ), and the long contours D and E (the third line usually ending on 0) are used with most of the 370 stanzas investigated, but not equally since the number of A'A(') stanzas is 180, almost the same as all the other together. Twenty 1a stanzas from Rímur af Thóri Hreu (No. 23a is stanza 5) are good illustrations of how Thóur uses several contours in a single performance. It consists of 1A'A, 1A''A, 1Eb-, 1BA, 1E-, 1Eb-, 1BA, 1A1', 1A'A, 2BA, 3A1A, 2A'A' , 3A' A and 1A'A1' stanza. Eb has a different second line from E such that the first half is B ending on level 1 (like the first half of No. 16). E is really a family of contours comparable to the family of A contours, and it could be regarded as consisting of two short contours in that the first half is sometimes exchanged for a different contour. Jónas Finnbogason seems even to use the second half of his E contour with the first couplet of a Id stanza, dropping the second measure (No. 24b3). Of course it could also be regarded as a new contour, and then the new contour might be used as well as E and wide D (No. 24a2) in the same performance. But it is difficult to draw definite conclusions from the limited material (20 I stanzas, 2 with different stemma ), and in the longest performance (5 Ic stanzas) he uses only E with the third line ending on level 1 (No. 24c). Thorgils Thorgilsson begins three I performances with a narrow variety of A'A' (No. 6a,b1) and then continues with CC , his chief contour (No. 6b4); in the first performance (9 stanzas) he uses the combination BC with the last four stanzas. Two performances (one of them 17 stanzas) use BC throughout, and six CC. Kristín Pétursdóttir uses A'A' with the first stanza and two varieties of E with the remaining three stanzas of a Ic text; in a later performance of the same text she uses a third variety of E with the first stanza. Besides this text she performs only two I stanzas with A'A1' (No. 10a) and E . Gumundur Ólafsson in his performance of the second ríma from Rímur af Svoldarbardaga (85 stanzas) uses narrow A'; and A'' in this way: 1A1''A', 1A'A', 1A''A', 13A'A' , 1A''A', 24 A'A' , 1A''A' , 3 A'A' , 1(A''A') , 6 A'A', 1 A''A' , 23 A'A', 1 A''A', 8 A'A'. The A´A´ stanzas do not change as much in detail as the A´´A´ stanzas, and stanza 46 (in parenthesis above) has almost a new melody compared, for instance, with the first stanza (see No.5). Gumundur has also performed another I ríma with the same contours in a similar manner, but a stanza resembling No. 46 does not recur. Most I performances of Sigurur Kristjánsson contain more than one variety of the A contours, and sometimes a narrow contour is combined with a wider one as in No. 9a. In many cases the varieties of A contours cannot be indicated by the three symbols for them, as for instance in Thórur's performance of "Breifiringavísur" (16 Ic stanzas, the seventh repeated after an interruption) where every stanza, except 1, 2, 14 and 16, is A'A' ´in spite of considerable differences (No.11, first seven stanzas).
If the parts of the contour corresponding to a verse line, "line-contours," are considered, the couplet-contour A is obviously a repeated line-contour. The symbol "a" will be used for this, and then the A contours may be rewritten:

