In the forms of lyrical verse hitherto considered by us we were able to trace some popular germs, considerably modified and highly developed though they might appear. But any such connection ceases in the numberless variations of verse and stanza, in which the unrivalled workmanship of the troubadours loved to shine. That this ease of inventing ever new and ever more complicated combinations frequently led to excesses of artificiality need not surprise us. Our admiration of the marvellous ingenuity displayed by the poets is mingled with regret at seeing it wasted on trifles.

The number and variety of these efforts would defy all attempts at perfect classification and nomenclature. The troubadours altogether were sparing in the use of technical terms, but even the later grammarians found it impossible to affix names to all the metrical refinements and tours de force in which Provençal poets delighted. It is amusing to observe the different attitude in this respect of the poets and metrical theorists of Northern France. The rhyming capacities of their language were as inferior as their own craftsmanship to the language and the art of the troubadours. In consequence they found it desirable ‘to make a little go a long way,’ and, for example, dubbed with the sonorous name of ‘Chant Royal’ the mere repetition of the rhymes of a somewhat complicated stanza throughout a poem of moderate length: a feat performed almost unconsciously by the troubadours in numberless canzos and sirventeses. The ballade made celebrated, although by no means invented, by the genius of Villon, and which, by the way, differs as widely from the Provençal balada on the one hand, as it does from the Scotch ‘ballad’ on the other, is a similar contrivance of a still simpler nature. This simplicity, of course, by no means detracts from the poetic merit of these poems, and the manner, for instance, in which the refrain is used in both cases betrays considerable skill. But compared with the consummate workmanship of the troubadours, these efforts appear mere child’s play.⁠[14]

Of the elaborate rules of Provençal metrical science and practice, both as regards the rhyme and the construction of stanzas, full account will be given in the technical section. For the present it will suffice to name a few examples chosen for their quaintness and originality rather than for any extraordinary display of workmanship.

The most important amongst these is the sestina. It was invented by Arnaut Daniel, the master of ‘dear rhymes’ and ‘obscure words,’ of whom and of which previous mention has been made. For his propensities in that direction Arnaut himself tenders a very plausible excuse. He shifts the responsibility from his own shoulders to those of his lady. If she were kind to him, he alleges, melodious rhythms and pleasing simple verses would naturally flow from his pen. The lady’s cruelty therefore is answerable for involved sentences and harsh rhymes. The plea is not altogether without force. But Arnaut’s natural tendency towards the incomprehensible and strikingly original is at the same time established beyond a doubt. One of his favourite devices was to construct a stanza without a single rhyme in the stanza itself. But instead of this the close of the first line would match with that of all the other stanzas of the poem, the second line with the second, and so forth. In one poem, for instance, the last word of the opening line of the first stanza is larga, that of the second stanza embarga, that of the third descarga, and so on through all the corresponding lines of the poem. To modern and northern ears the consonance thus suspended for eight or more lines is hardly perceptible. But in the south and in the middle ages this was different. Even so great a master of form as Dante highly approved of Arnaut’s practice, and, what is more, avowedly imitated it (‘et nos eum secuti sumus.’—De Vulgari Eloquio, cap. 10). The result of this imitation is one of the sweetest love poems of the ‘Canzoniere,’ the sestina beginning, ‘Al poco giorno ed al gran cerchio d’ombra.’ Two other poems of the same kind are attributed to Dante by some commentators, although others doubt their authenticity.

The fundamental scheme of the sestina, as has already been stated, is that of blank-verse stanzas, being in the relation of rhyme to other blank-verse stanzas. But here this principle is carried to a climax. For not only the consonances, but the actual rhyme-words of the first strophe are repeated throughout the poem. The difficulty of writing sense and poetry under such conditions is increased by the curiously inverted order in which these words are repeated. To give the reader an idea of the ingenuity of this contrivance, it will be necessary to write down the concluding words of the six stanzas of a celebrated sestina by Arnaut Daniel in the order in which they occur. The number six—both as regards the stanzas of the poem and the lines of each stanza—is the orthodox one, and has given the name to the poem. A short tornada or envoi, however, is added, and in this the six rhyme-words of the previous stanza are once more repeated.

