Retraction notices are expected to be transparent about entities accountable for retractions and their reasons for retraction. No previous research on retraction notices has investigated accountable entities other than authors of retracted publications and their reasons for retraction from a cross-disciplinary and a diachronic perspective. Drawing on a dataset of 7650 unique retraction notices published before 2020 and indexed in the Web of Science Core Collection, this study identified 457 retraction notices that held four types of non-author entities accountable. Journal authorities (i.e., editors and publishers) were found responsible for retractions in 62.14% of the retraction notices, followed by unidentifiable entities (26.91%), third parties (6.13%), and dual entities (i.e., both authors and journal authorities, 4.81%). The three most frequent reasons for retraction were republication of valid publications (36.76%), unspecified reasons (26.91%), and publication of unpublishable work (17.07%).Accountable non-author entities were more identifiable in the retraction notices published between 2010 and 2019 than in those published before 2010. Retraction notices in soft disciplines were more likely than those in hard disciplines to identify journal authorities as accountable entities.These findings offer implications for handling retractions more effectively and efficiently.