| This dissertation combines a computational study of low-level structural patterns in the poetry of the Iliad with a general theory of oral-formulaic composition that both explains the origin of such structures and argues for their importance in the creation of subjective meaning for the poet and his audience. In the first part of the paper, evidence from the fields of cognitive science and linguistics, and theoretical language from the critical study of intertextuality are brought to support aspects of oral theory which currently match well with the evidence of the poems but lack explanatory mechanisms. A model of oral-formulaic composition is presented which emphasizes the role of pattern recognition in expert singers' ability to store and reproduce long works such as the Homeric epics. In this view, mnemonic links between poetic elements at various structural levels, particularly at the sub-word level, allow the trained singer to "chunk" the linguistic data, grouping suites of features in order to facilitate efficient storage and recall. The singer learns his material by subconsciously recognizing patterns in the poetic language, and in turn imposes these patterns upon the song in performance. This model agrees well both with current research in cognitive science and with the principles of intertextual theory, in particular in its earlier, structuralist form. While the cognitive model provides important grounding for oral theory in empirical evidence, the language of intertext provides helpful new perspectives that emphasize the ability of poetic discourse to create meaning for its participants.;The second part of the paper presents a detailed, quantitative study of two classes of low-level mnemonic features, metrical localization and phonemic anaphora. In each case, the work of previous scholars is replicated and extended through computational analysis of a digital text of the Iliad. While the effects identified by earlier scholars are confirmed by these methods, they are also shown to be much more complex and subtle than was originally believed. In the case of metrical localization, the tendency of words of a given shape to prefer a small subset of possible positions in the hexameter line, first recognized by Eugene O'Neill, Jr., is confirmed by an exhaustive study of the text. At the same time, patterns of localization are shown to be more specific than O'Neill understood, with words of the same shape localizing differently according to factors such as the presence of initial consonants. Phonemic anaphora, demonstrated in Hesiod by Berkley Peabody, is shown also to be present in the Iliad, and at least in some cases to be a significant motivation of verse structure. Perhaps most importantly for the theory outlined in the earlier part of the paper, neither of these classes of poetic features can be seen as working alone. Rather, in all the examples presented here these low-level features operate as part of a network of mnemonic associations; these features work with each other and with structures such as syntax, semantics, and above all with traditional formulas to shape the composition of the line.;The final portion of the paper explores the broader ramifications of the theory presented here, both for Homeric studies in particular and for the larger field of intertextual literary criticism. It is shown that contemporary oralist approaches to Homeric poetics demonstrate a reticence to attribute to formulaic or traditional composition too much of the content, and thus the meaning, of Homer's poetry. This creates an ambivalence towards the formulaic which ultimately represents a serious difficulty for a field dedicated to understanding the nature of traditional language, manifested in calls for a separate, specifically oral aesthetics, or even in scholars' explicit recusals from pursuing their investigations so far that the poet's creativity is in danger of being explained away. Such fears are unfounded, it is claimed, according to the intertextual/cognitive approach taken here. The formulaic structure of oral poetry is the product of a pattern-recognition process that must be honed by the singer through many hours of hard work and experience with the tradition. The cognitive science cited in this study's first part shows that the ability to detect and manipulate patterns of the sort identified here is only possible at the highest levels of expertise. It can thus be argued that even the formulaic elements of Homer's language are thoroughly Homer's, if not in the way that a Romantic notion of the free, creative artist might expect. From this point of view, a deeper understanding of the formulaic cannot detract from our understanding of who Homer was (and is) but can only enlarge it. |
