Skip to main content

The Uniqueness of Homer, Reconsidered

It has been almost a century since Milman Parry made the most significant advance in understanding the nature of Homeric narrative by focusing attention on the complex systems of expression (gathered in Parry 1971). The transformative power of his insight, achieved by applying a functional analysis to the materials assembled decades before in the 19th-Century single-line concordances of Prendergast and Dunbar and in Schmidt’s Parallel-Homer, has allowed scholars to see purpose behind the previously impenetrable surface variety of phrases, especially the stringent principles of formulaic economy.

Researchers since Parry’s day have explored the intricate subtleties of what is called the “oral-formulaic” element of the great poems, from literary interpretations to tracing Indo-European ancestry to elaborate statistical studies. But the interest in repeated word-groups has tended to obscure the surprisingly large volume of non-repeated expressions in the two epics, in particular, the multiplicity of ways in which identical or similar ideas can be formulated.

Michael Kumpf (1984) assembled the hapax legomena, bringing modern methods to a category which had already been explored by the Alexandrians. Commentators in the 1960s like Bryan Hainsworth spoke of “under-represented formulae,” offering a few hundred examples, whereas James Dee’s collection of hapax ezeugmena (2010) has ca. 4000 noun-epithet phrases, a clear indication that this group deserves investigation beyond a mere listing. Edzard Visser (1987) made a more auspicious start, examining the poet’s ways to say “X killed Y,” which opens up a broader avenue of approach--but one rarely taken thus far.

This paper follows Visser’s lead, describing a plan to create a different kind of reference database--a large-scale register of complete indicative main-verb sentences, which after all are the true “building-blocks” of epic narration. It would allow for the first time a synoptic overview of the full spectrum of Homeric expression, from fixed repetition to individualized novelty--something not directly visible in the old concordances or the KWIC (keyword-in-context) volumes of Joseph Tebben (1994, 1998). Such a repertory will expand the notion of the “uniqueness” of Homer, examined in a seminal article by Jasper Griffin (1977), by mapping the complementary territories of formularity and non-recurrence with unprecedented precision.

The opening section will summarize the problem with contrasting examples of extreme formularity and unique expressions, accompanied by statistical comparisons to illustrate the larger context. It will be apparent that there is more to Homeric narrative than the strictly formulaic, i.e. that “Homer coined,” to quote from an unpublished paper by the expert statistician William Merritt Sale (1990). The central part will consider the challenges of gathering a full-sentence database and the possible use of computer resources to facilitate the assembly; there is a distant precedent in Marvin Spevack’s venerable Shakespeare concordance (1969), which used a program to determine context for each lemmatized word. In this case, there would be no arbitrary cut-offs: the start- and stop-points would be the beginning and end of each utterance, as excerpted from a tagged text.

The final section will offer preliminary data (completion of this large-scale task being uncertain by the time of delivery), along with a handout to demonstrate the advantages of this presentation, highlighting the range of originality that is possible within the confines of a formulaic tradition.