Regular readers will recall I do not think so-called Gettier problems have any force in epistemology. The “problems” only crop up because people make two mistakes: forgetting the difference between ontology and epistemology, i.e. what is and what we know about what is, and forgetting the difference between necessary and local truths. A necessary truth is the law of non-contradiction, which we know is true via a form of induction. A local truth is “George wears a hat” given “All Martians wears hats and George is Martian.” We know this is true via a form of induction, too, helped by deduction.
Best to think of Gettier “problems” through example. This is quoted from Duncan Pritchard’s paper “Virtue epistemology and epistemic luck, revisited” (which I learned of from George Shiber) in which Pritchard quotes Linda Zagzebski (ellipses original):
Suppose that Mary has good eyesight, but it is not perfect. It is good enough to allow her to identify her husband sitting in his usual chair in the living room from a distance of fifteen feet in somewhat dim light […] Of course, her faculties may not be functioning perfectly, but they are functioning well enough that if she goes on to form the belief My husband is sitting in the living room, her belief has enough warrant to constitute knowledge when true and we can assume that it is almost always true […] Suppose Mary simply misidentifies the chair sitter, who is, we’ll suppose, her husband’s brother, who looks very much like him […] We can now easily amend the case as a Gettier example. Mary’s husband could be sitting on the other side of the room, unseen by her. (Zagzebski 1996, 285-7, emphasis in the original)
We’re meant to be suspicious that Mary has true knowledge because she flubbed and misidentified the chair sitter, yet the proposition of interest “My husband is sitting in the living room” is still true (it is said). But you can see the mix-ups I mentioned. Mary’s truth is a local one, conditional on the premise “I see a man I identify as my husband sitting in his usual chair”. To us, from our perspective, we see Mary’s premise is wrong. But to us, from our perspective, the proposition “Mary’s husband is sitting in the living room” is also locally true conditional on another premise (un-assumed to Mary but assumed by us) “Mary’s husband is sitting in that chair over there.”
The proposition “My/Mary’s husband is sitting in the living room” is not a necessary truth, which is to say, it is contingent. And all our knowledge of continent propositions is in some way based on local truths, which means Gettier “problems” are no problem at all. Unless we’re meant to talk of knowledge only with respect to necessary truths, which is, I suppose, a definition of sorts. But then we’re in a world of hurt because in order to speak of what is, we first have to use our senses, which are contingent.
We have to also keep in mind our goal: is it ontological or epistemological? If it is only to say or guess or bet or decide whether or not the proposition of interest is true, then in the above scenario Mary wins and so do we. But if it is to remove as much as possible of the contingency of the premises for the proposition, then Mary loses and we win. Why? Because it is a local falsity that “Mary’s husband is sitting in his usual chair” given “Mary’s husband is sitting in that chair over there”. We haven’t removed all contingency because it is not necessarily true that Mary’s husband should be anywhere.
Enter epistemic luck. You’re filling out this quiz: “Progressivism is (A) Harmful, (B) Stupid, (C) Degrading, or (D) All of the above.” Suppose you went to a standard government-sponsored school, complete will all the latest educational theory, then you will not know the answer to the question, so you guess. Eenie, meanie, minee, moe…you pick D.
Now D is the right answer. And we, from our perspective, know D is the right answer conditional on the things progressives do, along with assorted other facts. But, lo, you had no idea D was correct. You were just lucky; you have hit upon knowledge via no effort. So once again, it is hinted, epistemology is in trouble.
Again, the goal of the analysis is mixed up. If the goal is to get the right answer, then you win. You don’t always win in tests that you haven’t studied for, but this time you do win. But you don’t, before the answer is revealed, know D is right. You have made a decision, and decisions aren’t probability (or knowledge). You only have the premise “There are four answers, one of which has to be true.” From that you reason, “D is true” has probability 1/4. (Or maybe you add a premise about tricky “All of the above” questions, but the point is the same.)
If the goal of the analysis was to examine your premises in relation to the decision you made, then we realize you did not have knowledge of D. To have knowledge “D is the right answer” your premises would have had to have been “The things progressives do, along with assorted other facts”. You didn’t know that, so you lose. This mix-up, incidentally, is why multiple-choice questions are an unreliable or at least a weak form of knowledge assessment. But they sure are easy to grade.
This example doesn’t quite capture all what is meant by epistemic luck. Here’s another example from Pritchard (also modified from another):
[C]onsider an agent who is forming a belief about what the temperature of the room is by looking at a malfunctioning thermometer, the reading on which is randomly fluctuating within a certain temperature band. Suppose further, however, that this way of forming beliefs about the temperature of the room is entirely reliable because, as it happens, there is someone hidden in the room—a ‘helper’—who controls the temperature of the room and who ensures that every time our protagonist goes to check the thermometer, the reading it gives is correct
Philosophers like say “agent” when they mean “man”. Anyway, the “agent”‘s “beliefs do not amount to knowledge” because “they do not get to the truth via their cognitive characters, but rather due to luck”. The same goal mix-up happens here. If we checking if the man had the right temperature, he wins; else if we’re checking if he had the right premises, he loses. Simple as that.
We still need to discuss the awful terminology which is “possible worlds.”