My research at McGill has focused on how to use language models to understand political language, especially in international diplomacy.

Pouliot, Vincent and Scott Robert Patterson. “Domesticating Wealth Inequality.” Global Studies Quarterly (2024).

  • This paper features a discourse analysis of “wealth inequality” in diplomatic debates at the United Nations General Assembly that spans from 1971 - 2018. In the paper, we use a word embedding approach to demonstrate how discourse on wealth inequality has become “domesticated” in assembly debates since the late 1970s. We think about domestication here in two senses - both to tame and to make a domesticate (as opposed to international). Please reach out if you would like to access the replication data.
  • This project was completed with open source tools and data. Replication materials can be found here.

Patterson, Scott Robert and Vincent Pouliot. “Placing Machine Learning into the Hermeneutic Circle.” Journal of International Relations and Development (2024).

  • This is the first paper of my dissertation. The paper has two goals. The first is to explore the relationship between machine learning approaches to text analysis and traditionally interpretive methods with the aim of unconvering where they can be complimentary. As a second goal, it also introduces a technique for discovering “turning points” in chronologically ordered text data. Feel free to reach out if you are having trouble accessing this piece.
  • This project was also completed with open source tools and data. Replication materials can be found here

Working Projects

  • Podcasts as Data: Mapping the Canadian podcast ecosystem.
  • LLMs as Classification Tools: Using fine-tuned GPT models to standardize Federal Election Commission data.