For more than two millennia the Iliad and the Odyssey were read as the work of a single poet, Homer, whose genius the literate tradition took for granted even as it disputed his biography. The oral-formulaic theory, developed in the 1930s by Milman Parry and extended by his student Albert Lord, unsettled this assumption not by adducing new manuscripts but by attending to a feature the poems had always displayed and scholarship had always overlooked: their pervasive repetition. Fixed epithets - swift-footed Achilles, wine-dark sea - and recurring whole lines had struck earlier readers as ornament or, at worst, as lapses. Parry proposed that they were instead the load-bearing architecture of a compositional technique. His argument turned on economy. The formulaic phrases, Parry showed, were distributed with a systematic thrift: for a given metrical slot and a given grammatical case, the tradition tended to supply one and only one epithet for a given hero, neither more (which would be wasteful) nor fewer (which would leave gaps). Such a system, he inferred, could not be the deliberate contrivance of a lettered author composing at leisure; it bore the signature of a poet composing in performance, drawing on an inherited stock of phrases to meet the relentless forward pressure of the meter. To test the inference, Parry and Lord turned to the living oral tradition of the South Slavic guslari, singers who composed lengthy heroic songs extempore, and found the same economy at work. The theory's reception exposed a fault line that persists. In its strong form, oral-formulaic theory implied that the Homeric poems were not composed by anyone in the sense the word ordinarily bears - that Homer names a tradition rather than a person, and that the search for individual artistry in the epics is a category error. Critics resisted this deflation. The presence of formulaic diction, they countered, establishes the medium of composition, not its ceiling; the guslari themselves varied enormously in skill, and a tradition that furnishes a poet with materials does not thereby compose the poem. That an epithet is inherited says nothing about whether its placement in a particular line is inspired. Defenders of the strong thesis replied that this rejoinder quietly reintroduces the Romantic author the theory had displaced, smuggling back through the vocabulary of skill and inspiration the very individuality the formulaic evidence renders invisible. The dispute, in this light, is not primarily about facts - both sides accept the pervasiveness of formula - but about the burden of proof. Does the demonstrable orality of the diction oblige us to withhold attributions of individual artistry until they are independently earned, or does the manifest power of the poems entitle us to presume an artist until the presumption is defeated? No decisive experiment can settle a question so posed, for it concerns what the evidence must be taken to license rather than what the evidence is. What Parry established beyond reasonable dispute is that the epics emerged from an oral tradition; what he did not establish, and perhaps could not, is that emergence from such a tradition is incompatible with the individual genius the poems have always seemed to display. The formula constrains the singer; whether it exhausts him is another question, and one the evidence leaves open.