Sensitivity from others

The sensitivity condition on knowledge emerges out of a simple but highly attractive idea: whether S’s belief that p amounts to knowledge depends on whether S would have so believed had it been false that p. On a (simplified but) standard possible-worlds framework, to ask whether S would have so believed is to ask whether S does believe that p in the nearest possible world in which not-p. To a first approximation, then, the sensitivity condition on knowledge tells us that, in order to determine whether S knows that p (in the actual world), we must consider what is going on in the nearest world in which not-p. As will be familiar, this is merely a first approximation of a sensitivity account: it must be modified to deal with a problem Nozick himself anticipated (Nozick 1981). His now-familiar example involved a grandmother (Granny) who comes to believe that her grandchild (Sonny) is healthy by seeing him in a hale state, under conditions where had Sonny been unhealthy his parents would have hidden him from Granny, telling her instead that Sonny was healthy (to protect her from news that would upset her). Nozick’s point was that Granny’s belief that Sonny is healthy, formed through seeing Sonny in a hale state, is knowledge, even though in the nearest world in which Sonny is not healthy Granny still believes him to be healthy (having been told as much). This led Nozick to relativize assessments of sensitivity to the method involved in the formation of the belief under assessment.