Replying to
@amapanda@en.osm.town @kevin@en.osm.town yeah, that's annoying, but the discussion of gender modelling have proved to be super complex because people can't seem to agree on anything when it come to gender, tend to take it very personally on top of a already combative and toxic culture. It is quite hard that a group of people who want to break boundaries also resist neat classification. There is a compilation of most of the discussions on https://www.wikidata.org/wiki/Wikidata:WikiProject_LGBT/gender and in the end, I think the choice was made to respect gender diversity by being precise, but this requires people making queries to be aware of the said diversity by being equally precise (back then the trick was to use this syntax: https://www.w3.org/TR/sparql11-query/#propertypath-arbitrary-length ). However, I also see that something got changed around that in the past, because neither Q2449503 nor Q6581072 had proper subclassing as of 1h ago while I am sure it was the case in the past. I added the right entries to be able to use P21 subclassing, so it should be doable now.