Speaker: Windsor, L. (University of Memphis)
Date: 9 April 2019
SMA hosted a speaker session presented by Dr. Leah Windsor (University of Memphis) as a part of its SMA Minerva Speaker Series. During this presentation, Dr. Windsor discussed how computational discourse analysis can help the Department of Defense and intelligence community better understand political trends in the world, particularly in areas that currently lack coverage and are difficult to observe. She presented three research questions: 1) How does what we say reveal who we are?; 2) Is it what we say, or how we say it?; and 3) What biases influence our understanding of international relations? She then explained that one can learn an astonishing amount about individuals’ thoughts, beliefs, and preferences through their language and word choice. This allows us to identify patterns over time and among groups and gives us a better understanding of the world. She also stated that political forecasting has played a significant role in shaping this discipline. Next, Dr. Windsor discussed sources of bias within text. She explained that research about language and politics often has information and/or document processing deficiencies, which can affect conclusions and the generalizability of findings. However, she emphasized the importance of overcoming these deficiencies in order to increase our knowledge of groups that are difficult to observe, such as violent extremist organizations and authoritarian regimes. She used her team’s research on the Rwandan genocide as an example. Dr. Windsor emphasized the importance of using primary source documents and considering word choice when examining a document, as well as considering the language in which it is written. She also discussed how syntax, semantics, sentiment, and topics can affect the interpretation of a document. To conclude, Dr. Windsor described her Minerva 2018 capstone project, the functionality of Coh-MetrixML (a program which allows researchers to analyze syntactic and semantic features within a document in its original language), and noted the directions in which she hopes to take her future research.