In the epistemology class I’m teaching, we’ve moved on from Bergmann’s book to Pritchard’s book Epistemic Luck.
Here’s a principle that Pritchard ultimately is going to defend.
Safety Principle
If S knows P, then S believes P in most nearby worlds where P is true
This is not to be confused with the Nozickean/Dretskean tracking/sensitivity type principles (e.g., where S knows P just in case S believes P in the nearest possible world where P is true and does not believe P in the nearest possible world where P is false.)
Pritchard introduces this principle and notes that we’re going to need to refine it in order to avoid some problems. (For example, it seems as though Kvanvig has a modfied Barn Country scenario that would apply to this principle) – so I haven’t thought carefully yet about whether Pritchard’s refinements will handle the two cases I’m about to present. Nevertheless, I want to get them down before I forget them.
Bear Hating Neuroscientists
Suppose you have a band of bear hating neuroscientists who want to wipe out all beliefs in the existence of bears. They put a chemical in the world’s water supply for 7 days. Through a series of very bizzare coincidences you never actually drink the water from the water supply, but there are MANY, MANY nearby worlds where you do.
You can fill in the details of the scenario however you like. Suppose you were accidentally locked in the basement for two days. You nearly got out on several occasions – so there are many nearby worlds where you are not locked in the basement. You have a case of bottled water which you very nearly didn’t get. You get locked in the basement again for a few days and you very nearly weren’t.
A week goes by, and you’re one of the few people in the world who still has the capacity to believe in bears. You walk out and see a bear. It seems that you know there is a bear in front of you, however, it’s not true that in most of the nearby worlds you believe that proposition when it’s true – given all of the many, many ways in which you narrowly escaped ingesting the chemical.
Lucky to Have Looked
Suppose you have attention deficit disorder, and never can focus on any one thing for a long period of time – you just happen to glance up at the sky when there is a shooting star. I imagine we can, given your ADD and some other facts about you, cook this case up so that there are many nearby worlds where it’s true that there is a shooting star there – but you don’t believe it. It seems like you still know that there is a shooting star.
Again, I quickly read through chapter 6 a while back, and I know he says earlier in the book that Safety must be tweaked. I just want to get these down, so we can assess whether his proposed tweaks successfully avoid these counterexamples.
I have a question about how the chemical in the bear case affects people. when a person drinks the water, does he lose cease to have the concept of a bear? If so, then it may be that he can still have bear thoughts. If confronted with a bear, he might think “holy crap, there’s one of those kinds things in front of me!”. If we think that complex demonstratives directly refer to the things that they successfully demonstrate, then the content of that thought will be pretty much the same as the content of “holy crap, there’s a bear in front of me!”. Moreover, if this is correct, then the lucky person who has avoided the chemical will, even in nearby worlds where he has drunk the chemical, have bear thoughts.
On the other hand, perhaps the chemical is supposed to make it so that a person who is confronted with a bear has no thoughts whatsoever about the thing in front of him. Perhaps, phenomenally, he would be just like a person who is not confronted with a bear. In this case, the person will in fact have a bear thought and fail to have any such thought in nearby worlds. But, now I worry about the possibility of the case. It seems that there has got to be some external condition that interrupts the causal connection between the presence of the bear and the person standing before the bear. It is a bit hard to see how an internal mechanism (such as a chemical) can interrupt such causal connections.
That being said, it might still be possible. Perhaps the chemical does not interrupt the causal connections outside of the brain, but rather inside the brain. Perhaps a person who has taken the chemical gets all the right kind of input, but the chemical stops the processing of that input into a thought. If this is the case, then I think perhaps the safety theorist can hold onto the spirit of his position by invoking a safety condition on the receipt of input rather than the thought that results from processing.
The shooting star case seems a bit better to me.