Modeling ‘Worth by Association’ in U.S. Book Reviews, 1905–1925

How can we use computational methods to uncover hidden patterns in “book talk” materials, such as book reviews, advertisements, interviews with authors, and user-generated content from online platforms? How have attempts at mediation in book reviews changed over time in terms of their tendency to discuss a book’s genre and its physical features, as well as reviewers’ willingness to render judgment on a book’s overall merit?

In this presentation, Lavin will discuss his ongoing research on how institutions and technologies mediate reading and reception, and how pursuing this research can contribute to a more comprehensive theory of mediated reception. He will highlight how his work engages with word embeddings, which represent one of the crucial technologies underpinning recent advances in Large Language Models (LLMs). He will discuss how word embeddings are being used, how they work, and what kinds of errors they often generate. Looking closely at word embeddings can help us understand and interpret the kinds of predictions these models make and teach us about the potential risks and rewards of AI.

Matthew J. Lavin is an Assistant Professor of Humanities Analytics in the Data Analytics Program at Denison University. He earned a PhD in English from the University of Iowa in 2012, a master’s degree in American studies at Utah State University in 2006, and a bachelor’s degree at St. Lawrence University in 2002. From 2012 to 2013, Lavin served as a Council on Library and Information Resources (CLIR) Postdoctoral Fellow at the University of Nebraska – Lincoln’s Center for Digital Research in the Humanities (CDRH). From 2013 to 2015, he was Associate Program Coordinator for the Andrew W. Mellon Foundation Initiative “Crossing Boundaries: Re-Envisioning the Humanities for the 21st Century” at St. Lawrence University. From 2015 to 2020, he was a Clinical Assistant Professor of English and Director of the Digital Media Lab at the University of Pittsburgh. Lavin’s scholarship focuses on how cultural analytics modes of inquiry can advance literary studies, book history, and digital humanities, while engaging critically with tools and methods that require adaptation or iteration to make sense in humanities contexts. His publications have appeared in Auto|Biography Studies, Cather Studies, CA: The Journal of Cultural Analytics, Digital Humanities Quarterly, Digital Scholarship in the Humanities, The Journal of Open Humanities Data, The Programming Historian, Studies in the Novel, and Western American Literature, among others.

Throughout 2025–26, CDH will host six public talks by leading scholars in Cultural Analytics. Open to all, the series invites audiences to explore the histories, theories, and practices that shape the evolving intersection of AI and the humanities.

Since the release of ChatGPT, conversations about artificial intelligence (AI) have generated both excitement and concern across the humanities. Much of the debate has focused on political and ethical questions—bias, labor, environmental impact, and intellectual property—as well as the effects of large language models (LLMs) on teaching and learning. Yet one question remains underexplored: how might AI contribute directly to humanities scholarship?

Modeling Culture is generously funded by an AI Lab Seed grant, with additional support from the Humanities Council and the Princeton Humanities Initiative. Thank you to the Department of English for co-sponsorship of the public talks.