- Oral Theory and the Limits of Formulaic Diction
It has been pointed out more than once that the theory of formulaic composition in the form originally given it by Milman Parry is far from homogeneous.1 Not only chronologically but also in content Parry's studies of the Homeric formulae fall into two parts: the French publications of the 1920s and the American publications of the 1930s. In the former, Parry showed, first, that there are formulae and formulaic systems in Homer, and second, that they are characterized by extension and economy. He identified the style he thus described as "traditional." It is only in the American publications of the 1930s that he introduced the hypothesis that the formulaic character of the Homeric style is to be explained by its being the characteristic style of oral composition. A by-product of this development was that for all practical purposes "oral" became identified with "formulaic," giving rise to the widespread view that 100% orality amounts to 100% formularity. From then on, to claim that Homeric diction is oral-traditional has become equivalent to claiming that all of Homer consists of formulae.
As is well known to every student of Homer, neither Parry's definition of formula nor the rules of repetition, economy, and extension that he introduced would apply equally to all Homeric expressions. This fact became obvious to Parry himself as soon as he moved from noun-epithet combinations to other parts of Homeric diction. A partial solution that he proposed was, first, to introduce the notion of so-called "formulaic expressions," that is, such modifications of the traditional formulae that adapt them to situations not provided for in the poet's stock of traditional phrases; and, second, to approach all the unique expressions that cannot be accounted for in this way as underrepresented formulae. But even when both formulaic expressions and underrepresented formulae are taken into [End Page 236] account, this would not change the fact that, insofar as one proceeds from Parry's definition of the formula, the claim that 100% of Homer is formulaic is highly problematic.
A priori, there were two ways out of this difficulty: (a) to abandon the idea of 100% formularity in Homer, or (b) to loosen the criteria for identifying the basic unit of Homeric composition so that they may apply equally to all Homeric expressions. Yet insofar as in the minds of most scholars "oral" became equivalent to "formulaic," adopting (a) would have amounted to recognizing that while some parts of Homer are oral others are not. This is why both Parry himself and his disciples Albert Lord and J. A. Notopoulos chose the other option, with most of the Parryists following their lead. As a result, not only single words regularly placed in the same metrical positions but also recurring syntactic and phonic patterns not identical in their wordings gradually began to be treated as formulae. Soon enough, the tendency to stretch the definition of the formula so that it might apply equally to all Homeric expressions began to be felt unsatisfactory by many. This led to the efforts, discussed in detail below, to replace the formula with another, more flexible, unit of composition.
In all this, it has largely been overlooked that abandoning the original definition of the formula while preserving the thesis of the oral character of Homeric composition inevitably leads to a fallacy, for the simple reason that the only foundation we have for the hypothesis of oral composition of the Homeric poems is the Homeric formula as identified by Parry.2 This fallacy is characteristic of those recent developments in Homeric scholarship that seek to modify the formula in such a way that it is replaced by other entities that are ostensibly more apt to supply a unified explanation for the totality of the Homeric text. In Section I of this paper, I will dwell on two such developments, both of them issuing from the application of contemporary linguistic theories to the phenomenon of Homeric language—the "generative approach" launched by Michael Nagler (1967 and 1974) and the "nucleus-periphery" theory introduced by Edzard Visser (1987 and 1999) and further developed by Egbert...