The foundation of poetry, that is of the expression of human feelings in verse, is rhythm. According to Aristoxenus, the greatest metrical scholar of Greece, rhythm is the division of time into equally recurring shorter and longer parts; but it becomes perceptible only by being applied to certain movements performed in this time (τὸ ῥυθμιζόμενον). This object of rhythm is, of course, different in the different arts. In music, it is the notes of a melody (μέλος); in dance, the movements of hands and feet (σωματικὴ κίνησις); and in poetry, the words (λέξις). In ancient, and especially in Greek, poetry, rhythm was the first and chief principle of verse; and Greek poets observed this principle with the greatest consistency, measuring their lines exclusively according to the length or shortness of the syllables, without taking any notice of the rhetorical accent of each word, which depends, of course, greatly on its meaning. This metrical system, founded entirely on the beauty of sound, agrees perfectly with what we know of the plastic art of the Greeks, where also the graceful and harmonious form predominates over the emotional expression of the features. There is something analogous to be found in the primitive poetry of the Teutonic nations. The aim of their poets was to impress the audience by the strong and heroic sound of their verses; and in consequence the principle of their metrical system was purely rhythmical. In Beowulf, as well as in the Hildebrandliet, or the Wessobrunner Gebet, each line contains a certain number of long and highly-accented (hochbetont) syllables, which are further emphasised by alliteration. The rhetorical importance of these syllables does not in the least influence their metrical value.

Latin poetry was not at first equally strict. The earlier Roman poets always tried to make the rhetorical and the metrical accent coincide. This was the more easy for them, as their rules of quantity were not yet clearly defined. Only the later Roman poets, and among them especially Horace, who were under the influence of Greek literature, introduced the accurate rules of Hellenic prosody into their own language, and at the same time made the metrical accent quite independent of the rhetorical. A remarkable sign of the difference between the Roman and Greek metrical systems is the way in which the two nations used the most important terms of rhythmical art, arsis and thesis. Aristoxenus, founding his metrical system entirely on the rhythms of dance and music, called arsis the weak part of the metre, because there the dancer raised his foot (αἴρω), and thesis the strong part, when the dancer trod the ground (τίθημι)—exactly contrary to the modern use of these words made familiar by Bentley. The best Roman metrical scholars, such as Atilius Fortunatianus and Terentianus Maurus, on the other hand, led by the rhetorical accent of their language, called arsis the first, and thesis the second, part of the metre, whether weak or strong, following, however, in this the metrical ἐγχειρίδιον of an unknown late-Greek author.⁠[32] The only exception is Martianus Capella, the author of ‘De Nuptiis Philologiæ et Mercurii,’ a work considered, during the middle ages, as a standard authority for all the branches of human knowledge. In his translation of Aristides Quintilianus, he adopted from him the use of arsis and thesis, although it was in direct contradiction to his own definition: ‘Arsis est elevatio, thesis depositio vocis ac remissio.’ The introduction of Greek prosody into the Latin language was simply a matter of art; and its reign could last only so long as the great poets of the classic period kept down the influence of popular poetry. As soon as the unlimited sway of these grand traditions ceased, the original tendencies of the Roman language began to oppose the Greek-Augustan orthodoxy; and this struggle, which lasted for many centuries, ended in the complete overthrow of the ancient prosody. It would lead too far to follow the traces of this process through its different phases; it is enough to say that, at the beginning of the middle ages, the rhetorical as against the metrical accent had more than reconquered its original rights in Latin poetry. In the grand religious songs of mediæval monkish poetry, such as ‘Dies iræ, dies illa’ or ‘Stabat mater dolorosa,’ the verses are measured entirely according to the modern principle of rhetorical accent. Even where the mediæval poets tried to keep up the appearance of ancient versification they could not abstain from yielding to the powerful influence of rising mediæval art. The best example of this fact is the favourite metre of monkish scholars, the Leonine hexameter. The poems written in this metre—as may be seen by the following two lines from the poem ‘De contemptu mundi,’ of the eleventh century,

utterly neglect the fundamental rules of ancient prosody. The same might be said, even in a higher degree, of Godfrid of Viterbo. He goes so far as to join two leonine hexameters and one pentameter in a stanza; for example:

The principle of dividing the stanza into three parts which is the basis of Italian and German strophes, can be easily recognised here; and the mediæval poet might have written his sham hexameters much more properly in this way:—

One of the most striking features of this rising poetry is the rhyme—an element quite independent of the metrical principle, and founded entirely on the sound and rhetorical accent of the words. This rhyme is used no longer as an occasional effect, in different places of the verse, but defined by the strictest rules of art. It has been a favourite subject of investigation with literary scholars, to determine who first used the rhyme. Monkish mediæval poets and Provençal troubadours have found enthusiastic defenders of their claims to this great invention. It appears however that the question itself was a mistake. Nobody invented the rhyme: it has existed as long as poetry itself. Horace and Homer knew it as well as Byron and Goethe; but the rhythmical principle prevailed too largely in the Latin and Greek languages to allow the rhyme, as a rhetorical element, to attain that influence which it gained by a natural process, when verses began to be measured according to the modern principle of rhetorical accent. Wilhelm Grimm, in his monograph ‘Zur Geschichte des Reims,’ has collected with great care the numerous instances of rhyme in the classic Roman period. The rule is, as Grimm shows, that the chief cæsura in the third foot of the hexameter rhymes with the end of the verse; but in other places also the rhyming words may be found. Grimm, however, decidedly goes too far when he sees an intentional rhyme in all these cases. The Latin language, owing to its consonant final syllables in declensions and conjugations, possessed an immense quantity of rhyming material, and moreover each adjective had to agree with its noun, if it followed the same declension. It is therefore difficult to see how the poet could have avoided bringing into the same verse very often two or even more words ending in the same way. In a verse, for instance, like that quoted by Grimm from Virgil’s ‘Bucolics,’

no Roman poet could have intended, nor a Roman ear have noticed, a rhyme between ‘tristia’ and ‘bella;’ especially as the different metrical value of the two syllables modified the sound of the two a’s. Grimm seems not to have been able to free himself altogether from the propensity of biographers to overrate the importance of their heroes. However, in innumerable other cases rhyme has decidedly been used of set purpose by the Roman poets, especially where the corresponding words are found either in the chief cæsura and the end of the same verse, or at the end of two verses following each other. Of both cases an example may be cited from Horace, whose fine ear and ability to avail himself of beauties of rhythm and sound make him an important witness for the intentional use of rhyme.

