Safety, Sensitivity and the Value of Knowledge

Epistemologists who wish to provide a modal analysis of knowledge are divided over whether the safety or the sensitivity condition offers a correct necessary condition for knowledge.  The two are usually parsed in (something like) the following way:

Safety: S’s belief p is safe iff in nearly all (if not all) nearby possible worlds w in which S forms a belief p using the same belief-forming methods as the actual world, p is true.

Sensitivity:  S’s belief p is sensitive iff in the closest possible world in which S forms her belief using the same belief-forming methods as the actual world and in which p is false, S does not believe p.

Sosa suggested the following scenario, which has been influential in encouraging philosophers to favour safety over sensitivity:

On my way to the elevator I release a trash bag down the chute from my high rise condo.  Presumably I know my bag will soon be in the basement. But what if, having been released, it still (incredibly) were not to arrive there? That presumably would be because it had been snagged somehow in the chute on the way down (an incredibly rare occurrence), or some such happenstance. But none such could affect my predictive belief as I release it, so I would still predict that the bag would soon arrive in the basement. My belief seems not to be sensitive, therefore, but constitutes knowledge anyhow, and can correctly be said to do so. [‘How to Defeat Opposition to Moore’: 145-6]

Now, I’m not convinced that the condo owner really has knowledge in this case—as opposed to a true, justified belief, or knowledge of some closely related proposition, such as that it is highly probable that the bag is in the basement—and I also think that examples with the same structure as Sosa’s offer some evidence against the safety condition and in favour of sensitivity.

For instance, I once went to meet my supervisor on the fifth floor of our building, only to discover that his office was not there: he had moved to the sixth floor.  Two things strike me about this.  Firstly, it seems to me perfectly reasonable to hold that I didn’t have knowledge that he was on the fifth floor, even when this was true and I was justified (and safe) in believing it.  I was not properly connected to the facts: when the justified, true (and safe) belief became a justified, false belief, I was oblivious to the change.  Secondly, and relatedly, it is often thought that we value knowledge resides (at least in part) to its connection to felicitous action.  A mere true belief is, all things being equal, less valuable than knowledge because the vagaries of the world ensure that it has a tendency to turn into a false belief.  Known propositions are the sort that allow us to successfully navigate our environment, and false beliefs will (in ordinary circumstances) lead us to behave infelicitously (e.g. by going to the fifth rather than the sixth floor).  The thought here is that there are many scenarios akin to the one above, and, hence, sensitivity seems to be a necessary condition for the sort of belief that allows a person to successfully navigate her environment.