Ingjaldi, Hárfagra, Búseyru, Sigvalda (2), Danmarkar, Kolfinnu, Svíþjóðu, Aleifi (2), Klifsandi, Englandi, Rögnvaldi, Ögmundar, Ástríði (all in second lines). In these cases the first syllable taken by itself is of all possible types (1) Dan-, Kol-, Sig-, (2) Vé-, Sví, (3) Gunn-, Rögn-, (4) Stein-, Hár-. But let us look at the remaining examples: 18 gladar pengill herdrengi (Second line) 28 armdags en litfagra دو 31 sverð fylgir menmyrði (First line). These four instances are significant enough: the first element of the compound has in every case a short vowel, and the type of line is exactly the same as it would be if we had her drengi, hug rekki etc. But four examples are scarcely sufficient to establish a point of so much importance as this will prove itself to be: let us therefore take the examples quoted by Prof. Sievers in Beiträge VIII. 75. The united evidence of these lines can hardly be called in question: strict metre evidently required that the first element of a compound in this position should be short, i. e. the line ought to end with. From this absolute rule, however, proper names were excepted 1), and we have seen above that this type of line was mainly employed when it was necessary to use a proper name. The reason for avoiding it as a rule in other cases probably was that there was a kind 1) The frequency of such exceptions may explain the rare occurrence of lines like Þórmód's orms torg í skjaldborgu (Fóstbr. s. p. 86); compare the exceptions in the next section (á haussprengi Hrungnis, etc.) of break before the last foot, so that a compound in this position was felt to be awkwardly placed. § 18. The fact elicited in the foregoing section, that such compounds as hugrekki, skýranni were under certain circumstances stressed on the second element raises the question of their value in other positions. A little investigation shows that the formula holds for one other place in the line, viz. when the first element immediately follows the first stress, as in jarl goðvörðu hjarli. Here again the distinction. between first and second lines is so important that it will be best to exhibit characteristic specimens side by side. First lines pá's dynfúsir dísar Second lines sundr gráklædi pundar blid lofskreyti at riđa bands jódraugar landa gifrs hlémána drífu Ilmr dagleggjar hilmis stafns flatbalkar Hrafna Taking the second lines first, we find that the normal type occurs 29 times in Skjaldekvad, and the compound has invariably a syllable with short or final vowel as its first element, e. g. brimróta, geðbæti, hámána, berdraugar, flatvallar, valkastar, strádauða, etc. A remarkable fact is that the negative prefix ó- is among these short syllables, e. g. seid ófagran vigra minn ókunnar þínum. To interpret these lines as belonging to Prof. Sievers' D-type, XX plainly does not account for the facts, because in that case there would have been no necessity for the skald to select a compound of this form. The only value which can reasonably be assigned to them is, so that the first element of the compound is weaker than if it were an independent word, for Ilmr dag leggjar would be and impossible except in skjálfhend. § 19. In the case of second lines of the above form the first syllable regularly contains one of the hendings (see the examples above), the only exception in Skjaldekvad being Egil's seið ófagran vigra. If we now turn to the lines. in which the first syllable has not full stress, the position of the first hending becomes a matter of primary importance for the metrical interpretation of these lines. If Prof. Sievers were right in assigning them to his C-type (X ≤ | ≤ X), the hending might as well be in the second as in the third syllable. A few examples, taken straight forward from Skjaldekvad, will show how the case really stands. til mjödgjarnir fornir It is unnecessary to multiply instances: of all the lines of this type in Skjaldekvad only two form exceptions to the general rule. These are Einar's line, 18 nu's afrendra jöfra and Björn's, 31 þvíat armleggjar orma, but the latter is perhaps corrupt, as the MS. has dagleggjar (not dagbæjar) in the following line. Two facts must therefore be taken into account in deciding on the metrical value of these lines, (1) the shortness of the first element in the compound, and (2) the position of the hending. These clearly point to one result, and the scheme can hardly be anything but ||; here again the two secondary stresses make up a full one, and the line is only a variation of the normal -~ |-~| It is not a serious argument against this view that in a few cases the first element of the compound in actually long. In the following list I give every such line which occurs in Skjaldekvad, so that the full evidence may be before the reader. Now, even if we assume that these 6 lines were originally composed in this form, they cannot disprove the cumulative evidence of the other 37 (+ 29 second lines) in which the first element is short. Moreover, it will be no ticed that the hendings are still in every case placed in the second element, so that the line still presumably ends in The only alteration required in our scheme, therefore, is to grant the possibility of as a first foot, as well as2). § 20. We now come to the other type of line which contains a trisyllablic word covering the second, third, and fourth places, i. e. such lines as en slóðgoti síðan or snjallr morðhamar gjalla. The former of these Prof. Sievers takes as C 3 × × ×; the latter as D 2XX. This involves the assumption that the difference of stress at the beginning of the line alters the stressing of the compound, a view perhaps conceivable in itself, but, as we shall see, not at all necessary. 1) I omit, as of doubtful authenticity, the following lines: þvít lautsíkjar lékum (laut- is conjectural), slapp Hitdæla kappa (? read Hit-, in any case it is a proper name and so may form an exception), þvít armleggjar orma (see above), en fúrþollar fellu (fúr- is conjectural). 2) It may be noted here that such lines tend to disappear in later dróttkvætt, as the skalds became more and more averse to placing one of the studlar in the second place of the line. Thus Torfeinar's verses (32 lines) contain 6 specimens of↓↓ ~~, Bragi's Ragnarsdrápa has = 3 in 88 lines, Þjóđolf's Haustlöng has 1 in 160 lines, Hornklofi's Glym drápa 2 in 64 lines, Einar's Vellekla 6 in 188 (also 3 which begin with, viz. ok oddneytir úti, ok rauđmána reynir, und sigrrunni svinnum), Hallfred's Erfidråpa 2 in 212 (one in and one in ), Sigvat's Vikingarvisur 2 in 116 (his Bersöglis visur none in 136), Einar's Geisli only 1 in 568 lines! The word in question may or may not be a compound: of the 115 examples of this type in Skjaldekvad 70 contain compounds, and 45 show words of other formation. The latter I shall leave out of consideration for the present. As regards the former, our previous distinction between first and second lines naturally holds good, as may be seen in the following lists. First lines. ok borróins barđa Second lines. hugr malmgnáar brugtit farmr dólgsvölu barma The form of the second lines, it is evident, bears a striking resemblance to the types discussed in § 14 above; syllable for syllable, úlfs tannlituor glitra does not differ at all from lands frækn jöfurr granda or jarlmanns bani snarla. This being so, there seems to be no reason against assuming that they are metrically identical, and that the type we are now considering is also, exactly as Prof. Sievers has it. Taking now the first lines, we see at once that the position of the first stuðill is, and must be, invariable. This was also the case in the lines discussed in the previous section, in which we were led to assume that the line began with . The same view is obviously tenable here, and if we set down the metrical scheme as we arrive at the same distinction between first and second lines as we are already familiar with, while we avoid varying the stress in compounds of exactly the same form 1). Here, as in the previous section, the position of the hending ought to be of some value as an indication of how such compounds were stressed. Examination of the various 1) The first-line type becomes less common in later verse; compare the note to the preceding section. |