Reasoning Together: Network Methods for Political Talk and Normative Reasoning
Dissertation Defense
Past Event
Sarah Shugars
Network Science PhD Candidate
Apr 10, 2020
Watch video
2:00 pm
177 Huntington Ave
Remote Participation via Zoom
Join Talk (Zoom)
11th floor

Democracy is fundamentally an interactive endeavor which relies upon the collaborative reasoning of its citizens. In formal governing bodies, online interactions, and informal conversations, people engage in political talk in which they use reasons to convince, persuade or justify. These interaction form a rich deliberative system which spans every level of political life, providing myriad venues for both everyday citizens and elites to clarify their thinking, share relevant information, persuade and be persuaded, and have a voice in determining what topics are important and how those topics should be discussed. At its best, such a deliberative system allows every citizen the power to both develop an informed opinion and play a role in developing the opinions of others. Of course, there is good reason to believe the ideal deliberative system has not been achieved — systemic power disparities, limits to human cognitive capacity, and tendencies towards blind partisanship all make a functional deliberative system appear desperately far away. However, if we are to accept as an ideal that citizens in a democracy should have their voices heard — that they should talk, reason, and work with others around matters of common concern — then we have an obligation to work towards that democratic ideal. This dissertation therefore aims to better understand several key elements of the deliberative system. Can a process of sharing reasons and ideas bring a group to better decisions? Do everyday citizen express themselves in individually distinctive ways or simply repeat party talking points? What drives people to engage in ongoing, repeated interactions around political topics? How can we measure phenomena related to the deliberative system in order to design and test possible interventions? This dissertation aims to address all these questions through a focus on three elements of the deliberative system: the exchange of arguments, the structure of individual-level reasoning, and the dynamics of ongoing conversation. Each of these elements has unique networked dynamics — i.e., they can be understood most fully by considering interconnections along a range of dimensions. Social networks influence the people and ideas we come into contact with, semantic networks influence the very words which are exchanged, conversation itself has a network structure as participants respond to each other’s comments, and the very notion of a "deliberative system" implies a richly interconnected symphony of dialogue. In short, the deliberative system cannot be fully interrogated without network methods capable of considering the interconnectedness of reasoning and engagement. Therefore, each chapter further aims to build methods and models capable of strengthening our understanding of these richly interconnected systems. Specifically, in this dissertation, I examine the effects of cognitive capacity and bias through an agent-based model, finding that agents can discover good solutions if they are willing to accept good information from others. I further explore the semantic and cognitive connections inherent in expressing a political view, finding that individuals express themselves in individually distinctive ways which are correlated with known personality traits and not merely a result of ideological talking points. Finally, I delve into the drivers of engagement in real world, online conversations, finding that these conversations are surprisingly predictable, often driven by emotional and topical salience, and occur over a range of ideological divides. Taken together, this work suggests that a functional deliberative system may indeed be obtainable, and it provides conceptual frameworks and methodological tools for moving towards such an ideal.

Dissertation Committee Members:

Nick Beauchamp (co-chair)
Assistant Professor of Political Science
Northeastern University

David Lazer (co-chair)
University Distinguished Professor of Political Science and Computer and Information Science
Northeastern University

Lu Wang
Assistant Professor of Computer Science
Northeastern University

Peter Levine
Academic Dean and Lincoln Filene Professor of Citizenship & Public Affairs
Tufts University

About the speaker
Sarah is a doctoral candidate working with Professors David Lazer and Nick Beauchamp. Her research focuses on American political behavior and develops new methods for natural language processing, network analysis, and machine learning. Her published work includes multiple studies of engagement in political conversations online, an examination of persuasion and argument strength in formal political debates, and a framework for building core civic skills through interactive games. Her substantive focus on political communication, public opinion, and civic participation is enriched by her methodological training; allowing her to identify methodological gaps and develop text and network methods needed to pursue research questions within these domains.