Target Based Speech Act Classification in Political Campaign Text
This work addresses the challenge of understanding pragmatics in political discourse for researchers in computational linguistics and political science, though it appears incremental as it builds on existing methods with domain-specific adaptations.
The authors tackled the problem of classifying speech acts and their targets in political campaign texts by proposing a new annotation schema and creating a novel annotated corpus from the 2016 Australian election. They evaluated various techniques, including contextualized word representations and semi-supervised learning, to model these as sequential classification tasks.
We study pragmatics in political campaign text, through analysis of speech acts and the target of each utterance. We propose a new annotation schema incorporating domain-specific speech acts, such as commissive-action, and present a novel annotated corpus of media releases and speech transcripts from the 2016 Australian election cycle. We show how speech acts and target referents can be modeled as sequential classification, and evaluate several techniques, exploiting contextualized word representations, semi-supervised learning, task dependencies and speaker meta-data.