Digital Humanities and Artificial Intelligence
Digital Humanities (DH) brings traditional humanities research into conversation with digital tools, texts and technologies, from computational literary analysis to the creation and study of born-digital projects. As digital technologies continue to reshape human culture and communication, Artificial Intelligence (AI) has emerged as a growing area of DH research, with our faculty exploring how these technologies can expand humanities scholarship while addressing their broader cultural and ethical implications.
Digital Humanities (DH) can, at first glance, sound like an oxymoron: humanities research is often associated with physical books, the past and understanding human nature and culture, all of which may seem digital-resistant. Yet this field of research has exploded over the last few decades as humanities scholars turn their research and analytical skills towards digital life in the twenty-first century. What exactly counts as (DH) work is an ongoing debate (so much so that there is a website that will give you a new definition of DH every time you refresh), but most DH work falls into one (or more) of three categories: 1) using digital tools (e.g., Natural Language Processing) to conduct traditional humanities work (e.g., literary analysis); 2) using traditional humanities approaches and methodologies (e.g., discourse analysis) to study digital texts (e.g., social media communication); 3) creating born-digital projects (e.g., online archives) based on traditional humanities research.
While we have scholars engaging with DH in all three ways, one recent avenue of interest has been Artificial Intelligence (AI). AI is built on many of the same tools (e.g., Large Language Models) used in DH work; English scholars are researching elements (e.g., accessibility and privacy) of this new digital tool; and scholars even create their own AI engines or use AI engines to develop new humanities-based scholarly materials (e.g., NotebookLM's Complete Works of William Shakespeare). For more information about AI research happening in the department, check out the Semantic Artificial Intelligence and Creativity Laboratory.
Our faculty engage in cutting-edge research in both DH and AI.
Featured Publications
What is eye tracking? Why is it important for linguistics? How can I use it in my own research project? Answering these questions and more, this book guides you through one of the most exciting and innovative research methods in the field of linguistics. Divided into three parts, the chapters first offer an historical introduction and a foundational overview to the neurology and physiology of the eye and the common measurements and tools used in eye tracking. They then provide a guide to the applications of eye tracking most pertinent to linguists (reading, the visual-world paradigm, social eye tracking, and classroom applications), followed by a step-by-step process to plan, execute, analyze and report your research project in eye tracking. The book covers topics such as reading, lexical and syntactic processing, mind wandering, second language acquisition, and AAC devices, and includes statistical tools and how to write up results. Each chapter also includes self-study questions and a range of applied case studies. Supported by a glossary of key terms, suggestions for further reading, and material to aid self-study, Eye Tracking in Linguistics is the only book you need to provide a solid foundation for your own research project.
The book provides a comprehensive discussion of the new humor that has appeared on the internet. The book is divided into five sections: First, the introduction, which explains the idea that humor has changed since the widespread adoption of the internet and social media. The introduction reviews the theoretical tools that will be applied throughout the book: a discussion of humor theory and memes and how they function. The discussion is kept engaging and readable but is nonetheless based on rigorous scholarship, presented clearly by a well-known humor researcher. The first comprehensive guide to humor in the age of the internet and social media, this book will make you laugh (for the examples) and will enlighten you (for the analyses). Hopefully.
This paper builds on a novel methodology of lexical semantics exemplified on lexical field theory by using several translations of Ken Kesey's One Flew Over the Cuckoo's Nest. The present study, a large-scale collaboration, presents and compares the results for laugh, smile, grin, giggle, and other words for laughter behaviors across 14 languages and in extensive detail. The key results answer the question of what semantic dimensions the vocabularies of the various languages distinguish as marked by lexical contrasts and can inform future research in humor as well as translation studies. Based on our findings, a key marking emerges for audible (e.g., laugh) versus non-audible (e.g., smile) behaviors, as Indo-European vocabularies treat smiling as a less marked variant of laughing, e.g., German lächeln, Italian sorridere, Polish uśmiech, Turkish gülüm, but further orthogonal dimensions are documented as well, for example, aggressive, concealed, loud, or suppressed behavior. An updated hierarchy of these semantic features is proposed, and the results are presented in graphic visualizations, which also help illustrate idiosyncrasies of individual languages that go against the general trends. Exceptions to these general trends include lemmata that can cover both audible and inaudible behavior straddling what we claimed is the most important distinction (e.g., Danish grine). Finally, we outline a probabilistic method to compare word senses across languages based on aligned corpora large enough for computational approaches.
