It is not the case that The inference from 'a polynomial algorithm exists for one NP-complete problem' to 'polynomial algorithms exist for all NP problems' treats existence as a transferable property, but Kripkean possible-worlds semantics demands we specify in which worlds and under which interpretations such algorithms exist.
?Set your confidence on the premises below to see your aggregate.