Search for contacts, projects,
courses and publications

Mining argumentative patterns in context. A large scale corpus study of Earnings Conference Calls of listed companies

People

 

Rocci A.

(Responsible)

D'Agostino G.

(Collaborator)

Lucchini C.

(Collaborator)

Yaskorska-Shah O.

(Collaborator)

External participants

Reed Chris

(Third-party co-responsible)

Moslemnejad Somaye

(Third-party collaborator)

Abstract

The proposed research is a large-scale study of argumentative patterns in a corpus of quarterly Earnings Conference Calls (ECCs), a key dialogical genre in the financial communication of listed companies. Traditionally, the study of argumentation in context has relied on the analytical reconstruction of individual discourses examined in relation to descriptions of activity types in order to outline how the goals, incentives and procedures of activities constrain the issues and the material and procedural starting points of argumentative discussions. Recently, researchers have advocated a shift towards larger corpus studies not only to test and refine hypotheses on the contextual constraints on argumentation, but also to map inherently more complex networks of arguments in multi-party discussions that shape broader debates in society. This shift is also necessary in order to start addressing the effects of argumentation on the context itself. Corpus research on argumentation, however, requires the development of theory-based annotation schemes and time-consuming annotation by trained analysts, which risks being an insurmountable bottleneck. Furthermore, it requires quantitative analytics to relate analyzed arguments to contextual parameters and, if the effects of argumentation are to be addressed, the possibility to measure change in contextual parameters at appropriate points in time. Recourse to Argumentation Mining and related Argumentation Analytics appears promising, but tools and techniques developed in this growing field have so far seen limited application to research on discourse in context. The project seeks to demonstrate how these challenges can be met with an Argumentation Mining approach designed to investigate the interdependency of argumentation and activity type through the notion of argumentative pattern (AP), which refers to significant constellations of argumentative moves whose occurrence can be explained in view of the goals and rules of the activity type.

Additional information

Start date
01.09.2021
End date
31.08.2025
Duration
48 Months
Funding sources
SNSF
External partners
University of Dundee
Status
Active
Category
Swiss National Science Foundation / Project Funding / Humanities and social sciences (Division I)