This chapter discusses two kinds of computational approaches to humor: knowledge-based humor models and machine-learning models. Both require an algorithm to manipulate humor as input or produce it as output. Such an algorithm is called a humor model because it needs to capture what is necessary and sufficient to create functional (that is, amusing) humor. A model can be built, first, on the basis of theories (like the GTVH) that are formalized into the relevant symbolic resources and processing steps, for example, topic lists, joke generation templates, and instructions for using them. The symbolic steps and resources of these theoretical or knowledge-based humor models are explainable and adaptable. As such they are of interest to humor research and will be the initial and general focus of this chapter. In the other approach, the computer builds an abstract model by “machine learning” humor through exposure to large amounts of relevant data. This probabilistic approach is currently favored and labelled “artificial intelligence”. Machine learning models don't have transparent components that would be explainable in a way relevant to humor research. But they can be finetuned and harnessed for specific steps and subtasks of symbolic humor models and will therefore be introduced in general and explained in their relevant roles in the second half of the chapter.
Writing center studies is currently reckoning with the complicit relationship between writing centers and the state. Here, we continue that work, inquiring specifically into digital surveillance as a colonial technology. Using the framework of feminist, queer, and decolonial surveillance studies, we critique examples of data collection in our Writing Center and offer takeaways to help writing centers reduce their dependence on data-based surveillance.
Read More: “Writing Centers are Watching: Surveillance, Colonialism, and Data Tracking”
Revising Marxist theories of circulation with affect theory, this article establishes a new model of rhetorical analysis that positions rhetorical exchange as a circulatory infrastructure of late capitalism. By measuring the value produced by rhetors and audiences in rhetorical exchange, we can see how the daily rhetorical activity of neoliberal subjects captures our behavior, positioning us a raw material for late capitalists. This new theory of rhetorical circulation is tested and revised by a qualitative study on the mundane communication of neoliberal subjects, in this case, the group chat of one fantasy football league. Fantasy football communication creates an ambient backdrop for its users, leading to quotidian rhetorical exchanges in clearly defined social networks. The study shows the contours of rhetorical exchange in one league’s GroupMe chat. I found that, in exchange, subjects transform their investments into social and cultural capital (Bourdieu’s capital forms). Ultimately, subjects can produce what I call affective capital, a uniquely neoliberal capital form. I find that the immense value of affective capital produced by league members in rhetorical exchange points to the reasons why neoliberal subjects repeatedly return to platforms that harvest our data.
When fans rewrite characters, how do they engage that character’s identity and the social constructions around it? Fan fiction writers resist, replicate, and create oppressive social systems by changing characters between published and fan texts. As such, fan studies scholars have long been interested in how fans construct characters, an interest that has often been paired with readings of race, gender, and sexuality. Digital humanities can help confirm and nuance extant fan studies scholarship around specific characters popular in fan fiction. We used Word2Vec software to mine the text of 450 pieces of fan fiction based on JK Rowling’s Harry Potter series. By focusing on the depiction of Hermione Granger in both Rowling’s novels and Harry Potter fan fiction, we tested how text mining character names can reveal properties closely tied to a specific character through the relationships between the target name and other characters. Analysis via Word2Vec found that” Hermione” is used grammatically and contextually differently in the books (in which she is most like Harry and Ron) than in our fan fiction corpus (in which she is most like other girls/women). This difference suggests that these fans have a specific reading of Hermione that is communally understood even if Rowling’s diction offers a different reading.
Read More: “Text Mining, Hermione Granger, and Fan Fiction: What’s in a Name?”
In this paper we examine the order of processing of multimodal tweets (text + image). Using an eye tracker, we collected a sample of 36 participants reading 25 humorous tweets. Our conclusions show that the processing of multimodal humorous tweets is in line with the processing of other multimodal texts. The participants were significantly more likely to start from the image, followed by the caption. Other elements, such as the tweet's “author” (the user name) or elements outside the tweet's frame, attracted significantly less and later attention. The participants spent significantly more time gazing at the caption, before moving on to another area. The longer the participants spent looking at the tweet, the less predictable their gaze direction became.
Read More: “On the Order of Processing of Humourous Tweets with Visual and Verbal Elements”
