Language and Chronology, Toner and Han apply innovative Machine Learning techniques to the problem of the dating of literary texts. Many ancient and medieval literatures lack reliable chronologies which could aid scholars in locating texts in their historical context. The new machine-learning method presented here uses chronological information gleaned from annalistic records to date a wide range of texts. The method is also applied to multi-layered texts to aid the identification of different chronological strata within single copies.
While the algorithm is here applied to medieval Irish material of the period c.700-c.1700, it can be extended to written texts in any language or alphabet. The authors’ approach presents a step change in Digital Humanities, moving us beyond simple querying of electronic texts towards the production of a sophisticated tool for literary and historical studies.
Gregory Toner is Professor of Irish at Queen’s University, Belfast. He has published widely on medieval Irish literature and language and is editor of the electronic Dictionary of the Irish Language.
Xiwu Han, Ph.D. (2006), Harbin Institute of Technology, China, has published numerous peer-reviewed articles on computational linguistics and machine translation.
Table of contents
List of IllustrationsAbbreviationsIntroduction
Automated Dating Methods
How to Read This Book 1
Dating Texts: Principles and Methods
Texts by Known Authors
Computational Approaches to Text Dating
A Brief History
The Problem Stated
Trials in English and Medieval Irish Texts
Dating English Texts
Dating Medieval Irish Texts
Dating Long Documents
Building a Datable Medieval Irish Corpus
Dating Long Documents
Establishing the Date of Composition
Transmission and Manuscript Dates
Focussed Dating Predictions
A Temporal Model
Towards a Tool: Computational Chronometrics
Applicability to Other Literatures
Appendix A: Conventional Dating of Texts Used in This Study
Appendix B: Machine Learning
Classification, Regression and Clustering
Other Relevant Statistics
Scholars concerned with dating ancient and medieval texts, e.g. in Biblical Studies, Anglo-Saxon and Celtic. Also, Digital Humanities scholars interested in high-level applications of machine learning for solving complex problems.