[EnderIII]

My problem with Nozick:

Condition (4) of Nozick seems potentially problematic in that it is too stringent of a restriction. I am thinking of cases in which condition (4) is not satisfied, but I still want to maintain that X knows something.

This will happen in cases where P is still true at the nearest possible world, but X no longer believes that P. An example will hopefully illuminate what I mean by this.

P is the proposition that “Emily cooked eggs this morning”.

In the actual world this is true (satisfies condition (1).

Sam believes that P. (satisfies condition (2).

If P weren’t true, then Sam would not believe that P. (satisfies condition (3).

When Emily was getting out of bed this morning, Sam groggily asked her what she was doing to which she replied, “ I’m going to make some eggs”. She then proceeded to cook eggs and Sam got out of bed in time to eat half of the eggs with her.

n the nearest possible world that Emily did not cook eggs this morning, she decided that a few minutes more of sleep were better than making breakfast and had cereal instead of eggs. So Sam has no reason to believe that P, if Emily did not actually cook eggs this morning. So the counterfactual necessary for condition (3) holds true.

But, in some cases the counterfactual:

If P were true, then Sam would believe that P.

Might be false, but I still want to maintain that Sam knows P.

In the nearest possible world at which P is still true, instead of groggily asking Emily what she is doing, Sam does not wake up. Emily then gets the idea that she will surprise Sam with breakfast in bed. So she cooks eggs, making P true, but because she is so excited about surprising Sam, she clumsily spills the eggs on the floor. She cleans up the mess so there is no evidence of eggs having been made, so she won’t be embarrassed. Sam now does not believe that P, even though P is true. This example shows that condition (4) is false for this case.

It seems to me to be too strict and limiting to have a definition of knowledge that precludes us from saying that Sam knows P, because condition (4) fails in this manner. I think I want a proper account of knowledge to include that Sam knows that P. There also is an issue of accessibility, in that we might not know when X knows P because we do not have proper information about the nearest possible world at which P is true. Not knowing when X knows may be problematic.