CLJun 8, 2018

ChangeMyView Through Concessions: Do Concessions Increase Persuasion?

arXiv:1806.03223v121 citations
Originality Incremental advance
AI Analysis

This work addresses a gap in discourse studies by empirically evaluating a theoretical claim about persuasion, though the results are incremental and specific to online forums.

The study tested whether argumentative concessions increase persuasion by analyzing their distribution in persuasive vs. non-persuasive comments from the ChangeMyView subreddit, finding that such concessions are almost equally likely in both types of arguments, contradicting theoretical assumptions.

In discourse studies concessions are considered among those argumentative strategies that increase persuasion. We aim to empirically test this hypothesis by calculating the distribution of argumentative concessions in persuasive vs. non-persuasive comments from the ChangeMyView subreddit. This constitutes a challenging task since concessions are not always part of an argument. Drawing from a theoretically-informed typology of concessions, we conduct an annotation task to label a set of polysemous lexical markers as introducing an argumentative concession or not and we observe their distribution in threads that achieved and did not achieve persuasion. For the annotation, we used both expert and novice annotators. With the ultimate goal of conducting the study on large datasets, we present a self-training method to automatically identify argumentative concessions using linguistically motivated features. We achieve a moderate F1 of 57.4% on the development set and 46.0% on the test set via the self-training method. These results are comparable to state of the art results on similar tasks of identifying explicit discourse connective types from the Penn Discourse Treebank. Our findings from the manual labeling and the classification experiments indicate that the type of argumentative concessions we investigated is almost equally likely to be used in winning and losing arguments from the ChangeMyView dataset. While this result seems to contradict theoretical assumptions, we provide some reasons for this discrepancy related to the ChangeMyView subreddit.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes