23 August 10:00 - 11:3
Causal inference is hard, and everyone knows it. It is less recognized that descriptive and comparative scholarship also rely upon causal inference. How data are sampled and curated influences how we should process the data, in order to accurately describe or compare the people, times, and places of interest. I'll present some examples to illustrate the problems that ignoring causal structure can create, along with some solutions.
Bio: I am an evolutionary ecologist who studies humans. My main interest is in how the evolution of fancy social learning in humans accounts for the unusual nature of human adaptation and extraordinary scale and variety of human societies. Humans are more widespread and successful than any other vertebrate. Simultaneously, humans are unlike any other animal in that we cooperate in very large groups of unrelated individuals. I and my colleagues use formal evolutionary models, experiments and ethnographic fieldwork to address these puzzles.
24 August 9:30 - 11:00
We are living in an era of abundant, fast-circulating and easily twisting information. Among all other stories, nation bound historical narratives shape our identities and worldviews and through that have a potential to shake politics and international relations. Popular historical stories circulating in the world wide web contain a diversity of historical narratives, including ones that consciously contest the academic research and spread simplified understandings based on conspiracy theories, and that thus can be defined as ‘pseudohistory’. This plenary introduces an ongoing project that seeks to understand the process of pseudohistorical content development in the current era. The project explores the global topic through a case study of circulating pseudohistorical narratives on Russian medieval history in the Russian language web. The used data contains 1,5 million websites, blogposts and discussion forum posts addressing the topic of the origins of Russian state in the middle ages. The project utilizes text reuse detection, network analysis and topic modeling in its effort to detect the structures and dynamics of evolution of fake historical narratives.
Mila Oiva (milaoiva@tlu.ee) is a cultural historian enthusiastic about identifying patterns of transnational circulation of knowledge and ideas in a long temporal perspective. Her current ongoing projects are related to circulation of pseudohistorical narratives in the world wide web in the 2000s, newsreel production in the Soviet Union in the 20th century, and circulation of news in the 19th century. She works as a Senior Research Fellow at CUDAN Open Lab at Tallinn University. Her most recent publications include for example Digital Readings of History. History Research in the Digital Era. Helsinki: Helsinki University Press, 2020 (open access, co-edited together with Mats Fridlund and Petri Paju), Yves Montand in the USSR: Cultural Diplomacy and Mixed Messages. Palgrave Macmillan, 2021 (coauthored with Hannu Salmi and Bruce Johnson) and “Topic Modeling Russian History.” In The Palgrave Handbook of Digital Russia Studies, edited by Daria Gritsenko, Marielle Wijermars, and Mikhail Kopotev. Palgrave Handbooks. London: Palgrave Macmillan, 2021.
Link to the recording: https://youtu.be/lVzPAYtaKmk
25 August 9:30 - 11:00
In recent years, virtual reality (VR) equipment has become much more accessible and affordable for consumers as well as researchers. In this keynote, I will introduce how this technology can be used in the humanities and social sciences to study human behaviour directly in the lab. In the past, experimental rigour and ecological validity have often been conceived as the opposites of a methodological continuum, where researchers would have to sacrifice ecological validity to assure sufficient control. However, VR experiments can provide participants with immersive and realistic tasks that nevertheless allow researchers to tightly control all variables involved. I will show some examples from my own research in the area of language evolution that applies this approach. I used interactive VR experiments to study the impact of the environment on spatial referencing systems. In the real world, cross-linguistic variation has been proposed to interact with the local environment (such as mountains or rivers), but also other sociocultural variables. My experiments were able to show how participants’ spatial referencing strategies were systematically affected by the topographic environment while they had to solve a collaborative task in environments with different affordances. Beyond that, I will discuss the further potential of VR for similar studies into linguistic diversity, cultural evolution and the digital humanities more generally and provide some pointers on how to set up such experiments using modern hardware and 3D engines.
Jonas Nölle (jonas.noelle@glasgow.ac.uk) is a cognitive scientist and evolutionary linguist. He is currently a research fellow at the Social Psychophysics lab at the University of Glasgow in Scotland, where he investigates multimodal communication and specifically the role of facial expressions in social interactions as part of Rachael Jack's Facesyntax ERC project. Previously he worked with the Interacting Minds Centre at Aarhus University and completed a PhD at the Centre for Language Evolution in Edinburgh, where he pioneered the use of interactive VR experiments to study language and communication in realistic and immersive laboratory settings. His research focuses on the study of communication in task-oriented face-to-face interaction, the emergence and evolution of novel communication systems, and the interaction of culture, cognition and the environment in shaping conventions and human behaviour.
Link to the recording: https://youtu.be/gC0wohPv6VE
26 August 9:30 - 11:00
We are increasingly using computational text analysis to explore humanities and social science questions. A common and crucial step, then, is measuring social or cultural concepts using computational methods. Recent advances in Natural Language Processing (NLP) are promising. However, there’s increasing evidence that NLP systems are brittle and that results are sensitive to small design choices. In this talk I will discuss two recent studies looking at these topics. I’ll focus on the measurement of biases in NLP and on a recently developed test suite to interrogate hate speech detection systems.
Dong Nguyen is an assistant professor at the department of Information and Computing Sciences at Utrecht University (NL). Previously she was a research fellow at the Alan Turing Institute (UK). She holds a PhD from the University of Twente (NL) and a master’s degree from Carnegie Mellon University (USA). She is interested in developing NLP methods to explore questions from the social sciences and the humanities. At Utrecht she leads the NLP and Society Lab.
Link to the recording: https://youtu.be/nWIHoG5OQqQ
26 August 14:00-15:00
In this lecture I will introduce TEXTA and give an overview of a project we did last year for the National Library of Estonia. We will describe the process of developing our solution for automatically proposing subject indices (keywords) for books.
Moderator Peeter Tinits
Link to the recording: https://youtu.be/W2jyfAySWyA