4 February 2024
A RESEARCH PROJECT
I propose to quantitatively assess the global conflict levels over time by conducting sentiment analysis on historical news archives. The underlying hypothesis is that public discourse varies in intensity, similar to how the temperature of a gas changes. This "temperature" of societal debates could reflect but also influence global events. My job exposed me to many situations in which groups of people are driven by different interests and aims. Often, I observed how language – for example the jargon of economic organisations or political militants – shapes common ground and sets the limits of cognitive domains, paving the way for friendly or adversarial relationships. Working in a field that is often the mean through which the different parts of the society talk to each other, I developed an interest in exploring the consequences of language variations on decision making and behaviour.
The specific inspiration for this project was the European Union's rapid response to the occupation of Ukraine, which I hypothesise was influenced by a pre-existing high intensity in global discourse. This intensity, or "temperature," might have been escalated by various crises over recent years, such as the Greek financial crisis, migration issues, Brexit, and the COVID-19 pandemic. While the world is witnessing a new arms race, the primary aim of the project is to develop a predictive model to anticipate future global inclination to conflict. This involves answering questions like: How does the current level of public discourse compare to past decades? Can we graphically represent the "temperature" of public debate from over the years? Is there a correlation between peaks in this graph and major global conflicts?
As for data, I plan to make use of the Reuters news archive, which provides a comprehensive global news coverage dating back to the 19th century. The archive's straightforward journalistic style, with emotional language mostly found in direct quotations, offers a solid base for robust sentiment analysis. This analysis will focus on identifying shifts in the tone and sentiment of news content, with the aim of mapping the intensity of global discourse over time.
This project not only holds potential academic value in the fields of natural language processing and conflict studies but also practical implications in predicting and understanding global trends and conflicts. The findings could offer insights into how historical events have shaped current global discourse and potentially help predicting future global challenges.