How can poems, built with such restrictive rules and materials as the Homeric poems, still produce a sense of
artistic unity and great quality?
If unity exists, if the poem as a whole is to make sense, the source of this may lie in tradition (1). This is no doubt partly true, but I would like to shift perspective a little and find the source of the unity as well as of the formulaic system in the rhetoric, that is to say: in the intentions of the poet. I know that to state it like this, goes against many axioms of literary science but I propose another set of axioms (see the introduction page) that may shed some light on the issue of unity.
In the beginning of the 20th century, scholar finally noticed that the Homeric poems are for a large part made uo of repeated material, phrases, whole lines and scenes, and that these repeated phrases can be found, most of the time, at the same place in the hexameter. This, and field research in Yugoslavia, led to the recognition that the craft of the aoidos is not literary but oral. The implications of this discovery were and are still debated extensively but the basic fact as stated above seems undeniable. However, the conclusions we can draw from this, especially regarding interpretation, are are not so clear cut. What we call formulas is to a degree a product of our own literate habits. We think language is made out of words and when we write it down we add spaces between the words. But this is not how a singer deals with them. To him, a phrase (formula) is not a collection of words but music: they are phrases with a rhythm, (most likely) some kind of melody and a collection of meanings. The rhythm automatically suggests a place in the hexameter. They are seen and remembered as units, the building blocks of poetry. Hence, in analogy to 'phoneme' and 'morpheme' I would like to call them 'poeme' rather than formula because for singers like Homer, these were the stones that their temple was built with.
This does not mean that the artist is My hypothesis is that the 'system' we find in Homer, is mostly an artifact of individual memory: a collection of poemes from the memory of one man, selected for use in one poem. Naturally a lot of these phrases are traditional and some are very old indeed but they are selected for their meaning and applicability in what the poet wishes to convey. He does not need or want more of them than he actually uses. Part of the function of the repetition of poemes may be, that it makes it easier for the public to recognize and remember them.
All this still seems very restrictive. To see why this is our problem and not the poet's, we need to be aware of another basic of Homeric poetry: these are not novels. In a modern novel, we tell a tale about a man, Homer tells us about the man. Not about Achilles but about the idea of Achilles(1).
Important for understanding this kind of poetry is to recognize that everything that is in focus at any moment, becomes generic. 'Achilles' the character is as generic as his epithets and poemes. Every ship, arrow, character or important concept like running, rowing and so forth can be lifted out of the immediate context, turn into a metaphor and become a part of the subtext, the pseudo-hidden meaning which the poet is trying to convey. Which brings us to the next item:
To quote Umberto Eco: Allegory is metaphor which could be taken literally. We suspect its presence:
a) because of convention
b) because there must be relevance
Convention is hardly accessible to us. All we can do is try to reconstruct some of it. But there is something else: allegory in Homer works 'by likeness(4)', surface likeness, not
farfetched ones, taking it as an axiom that there are no dead metaphors in Homeric poetry. Everything is
consciously and purposefully included. In isolation this reconstruction of meaning is naturally a dangerously
subjective and unfalsifiable step. But we do know that they thought like that, I would go so far as to say that the
ancients always were aware of any metaphorical meaning in their words. You could say that their 'is-like-a'
was a more important classification operator to them than our 'is-a'.
That leaves the touchstone of 'relevance'. Interpretation of any single part must make it relevant to the unity that we seek. Here, I trust that if there are errors on the way, or if the whole concept is flawed, this will inevitably show up as the impossibility to unify.