where an accent means level 1, no accent 0, and an accent in parenthesis 1 or 0; placed before the letter the accent refers to the beginning, and after the letter to the end of the line-contour. Applied to "Breifiringavísur" the first lines of the first five stanzas are 'a' , those of 6 and 7: a' , 8: 'a' , 9-11: a' , 12: 'a' , and 13-16: a' . But this analysis is unsatisfying, since it is difficult to see why the first lines of 3 and 4 should be classified with the first lines of 1, 2, 5, 8, and 12 rather than those of 6, 7, 9-11, and 13-16, as movement between levels occurs in both instances. Considering the performance as a whole, the first half-line or measure is variable while the second is relatively stable. Written in terms of measure-contours, where å corresponds to a , and accents are used as before, the second measure is 'å' while the first measure of 1, 2, 5, 8, and 12 is 'å' ; 3 and 4: 'å ; and 6, 7, 9-11, and 13-16: å' . Since the second line is always a (except perhaps in 7) the contour of the first couplet varies primarily in the first measure. The same is not true of the second couplet which is a'a and in 2, 14, and 16 where the sixth measure has a special contour and the seventh is 'å ; here the second measure of third and the first measure of the last line vary together. The sixth and seventh measures of 2, 14, and 16 could be interpreted as a diminution of the B contour, and the sixth measure regarded å rising from level 1. Jónas Finnbogason's I stemma (No. 24) analyzed in this way would consist of various å contours rising from level 1 follow 24b1 and the repeated line of No. 24c this combination resembles a diminution of B . The only exceptions are the contours of the second line of No. 24a2 and the second measure of No. 24b2, which looks like a diminution of the former and when completed by the third measure appears similar to Thórur's third line in "Breifiringavísur" 2, 14, and 16.
Interpreted in terms of A , a , and of course å contours, B-E are related in manifold ways. The first line of B may be regarded as a from level 1, and C as a special variety of A where level 1 is not reached until the end of the first line. The second and third line of D together form A'' in most cases; the first line is 'a and the last (')a . Thórur usually divides the first and the third lines into two 'å contours (No. 21a1) similar to those at the beginning of stanzas 3 and 4 of No. 11. No. 21, from Vestfirir, is limited to a repeated 'å , with the exception of the fourth and the last measure. The fourth and the fifth measure combine into 'a , and the stanza-contour somewhat resembles Thórur's D contour. The Id stanza No. 20 also has a contour which is possibly a variety of D (compare No. 19) but with the second measure left out. At the same time the contour of the second couplet could be interpreted as B forming part of D, and the contour of the whole is close to some varieties of E as well. Since the relationship between F, the II contour most used, and C and A is obvious, the entirety of kvæaskapur in Breiafjörur seems to be dominated by A, a, and å contours. In keeping with this view, the first syllable of the second line in stanza 7 of No. 11 could be seen as a prolongation of a' into the second line, but it is also possible to regard the contour of the third measure as an inversion of å a new variety. It is noteworthy that a narrow version of the contour appears on level 1 in the second and sixth measure of the stanza, but nowhere in preceding stanzas. Inverted å characterizes the A' contour of Kristín Pétursdóttir (No. 10) and the narrow A' contour of Gumundur (No.5), and it appears occasionally in the performances of other informants. The last line of Jónina Ólafsdóttir's Id stanza No. 17b is similar to Thórur's second line in the seventh stanza of No. 11 except that it is narrower. If the whole stanza is compared with No. 17a1 it is easily seen that the long line of Id is a diminution of the first three lines, each measure corresponding to a line. The second line of Id, on the other hand, is an almost exact augmentation of the seventh measure of stanza 1, which has the same contour as the fifth measure of 2 and, on a higher level, as the first measure of 1. The contour of the third measure of 1 and 2 and the seventh of 2 is related to it and to the inverted å in the last line of the Id stanza at the same time.
At this stage the ambiguity of classifying by contours should be clear. The contours do not exist in reality as strictly limited units, mechanically manipulated, but rather as types of melodic motion with fluid boundaries both in time with regard to lines, feet, and measures, and in space with regard to pitch areas or levels. Thórur has stated on several occasions that the bragur (stemma) should not be changed during a performance. His I contours discussed above therefore seem to belong to the same stemma. On the other hand, different stemmur can be based on contours of the same class, and their contours can even be identical.
Thus Karl does not think he learned No. 31b from anyone -- it is his own stemma, but wife Ingibjörg says it is the same as No. 31a, and that it probably stems from her father.
Evidently the concept of stemma or bragur depends to a great extent on other factors than contour and metres.

BRAGIR: Footnotes

1 The basic metres are analytical constructs suggested as a basis for classification by the performance practice. A completely different classification into 23 families of verse metres, apparently not taking the performance practice into account, is proposed by the Rev. Helgi Sigursson (1891) in his extensive treatise on rímur metres.

2 The musical examples are presented in approximately the same order as the contours in Fig. I.

3 Thórlófsson (1950), 180.

4 Nielsen (1972) has described Thóur's performances in metres Ia, Ib, and Ic in terms of eight models (a model corresponds to a stanza) and thirteen melodic units he calls intonations (an intonation corresponds to a verse line) which are combined within the models. This classification relates only partly to the one presented here, as it does not take combinations of parts corresponding to half a stanza (i.e., short contours) into account. His models also include much more detail than the contours, as individual pitches are roughly indicated.