Verified authors shape X/Twitter discursive communities

In this study, we address the challenge of detecting ``discursive communities'' on X/Twitter by focusing on the role of verified users as the main content creators in online political debates. The analysis centers on three major Italian political events in 2022 - the Presidential election, a governmental crisis, and the general elections - occurring before the introduction of paid account verification. We propose and compare two novel methodologies, MonoDC and BiDC, which exploit, respectively, the retweet network among users and a similarity network based on shared audiences, while integrating a maximum entropy null model to filter out the inherent noise in online social networks. Our results demonstrate that leveraging verified users-considered as indicators of prestige and authority-leads to significantly clear community partitions that closely reflect the actual political affiliations, outperforming standard community detection algorithms applied to the entire retweet network. Moreover, the comparison of different methodologies and user sets suggests that the status conferred by the blue verification tick plays a dominant role in shaping online discourse, with important implications for platform governance, especially in light of the recent shift to paid verification.