affords an excellent instance of rhyme in the chief cæsura; while the lines

prove even the existence of a sort of feminine rhyme in Latin poetry. Horace also shows how the Roman poets used the rhyme for onomatopoetic purposes. In the celebrated line,

the quick jump of the little animal could not have been better illustrated than by the rhyme of the two words immediately following each other—‘Schlagreim,’ as the German meistersingers appropriately call it. Though there can be no doubt that the Latin poets of the classic period knew and occasionally used rhyme, it never was to them of the same vital importance that it is to modern poets; and in fact it never could be, so long as the rhythmical accent preserved its unlimited power; for this is decidedly unfavourable to rhyme. In all those cases, for instance, where the chief cæsura of the hexameter rhymed with the end of the same verse, which, as we have seen, was the usual way, the two corresponding syllables had different metrical accents. In the line already cited,

the o of duro stands in the arsis, and therefore has quite a different sound from the o in aratro, which stands in the thesis. This becomes the more evident in those very rare cases where the rhyme in this position contains two syllables, or is, as we should say, feminine. In Horace there is only one instance of this; and indeed what could be the use of a rhyme which, if the verse were read according to rhythmical principles, would be scarcely audible?—

But by the same fact the destructive influence of rhyme on the rhythmical principle becomes evident. The line, for instance, already quoted from Godfrey of Viterbo, would, if properly scanned, have sounded thus:

But this way of destroying the feminine rhyme by the rhythmical accent certainly did not tally with the feeling of the mediæval poet; and it may be assumed that he accentuated fídus exactly like Gotefrídus, as if it were a trochee. This at the same time agreed perfectly with the rhetorical accent of the word. Reading the whole verse according to the same principle, the first part of it,

became quite trochaic in character, and the idea of the hexameter is utterly destroyed. This destruction of the rhythmical principle in mediæval Latin poetry was almost contemporary with the same phenomenon in Teutonic literature. Here also the dominion of purely rhythmical measurement and alliteration was victoriously contested by rhyme and rhetorical accent. At the beginning of the middle-high-German period, alliteration as a principle of art disappeared; and by the great minnesingers of the twelfth and thirteenth centuries it was used only occasionally, and without any strict rule, just as rhyme was by the old Roman poets. It was chiefly preserved in old alliterative formulæ, such as ‘Haus und Hof,’ or our ‘Kith and Kin.’ A beautiful example of this is found in the last stanza but one of the ‘Nibelunge Nôt’:

Rhythmical accent, however, existed side by side with rhetorical accent much longer in German than in any of the Romance languages; and traces of its influence may be found almost till the beginning of the modern high-German epoch.

In the Romance languages the decline of the rhythmical principle was even more complete than in the mediæval Latin or in any of the Teutonic idioms. The feeling for rhythm in those languages was so entirely lost that they were not able even to preserve the rhetorical accent in sufficient strength to make it of any avail for metrical purposes. Although in most of the poems written in the Romance languages there is a certain resemblance to the iambic or trochaic fall, yet the scanning of a whole stanza according to these metres would in most cases prove impossible. In modern French, which has gone farthest in neglecting the rhythmical difference between the syllables within the same word, there is scarcely a single line of the most finished poets which could be read metrically without altering even that remnant of rhetorical accent which has been preserved. In the following verse, taken from Boileau’s sixth Satire,

there are two striking examples of this fact; for the accent of the (if anything) iambic metre in the word peine is on a syllable which in prose is scarcely pronounced at all, and in commençant the last syllable is at least as long as the last but one. Where modern French poets try to introduce something resembling rhythm, they generally do so less by means of the rhetorical accent in words of several syllables than by putting the more or less important parts of the sentence, such as article and noun or personal pronoun and verb, in thesis and arsis respectively. In the main it may fairly be said that in Romance poetry metre is entirely founded on counting the syllables of the verse, and rhythm, properly speaking, has disappeared, except so far as it shows its influence in the combination of verses of different lengths in a stanza.

This leads us to another consideration, which is of the highest importance in studying Provençal versification. Rhythm showed its influence on the poetry of the troubadours, not only in the single verses, but also in the composition of several verses of different sizes and cadences into an organic whole—the strophe. The harmonious beauty and impulsive lyrical pathos of Pindar’s odes excite the same admiration as does the steady epical flow of Homer’s hexameters; and to the inheritance of the strophe, and its development into the stanza, mediæval poems, and especially the canzos of the troubadours, owe their greatest charm. To the relics of ancient literature already mentioned was added the rhyme, defined by strict rules and made obligatory; and this new principle contributed not a little to give variety and harmonious beauty to the mediæval stanza. In investigating Provençal versification, it will therefore be necessary to consider (1) rhythm, as shown in the manifold measures of verse, (2) rhyme, and (3) the mode in which by these two elements combined the stanza of the troubadours was formed.