Things to keep in mind:
The secret of ornament.
His argument was that the name we give to that supposed author disguised a bias modern folk have for writing and against listening. In Kanigel’s words, literary critics of the twentieth century associated reading and writing with “advanced civilizations,” and disliked the “repetition and stereotype” that characterizes oral poetry, a leaning that “blinded them to the fecund richness of illiterate cultures.” Although this theory had been modestly proposed by a number of scholars before Parry, and the groundwork was laid by the French scholar Marcel Jousse, who himself grew up amid the oral songsmithery of a largely illiterate community in France, Parry’s innovation lay in the scientific way he proved his inklings.
Using numerical methods, he counted exactly how many times the “ornamental epithets” so characteristic of Homeric epics—γλαυκῶπις Ἀθήνη (bright-eyed Athena), πολυφλοίσβοιο θαλάσσης (loud-roaring sea), and so on—appeared, and, crucially, where in the dactylic hexameter of the poem’s lines they cropped up. Such epithets were not, Parry showed, functionally descriptive at all; they tended to provide no new information about whatever story was being told, but instead existed for what he called the “convenience” of people performing the song. The epithets showed up in “prescribed position and order,” he wrote, essentially filling in metrical gaps wherever they occurred, “giving a permanent, unchanging sense of strength and beauty.”
—posted 1237 days ago
