The development of romance in these authors is not always and in all respects a gain. Even the pathetic stories of the Decameron (such as the Pot of Basil, Tancred and Gismunda, William of Cabestaing) seem to have lost something by the adoption of a different kind of grammar, a more learned rhetoric, in comparison with the best of the simple French stories, like the Chastelaine de Vergi. This is the case in a still greater degree where Boccaccio has allowed himself a larger scale, as in his version of the old romance of Flores and Blanchefleur (Filocolo), while his Teseide might be taken as the first example in modern history of the pernicious effect of classical studies. The Teseide is the story of Palamon and Arcita. The original is lost, but it evidently was a French romance, probably not a long one; one of the favourite well-defined cases or problems of love, easily understood as soon as stated, presenting the rivalry of the two noble kinsmen for the love of the lady Emily. It might have been made into one of the stories of the Decameron, but Boccaccio had other designs for it. He wished to write a classical epic in twelve books, and not very fortunately chose this simple theme as the groundwork of his operations. The Teseide is the first of the solemn row of modern epics; "reverend and divine, abiding without motion, shall we say that they have being?" Everything is to be found in the Teseide that the best classical traditions require in epic—Olympian machinery, catalogues of armies, descriptions of works of art to compete with the Homeric and Virgilian shields, elaborate battles, and epic similes, and funeral games. Chaucer may have been at one time tempted by all this magnificence; his final version of the story, in the Knight's Tale, is a proof among other things of his critical tact. He must have recognised that the Teseide, with all its ambition and its brilliancy of details, was a failure as a story; that this particular theme, at any rate, was not well fitted to carry the epic weight. These personages of romance were not in training for the heavy classical panoply. So he reduced the story of Palamon and Arcita to something not very different from what must have been its original scale as a romance. His modifications of Boccaccio here are a lesson in the art of narrative which can hardly be overvalued by students of that mystery.

Chaucer's procedure in regard to his romantic subjects is often very difficult to understand. How firm and unwavering his critical meditations and calculations were may be seen by a comparison of the Knight's Tale with its Italian source. At other times and in other stories he appears to have worked on different principles, or without much critical study at all. The Knight's Tale is a complete and perfect version of a medieval romance, worked out with all the resources of Chaucer's literary study and reflexion; tested and considered and corrected in every possible way. The story of Constance (the Man of Law's Tale) is an earlier work in which almost everything is lacking that is found in the mere workmanship of the Knight's Tale; though not, of course, the humanity, the pathos, of Chaucer. The story of Constance appears to have been taken by Chaucer from one of the least artificial specimens of medieval romance, the kind of romance that worked up in a random sort of way the careless sequence of incidents in a popular traditional tale. Just as the tellers of the stories in Campbell's Highland Tales, and other authentic collections, make no scruple about proportion where their memory happens to fail them or their irrelevant fancy to distract them, but go on easily, dropping out a symmetrical adventure here and there, and repeating a favourite "machine" if necessary or unnecessary; so the story of Constance forgets and repeats itself. The voice is the voice of Chaucer, and so are the thoughts, but the order or disorder of the story is that of the old wives' tales when the old wives are drowsy. All the principal situations occur twice over; twice the heroine is persecuted by a wicked mother-in-law, twice sent adrift in a rudderless boat, twice rescued from a churl, and so on. In this story the poetry of Chaucer appears as something almost independent of the structure of the plot; there has been no such process of design and reconstruction as in the Knight's Tale.

It is almost as strange to find Chaucer in other stories, as in the Franklin's Tale and the Clerk's Tale, putting up with the most abstract medieval conventions of morality; the Point of Honour in the Franklin's Tale, and the unmitigated virtue of Griselda, are hopelessly opposed to anything like dramatic truth, and very far inferior as motives to the ethical ideas of many stories of the twelfth century. The truth of Enid would have given no opportunity for the ironical verses in which Chaucer takes his leave of the Clerk of Oxford and his heroine.

In these romances Chaucer leaves some old medieval difficulties unresolved and unreconciled, without attempting to recast the situation as he found it in his authorities, or to clear away the element of unreason in it. He takes the framework as he finds it, and embroiders his poetry over it, leaving an obvious discrepancy between his poetry and its subject-matter.

In some other stories, as in the Legend of Good Women, and the tale of Virginia, he is content with pathos, stopping short of vivid drama. In the Knight's Tale he seems to have deliberately chosen a compromise between the pathetic mood of pure romance and a fuller dramatic method; he felt, apparently, that while the contrast between the two rivals admitted of drama, the position of the lady Emily in the story was such as to prevent a full dramatic rendering of all the characters. The plot required that the lady Emily should be left without much share of her own in the action.

The short and uncompleted poem of Anelida gains in significance and comes into its right place in Chaucer's works, when it is compared with such examples of the older school as the Chastelaine de Vergi. It is Chaucer's essay in that delicate abstract fashion of story which formed one of the chief accomplishments of the French Romantic School. It is his acknowledgment of his debt to the artists of sensibility, the older French authors, "that can make of sentiment," and it proves, like all his writings, how quick he was to save all he could from the teaching of his forerunners, for the profit of "that fair style that has brought him honour." To treat a simple problem, or "case," of right and wrong in love, was a favourite task of medieval courtly poetry, narrative and lyric. Chaucer in his Anelida takes up this old theme again, treating it in a form between narrative and lyric, with the pure abstract melody that gives the mood of the actors apart from any dramatic individuality. He is one of the Extractors of Quintessence, and his Anelida is the formal spirit, impalpable yet definite, of the medieval courtly romance.

It is not here, but in a poem the opposite of this in fulness and richness of drama, that Chaucer attains a place for himself above all other authors as the poet who saw what was needed to transform medieval romance out of its limitations into a new kind of narrative. Chaucer's Troilus and Criseyde is the poem in which medieval romance passes out of itself into the form of the modern novel. What Cervantes and what Fielding did was done first by Chaucer; and this was the invention of a kind of story in which life might be represented no longer in a conventional or abstract manner, or with sentiment and pathos instead of drama, but with characters adapting themselves to different circumstances, no longer obviously breathed upon by the master of the show to convey his own ideas, but moving freely and talking like men and women. The romance of the Middle Ages comes to an end, in one of the branches of the family tree, by the production of a romance that has all the freedom of epic, that comprehends all good and evil, and excludes nothing as common or unclean which can be made in any way to strengthen the impression of life and variety. Chaucer was not tempted by the phantasm of the Epic Poem like Boccaccio, and like so many of the great and wise in later generations. The substance of Epic, since his time, has been appropriated by certain writers of history, as Fielding has explained in his lectures on that science in Tom Jones. The first in the line of these modern historians is Chaucer with his Troilus and Criseyde, and the wonder still is as great as it was for Sir Philip Sidney:—

Chaucer undoubtedly did excellently in his Troylus and Cresseid; of whom, truly I know not whether to mervaile more, either that he in that mistie time could see so clearely, or that wee in this cleare age walke so stumblingly after him.

His great work grew out of the French Romantic School. The episode of Troilus and Briseide in Benoit's Roman de Troie is one of the best passages in the earlier French romance; light and unsubstantial like all the work of that School, but graceful, and not untrue. It is all summed up in the monologue of Briseide at the end of her story (l. 20,308):—

Dex donge bien a Troylus!
Quant nel puis amer ne il mei
A cestui[90] me done et otrei.
Molt voldreie aveir cel talent
Que n'eüsse remembrement
Des ovres faites d'en arriere:
Ço me fait mal à grant manière!