What Would Cicero Write?

Examining Critical Textual Decisions with a Language Model

Authors

  • Todd G. Cook, TGC Classical Language Toolkit (CLTK.org)

DOI:

https://doi.org/10.13135/2532-5353/6523

Abstract

Recent developments in Transformer language models now allow users to predict the probability of different sentences and to predict missing words more accurately than before. This new information and perspective can be used to form judgments on novel textual emendations and to further quantify existing historical editorial judgments. We examine the importance of analyzing an author’s corpus, and the impact of the Good-Turing theory of frequency estimation when predicting missing words. We will also outline some of the limits of what Transformer language models can do, and how to practically evaluate them.

Downloads

Download data is not yet available.

Author Biography

Todd G. Cook, TGC, Classical Language Toolkit (CLTK.org)

Todd G. Cook is a core contributor to the Classical Language Toolkit (CLTK.org), and he has studied Classics at California State Universities of Chico and Long Beach. He works as a data scientist and software engineer with years of experience writing educational software.

Downloads

Published

2021-12-31

How to Cite

Cook, T. G. (2021). What Would Cicero Write? Examining Critical Textual Decisions with a Language Model. Ciceroniana On Line, 5(2), 285–296. https://doi.org/10.13135/2532-5353/6523

Issue

Vol. 5 No. 2 (2021): “Cicero digitalis” Congress Proceedings edited by Alice Borgna and Mélanie Lucciano

Section

Papers

License

Authors who publish with this journal agree to the following terms:

  1. Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
  2. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.