Edmund Gettier: “Is Justified True Belief Knowledge?”
Since Plato, it had generally been agreed that there are
three criteria of propositional knowledge(that is, knowledge that something), individually necessary and jointly sufficient:
S knows that p iff
1. p is true (truth criterion)
2. S believes that p (belief criterion)
3. S is justified in believing that p (justification criterion)
Explanations of each:
Belief criterion:
Why necessary: You can’t know
something without believing it.
Why not individually sufficient: Not everything we believe is knowledge. I can (and do) believe many false things.
What is belief? Not an easy question to answer. It’s not simply a mental state like pain, because I can believe something without being aware of it (you believe that 2+2=4 all the time but only occasionally think of it)
Truth criterion:
Why necessary: you can’t know something false.
Why not sufficient (along with belief): you can believe something true without knowing it. Suppose I believe that I’ve won the lottery before they announce the results. Turns out I do win. Did I know it? Clearly not, so belief and truth are not enough.
What is truth? An even more vexed issue – there are three main theories:
Correspondence: Truth is in the correspondence of some proposition with the world.
Coherence: truth is a property of systems – that is, whether or not a proposition is true is decided by how well it coheres with other propositions in the system of which it is part.
Pragmatic: truth of a belief is dependent on how useful that belief is in reaching certain goals.
Justification criterion:
Why necessary: (see above explanation of why the other two are not sufficient)
Why not sufficient (with belief): I can be justified in holding a belief without it being true. I can be very justified in believing that if I drop a brick out of the window it will fall to the ground, but it’s still possible that a big updraft will lift it up. Suppose that happens. In that case I didn’t know that it would fall, even if I was justified in believing it.
What is justification? Chisholm, for example, defines it in terms of conforming to epistemic rules. We set out guidelines that tend to produce true beliefs, and then we can claim that a belief is justified if it conforms to those.
Gettier does not question that each criterion is necessary. What he purports to show is that they are not jointly sufficient. That is, that we can justifiably believe the true proposition p but not know p.
Two assumptions are required to prove this:
A1. It is possible for someone to be justified in believing something false (as argued above)
A2. If S is justified in believing p and p entails q, then S is justified in believing q.
The cases:
1. Smith
and Jones and the Job:
Smith believes (justifiably) of Jones both
a) That he will get the job
b) That
he has 10 coins in his pocket
As a result, Smith believes
c) The man who has 10 coins in his pocket will get the job
(c) is true, but not because Jones is that man, but rather because Smith himself has both 10 coins and the job in hand. By A1, Smith is justified in believing both (a) and (b) [and thus their conjuction (a) & (b)], even though they are false, and by A2 he is thereby justified in believing (c). BUT HE DOES NOT KNOW IT.
2. Smith,
Jones’s car, and the location of Brown:
Smith believes (justifiably)
a) That
Jones owns a Ford
As a result, Smith infers
b) Either
Jones owns a Ford or Brown is in
even though he has no idea that Brown is in
Turns out that (a) is false (Jones
is renting) but (b) is true simply because by
sheer coincidence, Brown is in
These are both cases where Jones has justified belief in a true
proposition (1(c), 2(b)), but we would not want to say that he has knowledge of that proposition. Thus, even if truth, belief and justification
are individually necessary for knowledge, they are not jointly sufficient.
Responses to Gettier:
1. find some means to show that the counterexamples do not work
2. accept the counterexamples and search for a fourth criterion
3. accept the counterexamples and alter the three criteria to
fix the problem.