AN AUTOMATED APPROACH TO SYNTAX-BASED ANALYSIS OF CLASSICAL LATIN

Author/​Artist
Field, Anjalie [Browse]
Format
Senior thesis
Language
English
Description
50 pages

Availability

Available Online

Details

Advisor(s)
Fellbaum, Christiane [Browse]
Contributor(s)
Baraz, Yelena [Browse]
Department
Princeton University. Department of Computer Science [Browse]
Class year
2015
Summary note
The main goal of this study is to analyze the style of Latin authors using automated tools. Many of the common methods in stylistic analysis are based on lexical measures, which do not work well with Latin because of the language’s high degree of inflection and free word order. Thus, this study focuses on analysis at a syntax level by examining two constructions, the ablative absolute and the cum clause. Because these constructions are both used adverbially, they are somewhat interchangeable, which suggests an author’s choice of construction is often more stylistic than functional. We first identified these constructions in hand-annotated texts by relying on the annotations. Then, we developed two methods for identifying these constructions in unannotated texts, a purely rule-based method and a method that involved probabilistic morphological tagging. These methods identified constructions with enough accuracy to distinguish between different genres and between different authors. In particular, these methods were able to determine which book of Caesar’s Commentarii de Bello Gallico was not written by Caesar. Furthermore, the usage of ablative absolutes and cum clauses observed in this study is consistent with the usages previously observed by scholars analyzing the same texts by hand. Overall, this study demonstrates the usefulness of automated syntax-based analysis in the study of classical literature and proposes a few methods for conducting this analysis.
Statement on language in description
Princeton University Library aims to describe library materials in a manner that is respectful to the individuals and communities who create, use, and are represented in the collections we manage. Read more...

Supplementary Information