Uhm, no.
1., you're still missing the point, that is whether being a fellow of the Linnean Society constitutes a type of membership or not. That's a simple question, for which there's a simple way to answer: Go to the society's website, and read that "there are three levels of membership: Fellows, Associates and Student Associates." Pretty straightforward, isn't it? You haven't given any rationale that could overwrite that statement.
2., by using the qualifier "subject has role" (P2868), it's absolutely clear that the date (year) given marks the beginning of the fellowship, not of a regular membership that may have existed before. Plus, if you claim that Victoria Braithwaite must have been member before she could have been elected fellow, then why do you want to delete that membership from the data set?
3., I explicitly cited your statement that "the date when someone becomes a fellow is NOT the date when he becomes a fellow", yet you still seem not to realize that this makes no sense, obviously. You may consider this to be a small typo, but it pretty much undercuts your whole point if you mix up the very two things (fellowship and membership) you want to treat separately, and then don't even realize it even if I explicitly point it out for you.