When we ask the reason for this curious separation between classical motifs invested with a nonclassical meaning, and classical themes expressed by nonclassical figures in a nonclassical setting, the obvious answer seems to lie in the difference between representational and textual tradition. The artists who used the motif of a Hercules for an image of Christ, or the motif of an Atlas for the images of the Evangelists acted under the impression of visual models which they had before their eyes, whether they directly copied a classical monument or imitated a more recent work derived from a classical prototype through a series of intermediary transformations. The artists who represented Medea as a mediaeval princess, or Jupiter as a mediaeval judge, translated into images a mere description found in literary sources.

This is very true, and the textual tradition through which the knowledge of classical themes, particularly of classical mythology, was transmitted to and persisted during the Middle Ages is of the utmost importance, not only for the mediaevalist but also for the student of Renaissance iconography. For even in the Italian Quattrocento, it was from this complex and often very corrupt tradition, rather than from genuine classical sources, that many people drew their notions of clasical mythology and related subjects.

Limiting ourselves to classical mythology, the paths of this tradition can be outlined as follows. The later Greek philosophers had already begun to interpret the pagan gods and demigods as mere personifications either of natural forces or moral qualities, and some of them had gone so far as to explain them as ordinary human beings subsequently deified. In the last century of the Roman Empire these tendencies greatly increased. While the Christian Fathers endeavored to prove that the pagan gods were either illusions or malignant demons (thereby transmitting much valuable information about them), the pagan world itself had become so estranged from its divinities that the educated public had to read up on them in encyclopaedias, in didactic poems or novels, in special treatises on mythology, and in commentaries on the classic poets. Important among these late-antique writings in which the mythological characters were interpreted in an allegorical way, or "moralized," to use the mediaeval expression, were Martianus Capella's Nuptiae Mercurii et Philologiae, Fulgentius' Mitologiae, and, above all, Servius' admirable Commentary on Virgil which is three or four times as long as the text and was perhaps more widely read.

During the Middle Ages these writings and others of their kind were thoroughly exploited and further developed. The mythographical information thus survived, and became accessible to mediaeval poets and artists. First, in the encyclopaedias, the development of which began with such early writers as Bede and Isidorus of Seville, was continued by Hrabanus Maurus (ninth century), and reached a climax in the enormous high-mediaeval works by Vincentius of Beauvais, Brunette Latini, Bartholomaeus Anglicus, and so forth. Second, in the mediaeval commentaries on classical and late-antique texts, especially on Martianus Capella's Nuptiae, which was annotated by Irish scholars such as Johannes Scotus Erigena and was authoritatively commented upon by Remigius of Auxerre (ninth century). Third, in treatises special treatises on mythology such as the so-called Mythographi I and II, which are still rather early in date and are mainly based on Fulgentius and Servius. The most important work of this kind, the so-called Mythographus III, has been tentatively identified with an Englishman, the great scholastic Alexander Neckham (died 1217); his treatise, an impressive survey of whatever information was available around 1200, deserves to be called the conclusive compendium of high-mediaeval mythography, and was even used by Petrarch when he described the images of pagan gods in his poem Africa.