This appendix presents a computational analysis of noun frequency distributions across three texts: Homer’s Odyssey, Zachary Mason’s The Lost Books of the Odyssey, and Madeline Miller’s Circe. The investigation employs natural language processing techniques to map what I term “agent-prints”—distinctive linguistic signatures that reveal how contemporary retellings negotiate their relationship with Homeric authority. This quantitative approach supplements the dissertation’s qualitative analysis by exposing underlying linguistic patterns that persist despite surface-level narrative innovations.
The purpose extends beyond descriptive statistics to interrogate a fundamental paradox: how texts that explicitly challenge epic conventions remain linguistically bound to patriarchal semantic fields. By visualizing noun distributions as constellations of agency, we can trace both movements toward and resistances against classical archetypes, revealing the algorithmic substrate that governs even radical retellings.
Using Python’s NLTK and spaCy libraries, I extracted and ranked the most frequent nouns from each text, treating these lexical items as nodes in semantic networks. The visualization strategy—mapping these frequencies as agent-prints—draws from Actor-Network Theory’s understanding of distributed agency, where meaning emerges through relational patterns rather than isolated terms. Each noun functions as an actant, generating effects that exceed its individual semantic value through accumulation and iteration.
The Odyssey’s frequency distribution presents the expected epic inventory: proper names of heroes (odysseus, telemachus, athena, penelope), spatial markers (home, house, sea, ship), and hierarchical positions (suitors, gods, king, father). This establishes what we might call the “algorithmic genome” of epic—the foundational semantic patterns that subsequent texts must negotiate.
Mason’s Lost Books maintains the heroic roster (odysseus, achilles, agamemnon) while introducing phenomenological markers (eyes, time, night, face). The persistence of men, man, and war alongside sensory terms suggests a hybrid discourse—postmodern fragmentation operating within classical masculine parameters. The enigmatic presence of white hints at Mason’s interest in absence, erasure, and spectral presences, yet these experimental gestures remain anchored to traditional epic terminology.
Miller’s Circe presents the most intriguing configuration. While father, men, gods, and odysseus dominate—revealing persistent patriarchal framing—the prominence of embodied terms (face, eyes, hands, voice) signals a crucial shift toward what the dissertation identifies as “logosomatic” narration. The term father occupying the primary position exposes how even feminist retellings orbit around masculine authority, unable to escape the gravitational pull of patriarchal language.
The computational analysis reveals three distinct yet overlapping semantic strategies:
Classical Preservation (Homer): The archetypal distribution establishing epic’s linguistic DNA—heroes, gods, geography, and social hierarchies forming a stable semantic constellation.
Phenomenological Supplementation (Mason): Retaining epic’s masculine framework while adding temporal and sensory dimensions, creating what might be termed “modernist epic”—formally experimental yet semantically conservative.
Embodied Relationality (Miller): Attempting feminist revision through somatic vocabulary while remaining linguistically tethered to patriarchal referents, producing a paradoxical “gynocentric narrative in androcentric language.”
The striking absence of explicitly feminine nouns from top frequencies in Circe—no woman, goddess, daughter, or mother in the highest ranks—suggests that feminist mythological revision faces a fundamental linguistic challenge: the available vocabulary for classical narratives remains structurally masculine. Even when centering female experience, the text must articulate itself through relationships to male figures.
This algorithmic reading demonstrates how computational methods can illuminate ideological tensions invisible to traditional hermeneutics. The persistence of patriarchal terminology across all three texts—despite their radically different narrative strategies—reveals language itself as a conservative force, preserving power structures even within revolutionary retellings.
Yet the variations are equally significant. Mason’s addition of temporal-sensory terms and Miller’s emphasis on embodied vocabulary represent different tactics for infiltrating epic discourse. Where Mason fragments narrative while maintaining linguistic stability, Miller preserves narrative coherence while attempting lexical transformation. Neither fully escapes the “algorithmic genome” of epic, but both create interference patterns—moments where contemporary consciousness disrupts classical authority through subtle semantic shifts.
The agent-prints thus function as diagnostic tools, revealing how deeply patriarchal structures embed themselves in language’s very infrastructure. They also suggest why contemporary retellings remain necessary: not to escape these linguistic constraints—perhaps an impossibility—but to make them visible, to expose the algorithms that generate our mythologies. Through computational analysis, we discover that the question is not whether we can tell stories outside patriarchal language, but how we can use that language against itself, creating what Audre Lorde warned might be impossible: dismantling the master’s house with the master’s tools—or at least revealing the blueprints.