Automatically Identifying Ideology, Argumentation Schemes, and Discourse Structure in Parliamentary Text

Graeme Hirst, University of Toronto

A long-term goal of argument mining and opinion mining is to interpret opinionated texts, such as parliamentary debates, to determine not just the position that the speaker or writer takes on a particular issue but also their reasons for doing so, the arguments they adduce, and the ideological framework within which their position is formed. I will discuss our research on several facets of this problem. First, I will describe our work on automatically identifying the left/right ideological position of speakers in the Canadian and European Parliaments from the vocabulary that they use, including the confounds that we discovered. I will then turn to the broader, long-term goal of understanding speakers' argumentation and ideological frameworks, and describe some of our research on two tasks that are components of that problem -- identifying argumentation schemes in text and identifying the structure of a textual discourse in terms of Rhetorical Structure Theory -- and I will present some (very) preliminary results. I will also outline the tri-national interdisciplinary project Digging Into Linked Parliamentary Data, which brings together computational linguists, digital librarians, political scientists, and political historians, and which forms a context for our work.

The talk will incorporate work by my students and colleagues Yaroslav Riabinin, Jory Graham, Magali Boizot-Roche, Colin Morris, Vanessa Wei Feng, Nona Naderi, and Christopher Cochrane.