EzekielRaiden
Follower of the Way
Okay. Now: How do we then make sense of a claim such as "I reject F because W' is unrealistic."?In fiction, it's true that Holmes is a consulting detective (a fact in W'). It's not true that Holmes is a consulting detective in W (the real world). Of these conflicting facts, W' is privileged so long as we're entertaining fiction. It's not ruled out that there is a consulting detective called Sherlock Holmes in W' just because there isn't one in W.
This becomes interesting when it's asked "What else is true in W'?" Upthread I gave an example. Could Sherlock Holmes have taken passage to New Zealand or Neptune? It's never established in the fictional world of Holmes that he takes passage to either, but one seems like a more "realistic" possibility because in W in the 19th century, consulting detectives were able to take passage to New Zealand, but not Neptune. By this and an abundance of examples like it, it appears evident that W is privileged only where W' is silent.
Speculating on the problems you identify, I can think of a few explanations consistent with the idea of cognitive models. One is where player A assumes F to be an established fact in W' whilst player B does not. This would be acute where A has a slightly different imagined world in mind than B... something that is readily possible. B will foreseeably argue from W (because to them, W' is silent) while A will expect what they take to have been established in W' ought to prevail. You can see the obvious permutations.
Another explanation is where W' has until now adhered closely to a genre (this relates to the "accessibility" quality I mentioned upthread.) Holmes is a consulting detective, but not an elf. And no dragons appear in Holmes' Europe. Player A might develop expectations along the lines of an expected genre, making them resistant to propositions that some F is "realistic" in W' because those propositions would not be "realistic" in other possible worlds fitting the genre, which they assumed W' to be a member of. Again, I think you will see the permutations.
So where W' is silent or where it is assumed to match other worlds but turns out not to, one would predict disagreements about what to count as realistic. I suppose that'd be most noticeable at the moment an F that formerly was not established in W' became established.
Because that sort of claim is precisely why, for example, a player might balk at Bastilles & Basilisks 7th Edition, because it contains elements that are "unrealistic". Or why someone might assert--as has happened in this very thread--that it is problematical ("unrealistic" is known to be a flawed word, so some variation like "lacking verisimilitude" etc.) to have a mechanic which resolves an action by having failure indicate that a bad situation has occurred as a result of the undertaken effort, for one reason or another.
In such things, W' itself is specifically being called out as wrong, bad, incorrect, unacceptable, etc., etc., specifically because it conflicts with W. Even when the actual content of the real world is more similar to the thing described in W' (hence my repeated reference back to the lockpicking example).