Hi, I participated into the edits that ended up with this mess, so I plead guilty /o\.
I’d say the problem is that we don’t really have a model at all. At best, there is some WikiProject that try to impose some rules they decided, with the notion of concensus decided by the people of the project. Some WikiProjects exists for some domains but are inactive and/or inefficient to impose rules. Apart from that there is constraints, that are decided by the sums of individual edits, for example, and occasionally discussions on project chat or other venue like the french «bistro». In my experience RfCs on the model does not usually reach a conclusion. In this case there is a WikiProject Award, that sets up some rule : https://www.wikidata.org
, but … I’m not sure how those rules came up and the rationale behind it are not explained.