#1 Seminar | 30 April 2024 (Tue)
From Corpora to Semantic Spaces: how computational methods can help us uncover word meaning change in historical texts
Date: 30 April 2024 (Tue)
Time: 3:00-4:30pm
Venue: Digital Scholarship Lab, University Library
Speaker: Dr. Barbara McGillivray
The growing availability of digitised textual collections and advancements in data-intensive methodologies from the computational sciences present a unique opportunity to apply these techniques in digital humanities. The investigation of lexical semantic change, i.e. the evolution of word meanings over time, represents a particularly promising endeavour. This area holds broad appeal and relevance extending well beyond linguistics, encompassing disciplines such as history, literary studies, and lexicography. Computational linguistics research has increasingly focused on this challenge in recent years, leading to the development of various algorithms aimed at automatically detecting semantic change from corpus data.
This seminar will explore computational methods for the analysis of semantic change in texts. Dr McGillivray will present her insights into the application of computational tools to investigate shifts in word meanings over time, drawing on her experience of working on contemporary texts and 19th-century English texts. This interdisciplinary approach bridges computational linguistics with digital humanities scholarship, offering novel perspectives on language evolution and cultural transformations through quantitative analysis.
#2 Workshop | 2 May 2024 (Thu)
From Corpora to Semantic Spaces: some practical examples of corpus analysis and computational processing to study word meaning change in historical texts
Date: 2 May 2024 (Thu)
Time: 3:00-4:30pm
Venue: Digital Scholarship Lab, University Library
Speaker: Dr. Barbara McGillivray
The question of what words mean and how they come to have multiple meanings has fascinated generations of linguists and philosophers. In recent years, computational linguists have addressed this research area as a detection challenge: how can we automatically identify whether a word has changed semantically just from its contexts of usage in corpus data?
The aim of this workshop is to support students and researchers interested in exploring the changing nature of the vocabulary in historical texts at scale, and to reflect critically on the limitations of these computational analyses. The workshop will cover the topics of corpus building and text processing. Dr. McGillivray will introduce the basic concepts of semantic spaces and distributional semantics and walk through some simple code for training computational methods for representing word meaning (word embeddings) from 19th-century English corpora, evaluating them, and using them in digital humanities research with reference to the corpus of Darwin’s letters from the Darwin Project.
About the Speaker
Barbara is a digital humanist and computational linguist. Her research interests lie at the intersection between computational and quantitative methods and research questions in the Humanities. She is particularly interested in the analysis of semantic change via quantitative and computational methods, computational models of word meaning, time-aware natural language processing, and computational lexicography.