Computational Stylistics in Poetry, Prose, and Drama pp 133–⁠152

Abstract

In “Unnecessary Duplicates,” renowned Melville editor Harrison Hayford discusses duplicate and vestigial settings, events, and characters in Moby-Dick,showing how they serve both wider aesthetic and structural effect in the novel. Their seemingly “unnecessary” nature prompts Hayford to offer several compositional hypotheses to explain their origins. With no extant manuscripts,Hayford relies on stylistic evidence, inductive reasoning, and Melville’s personal letters to posit three primary draft stages of the novel. As with Hayford’s duplicates, evidence of phenomena that produces data can be scant or absent. Multiple data sources may be patched together to paint a fuller picture of that phenomena. Model-based collaborative filtering methods can be used to characterize sparse data. Their outputs are recognizable in the recommendations of commerce and streaming websites. One such method, nonnegative matrix factorization (NMF), has proven successful in these tasks. Factorizing a table of numeric records approximates two factor matrices that potentially produced the table, thus estimating latent features not explicitly present in the data. NMF’s constraint of producing matrices with nonnegative values better isolates those features and produces more human-interpretable models. This chapter demonstrates how a probabilistic variant of NMF can be used to characterize the parts of speech in the sentences of Moby-Dick. Hayford’s ideas on Melville’s draft stages are explored and tested via the NMF model.

About

Armoza, J. (2023). Hayford’s Duplicates: Cobbling a Model of Melville’s Moby-Dick. In A. S. Bories, P. Plecháč, & P. Ruiz Fabo (Eds.), Computational Stylistics in Poetry, Prose, and Drama: (pp. 133–152). Berlin, Boston: De Gruyter. doi: 10.1515/9783110781502-008

DOI
http://doi.org/10.1515/9783110781502-008

Print ISBN
978-3110-781-41-0

Online ISBN
978-3110-781-50-2

Published under Creative Commons Attribution 4.0 International License (CC BY 4.0)