Theory of Knowledge
Fall 2006
Lecture Notes: The Gettier Problem

I. Review

The key ideas to get from what we’ve done so far: The Standard View and the TAK. Regarding the latter: objective theory of truth, justification amounts to objectively good reasons to believe (not moral value of believing).

II. The Gettier Examples

 A. Basic Idea

The basic idea is that Gettier thinks there are cases in which all three conditions of the TAK are met, but in which you don’t know the relevant proposition. He is arguing that the three conditions are not sufficient for knowledge, by giving counter-examples to the analysis.

B. Case 1—The Ten Coins Case

C. Case 3— The Sheep in the Field

D. Two Possible Responses

There are two ways you can respond to these examples:

(1) Deny that they are genuine counter-examples to the TAK. That is, you will want to show that Smith did not actually have a justified belief.

(2) Accept that they are genuine counter-examples to the TAK, and modify the TAK to take account of them.

We will start with response #1. In order to do that, we’ll want to get clear on the basic structure of the examples, which we’ll do next time.

E. The basic structure: start with some very strong reasons to believe something, that something turns out to be false, infer a truth. Note principles JF and JD.

III. Defending the TAK
●          The “Raise the standards” reply

●          Rejecting JF

●          Rejecting JD

If the standard view and SE are true, then none of these defenses works: Rejecting (JF) won’t work because for any belief and its evidence, there are a pair of cases in which the evidence is the same and in one case the belief is true (this is the typical case) and in the other (the unusual case) the belief is false. Given the standard view, the belief must be justified in the typical case. So it follows that is it is justified in the unusual case.

Pretty much the same thing shows that you can’t reject JD. [Further, as argued in the text, rejecting JD doesn’t entirely solve the Gettier problem since there are Gettier like cases in which you don’t reason through a falsehood.]

Raising the standards won’t help either. As long as there is any chance whatsoever that the belief is false, there will be possible examples in which it is false. There will also be possible examples in which it is true only by coincidence.

III. Modifying the TAK

There have been hundreds of attempts to add a fourth condition. We will not review all of them. Robert Shope’s book - something like The Analysis of Knowledge: A Decade of Research does. I picked two proposals for discussion.

●          No False Reasons - I think that this is pretty clear in the text. It’s too strong a condition. And, arguably, not enough to deal with all the Gettier like cases given the possibility of alternative routes to the final conclusion.

●          No Defeaters (Undefeated Justification) - I think that this is much harder to understand. Go through the steps in thinking about this

Before going through the proposal, a digression on if-then sentences (conditionals) will be helpful. [Notes for this are not available.]

1) Observation: In each Gettier case there’s a key truth that Smith is missing. Call this a “defeater.”

2) Idea: The existence of the defeater is the crucial fact. So, to have knowledge there must be nothing like that. That is, knowledge requires that there be no defeater.

3) Precise proposal: Definition of defeater: D defeats S’s justification for p iff D is a true proposition such that, if S were justified in believing D, then S would not be justified in believing p. Roughly and intuitively: imagine modifying things minimally so that S is justified in believing D. Then ask if, under those situations, S would be justified in believing p. If not, then, D is a defeater.

Revised definition of knowledge - JTB plus no defeater.

4) Problems: Key to appreciating this is not to think intuitively about the unexplicated idea of a defeater, but rather to look at the actual definition. Note: the original Gettier cases were cases that did satisfy the original 3 conditions but were not knowledge. So, the response was to add a 4^th condition on knowledge. The new examples show that the new condition requires too much - cases of knowledge that do not satisfy the new condition. In other words, given the definition, there are too many defeaters; there are defeaters even in cases of actual knowledge. The Radio case illustrates the fact that there can be truths such that, if Smith were justified in believing them, then he would be in a different situation, one in which he would not have the evidence for something he actually does know. The Grabit case illustrates the fact that there can be misleading defeaters - these are things that do not really block knowledge but are misleading truths such that, if Smith learned just them, he’d no longer be justified in his original belief.

            You might think that there’s some better definition of “defeater” that could be used that would avoid these problems. You might think that there is some way to revise the definition of knowledge, while sticking with the existing definition of defeater, that might be better. Those are legitimate things to write about (but hard).

The Modest Proposal at the end of the chapter is along the same general lines of these two ideas, but not subject to the same objections. It’s true that the idea of “essential dependence” is less than perfectly clear.