We scored statements from four datasets: Uber regulation in Taiwan and three climate policy topics from Austria. Each had two opposing opinion groups; bridging = the agreement rate from the less supportive one.
We saw no significant association between bridging and ambiguity.
Read about the rest of our findings in the paper: aclanthology.org/2026.emnlp-…
And if you're at EMNLP, come to my presentation tomorrow morning during session 6 on Sentiment Analysis, Stylistic Analysis, and Argument Mining!