It will be observed that the second stanza repeats the rhyme-words of the first in this order, 6, 1, 5, 2, 4, 3, and exactly the same relation will be found to obtain between each stanza of the poem and its predecessor. Whether there was some hidden significance in this sequence it is impossible to tell. But one is inclined to suspect that it must have been some such attraction which induced the great Dante to imitate Arnaut Daniel’s device with perfect accuracy. Or was it the pseudo-scientific regularity of the scheme, so fascinating to the mediæval mind, which attracted him? Anyhow, the fact is undeniable that Dante’s poem in question, although infinitely superior, by its poetic beauty, to anything that Arnaut Daniel ever wrote, is, as regards its metrical scheme, an exact copy of the troubadour’s sestina. Only in two minor points has Dante dared to deviate from his model, in points, too, which do not materially interfere with the position of the rhyme-words. These are, the length of the opening lines of each stanza, which in Provençal are by one foot shorter than the other verses, while in Italian they are of equal size; and the arrangement of the rhyme-words of the tornada. But Dante’s licence in these details makes his strict adherence to the essential idea of the form all the more significant.

It is interesting to note that the preference for the sestina has not been confined to mediæval poets or Romance languages. Mr. Swinburne, to mention but one instance, has essayed the form with excellent results, both in French and English. But the model he has followed is not derived from the Provençal original, nor yet from the Italian copy, but from a modified French version of the scheme. This modification consists chiefly in the use of the rhyme within the single stanzas themselves, which is wholly at variance with the original meaning of the form. Banville suggests that the stanza in this altered condition has been imitated from Petrarch. But he is quite mistaken. Petrarch, although he ostentatiously avoided reading Dante’s works, has in this instance exactly followed Dante’s example. Besides, he was too well acquainted with the musical significance of the stanza in question, not to know that all the lines must be rims escars, or, according to Dante’s terminology, claves, that is, unmatched by rhymes in their own stanza. For a fuller account of these details I must again refer the reader to the technical portion of this book.

In connection with the sestina and its history in the lingua di sì, it may be mentioned that another important form of Dante’s, and generally of Italian, poetry, the sonnet, seems to have been of indigenous, not at least of Provençal, growth. The structure of a stanza of fourteen lines containing the well-known number and arrangement of rhymes, is in perfect accordance with the metrical principles practised by the troubadours, but the only specimen of the sonnet in the langue d’oc was written by an Italian poet, Dante da Maiano. It is by no means a master-piece, and remarkable chiefly for the fact that all the rhymes are of the male or monosyllabic order, an arrangement not unfrequent in Provençal, but unprecedented in Italian; the latter circumstance being perhaps the reason why an Italian poet, writing in Provençal, adopted it. Against the Provençal origin of the sonnet would also seem to speak the fact, that the word is used without any technical restriction, merely as equivalent for a song:

‘I make a sonnet evil or good, what about I don’t know myself,’ says Guiraut de Bornelh, wishing to illustrate the wayward mood of a distracted lover. Of the poem of sixteen lines he thought no more than did Burns when he described Tam O’Shanter as ‘crooning o’er some auld Scots sonnet.’

The exact antipodes of the sestina is the descort, Anglicè discord, or dissonance. In the former everything is fixed by rule—position of rhymes, number and length of lines and stanzas. In the latter absolute liberty prevails regarding all these points, or rather it is the ambition of the poet to create the most bewildering divergence. But sometimes even the most glaring contrasts of metre are found insufficient, and an additional discordance of idiom is resorted to. Rambaut de Vaqueiras has employed no less than five different languages or dialects to complain of the cruelty of his lady. For, like the harsh rhymes of Arnaut Daniel, all these dissonant contrivances were attributed to the feeling of unrequited love, and Guiraut de Salinhac, in a very pretty little poem, distinctly says, ‘I should not compose a Discord if I met with accord and accordance at the hands of her I love.’ The inventor of this curious device is said to have been one Guerin d’Apelier, a poet not otherwise known to us. His claim to immortality may appear somewhat slender under such circumstances.

Akin to the elaborate confusion of the descort and about on a par with it as regards artistic merit, is the sudden lapse from poetry into prose, for which Rambaut of Orange is more especially responsible. Of Rambaut and his disastrous love-affair with Beatrice de Die, the poetess, we shall hear more hereafter. As a poet he belongs to, and is indeed amongst the earliest representatives of, the artificial school which culminates in Arnaut Daniel. Rambaut is by no means without skill, and according to his own statement, ‘no poet s work from the time Adam ate the apple was worth a turnip compared with his.’ But his devices frequently take the form of mere eccentricities, and he never induces us, perhaps never intends us, to forget the amateurish quality of his work. The mixture of poetry and prose alluded to in the above remarks well illustrates the lawless tendency of the noble poet.

The explanatory nature of these prose interludes induces Raynouard to class Rambaut’s poem with the ‘pièces avec commentaire’ (‘Choix,’ vol. ii. p. 248). To add a kind of commentary to poetic work was a not uncommon custom in the middle ages. Dante’s ‘Vita Nuova’ is a prominent case in point. In Provence, where a whole school of poets took pride in writing as incomprehensibly as might be, some such assistance to the weaker brethren became all the more indispensable. In most cases no doubt the joglar supplied the want by adding, after the recital of a poem, such explanatory notes as might seem most adapted to the intellectual level of his audience. Of Guillem (not Peter, as Raynouard calls him) de la Tor, the joglar, and friend of Sordello, we are told in the manuscripts that ‘he knew a great many canzos and was clever and sang well. He also was a poet; but when he wanted to recite his canzos he made his commentary longer than the poem itself.’ From the expression used by the biographer, sermo de la razo, we are led to conclude that Guillem’s long-winded explanations were couched in prose. This, however, was not always the case. We know of troubadours who good-naturedly took the trouble to elucidate the darknesses of brother bards by means of poetic glosses. Guiraut Riquier, the scholar and poet, here again shines by his example. The nature of these commentaries is well illustrated by a stanza of one of his poems which the reader will find translated in Raynouard’s ‘Choix’ (ii. 252). Guiraut de Calanson, in one of his poems, speaking of the palace of love, says that four steps, or degrees, lead up to it. Guiraut Riquier explains that these steps are ‘honour,’ ‘discretion,’ ‘gentle service,’ and ‘good sufferance,’ much to the edification, no doubt, of mediæval readers, and especially of Count Henri de Rodez, who, under his hand and seal, testifies to Guiraut’s explanation being trustworthy and to the point. The five portals of the palace and the mode of opening each individually, also find a circumstantial explanation at Guiraut’s hands.

Of the breu-doble (double-short), again, Guiraut Riquier is the inventor, and, as far as I am aware, the sole representative. In the poem of the kind which we possess from his pen he complains of the cruelty of his lady, to which he, in imitation of other troubadours, ascribes his adoption of this new mode of utterance. ‘As she will not accept my canzos at their worth,’ he says, ‘I write this breu-doble.’ There is nothing very remarkable about this form, which, for that reason perhaps, has met with little approbation amongst the elaborate rhymsters of the later epoch. The name ‘breu-doble’ has been a puzzle to modern scholars. Raynouard is inclined to derive it from the shortness of the poem, which, however, would by no means account for the ‘doble.’ To me it seems more likely that allusion is made in a slip-shod way to the last verse of each stanza, which, although not exactly half the length of, is at least considerably shorter than, the remainder of the lines, from which it also differs by its rhyme.

Of greater importance than the breu-doble is the retroensa, also known chiefly through Guiraut Riquier’s agency. The only striking feature of this form is the refrain which, against the usage of Provençal poetry, consists of more than one line. An exceedingly pretty poem, called in the MS. ‘The First Retroensa of Guiraut Riquier made in the year 1276,’ is devoted to the praise of the Catalans, renowned in the middle ages as models of knightly courtesy. ‘As my star has decreed,’ the poet says, ‘that good should not come to me from my lady, as nothing I can do will please her, as I am too weak to tear myself from her, it is time that I should be grounded in the ways of true love; and of these I can learn enough in gay Catalonia amongst the brave Catalans and their sweet ladies.’ On these he proceeds to shower every imaginable compliment through a number of stanzas all bearing the harmonious burden:

Like the descort and many other metrical creations of the troubadours, the retroensa was known to the poets of northern France. The name at least occurs in the literature of the langue d’oïl; but it must be confessed that, for instance, the religious song in praise of the Virgin, expressly called by the poet a retrovange novelle, has neither in substance nor form anything in common with Guiraut Riquier’s poem. Even the refrain has disappeared. There may perhaps have been some musical reason to account for the adoption of the name. But on that point we are, alas! completely in the dark. It is unnecessary to enter into the numerous and for the greater part arbitrary distinctions in which the subtle minds of grammarians and metrical scholars were wont to delight. Most of the divisions thus created, such as the escondigz (justification), the comjatz (literally leave-taking, i.e. the song in which the allegiance to a cruel lady is renounced), or the torneys (tournament song), and many others never seem to have attained distinct formal development, and the remaining specimens are very few in number.