middle syllable or to the root-syllable of the second part of a compound: góde páncòde, wine Scýldinga múrnènde mód —

sá-nassas

féond máncỳnnes, side wéorð-mỳndum þáh, wórold-àre forgeaf etc., so too in verses like wis-fast wordum, fólc-stède frætwan, Gréndles gúð-cræft, gúð-rìnc gold-wlànc, and in those verses mentioned above, which contain three independent words, such as blæd wide spràng, flóta stille bad, wlánc Wédera leod, grette Géata leod, átol ýđa geswing etc. By thus scanning, however, the strict two-beat theory breaks down and is recognized as inadmissible, for a nebenhebung' or 'weaker hebung' counts in the whole Germanic prosody as a 'hebung' and never as a 'senkung'. The verses above contain, therefore, as Sievers himself admits, three and sometimes four hebungen or beats.

thesi

§ 25. Criticism of the Two-Beat Theory

(conclusion).

Another great objection to the two-beat theory results from the observation that in the later rimed verse, e.g. of Otfrid and, as we shall see later, also of Lagamon (§ 107 ff.) the verses are always masculine, that is they end with a stressed syllable. When words like múatèr, drúhtìn, hímilè, wóroltì are at the end of a verse, they have in addition to the chief stress (haupthebung) on the root-syllable also a secondary stress (nebenhebung) on the inflexional or derivative syllable. This weaker stress

counts as a full beat for the verse. In view of this it would be strange if, according to the twobeat theory, in the far older alliterative verse inflexional and derivative syllables of disyllabic words with a long root-syllable, or of trisyllabic words with a short root-syllable, such as hwile, wolcnum, frætwan, modor, dryhten heofone, warode, fremede etc. were mere 'senkung-syllables' and of no further importance for the rhythm of the verse. In the same way in Otfrid's verse words with three syllables, of which the root-syllable and middle syllable are long, such as éwärtò, drúrềntì, ántwürtì etc. have three beats at the end of the verse, whilst, according to the two-beat theory, either only one beat or at most two beats are allowed the corresponding OE. words yldesta, murnende, lidende, mancynnes etc. (see above p. 29).

The inflexional and derivative syllables of such words cannot have become capable of bearing a beat at a later period, after having become weaker, if at an earlier period they could not be used as full beats of the verse. If, therefore, Otfrid and Lagamon could use these inflexional and derivative syllables as full beats, then they must be assumed as full beats for the much earlier alliterative verse. We must, therefore, scan: hwilè, wólcnùm, frætàn, mnigò, wdrođè — ýldestà, Scýldingà, múrnendè, máncŷnnès etc. If, however, we are compelled to exceed the measure of two beats, then only double that number, four, can be con

sidered; for the number three is quite unknown in the oldest Germanic prosody, it contradicts the origin of the rhythm of verse from the rhythm of marching and dancing.

NOTE. In my Studien zum germanischen Alliterationsvers I, § 12 I have shown that the conclusion of the OE. poem Phoenix, which contains English in the first halfverses and Latin in the second half-verses, provides no proof of the correctness of the two-beat theory. Schipper's reference (Grundriss, 9 f. 80) to the judgment passed on alliterative verse by King James I and Bishop Percy is of no account, for Englishmen in the XVI-XVIII centuries could have had no correct conception of the historical development of the older English language and prosody. This we see from Dryden's remarks on Chaucer's verse (§ 189).

§ 26. Insufficiency of the Two-Beat Theory.

After the work of Vetter and Rieger at the be ginning of the seventies the two-beat theory was generally adopted. A few scholars only, e.g. Oskar Schade in Königsberg, held fast to Lachmann's four-beat theory, unfortunately without properly exposing the weaknesses of the two-beat theory. But the difficulties in the way of a strict application of the two-beat theory, especially the arbitrariness in arranging the theses and the extent of the theses, made themselves felt by the supporters of the theory.

The two-beat theory could not provide a real scheme for scansion. It could fix the number of

beats only; everything else was left to the arbitrariness of the poet, who, as the supporters of the theory insisted, had an opportunity of putting his poetic feeling into practice by a wise restraint. If, however, the laws of OE. prosody, as they hovered before the poets, had really been so unfixed, then we should surely find a greater looseness in the use of alliterative verse in the poems handed down to us than is the case. The great freedom, which the two-beat theory of Vetter and Rieger gives the poets, cannot therefore have really existed.

$27. Sievers.

These considerations induced Sievers once more carefully to examine the OE. verse structure. He began with Beowulf. He says (PBBeitr. 10, 218) that he had always had "an uneasy feeling when reading [OE. verse] that a definite rhythmical something made itself heard, which required a definite statement", although he could not establish the laws of this "definite rhythmical something". Thus Sievers, who approached the subject from the two-beat theory, observed that the use of the so-called 'auftakt' (anacrusis), was not so arbitrary as had been assumed; but that in certain cases, viz. where no thesis followed the first or second arsis, the presence of the anacrusis was as necessary as its absence was usual when a thesis followed both the first and the second arsis, cp. e.g.

1

in | gear-dagum, on | bearm scipes or purh | mīne hand, him on | bearme læg with lange hwile, geong in geardum, weox under wolcnum etc. Sievers found, too, that the final thesis of a verse was always monosyllabic: lange hwile, gewāt him pā to warode, on bearm scipes etc., whilst the middle thesis and the anacrusis could consist of two or more syllables: wēox | under | wolcnum, sæg dest. from his side him on | bearme læg, peah he him | leof ware etc. And so he arrived at a new theory with regard to OE. alliterative verse, by which it is possible to regulate the construction of the half-lines according to definite laws.

Sievers first founded his theory on the verse of Beowulf and published his results in a long essay Zur Rhythmik des germanischen Alliterationsverses (PBBeitr. 10, 209-314 and 451-545). Later he developed his system further and extended it to the alliterative poetry of all Germanic peoples in his Altgermanische Metrik (Halle 1893) and in the section on Altgermanische Metrik in the first edition of Paul's Grundriss der germanischen Philologie II, 1, 861-897. Unfortunately Sievers has taken no notice of any criticism of his book nor of any works on old Germanic prosody, which have appeared since 1893. The revision of his article on old Germanic prosody in the second edition of Paul's Grundriss he has left to others (Kaufmann and Gehring); see above § 9.