What is knowledge?
We start the course by considering the question: What is knowledge? In particular: What is propositional knowledge?
Actually, we will consider a slightly different question: What is knowing? In particular: What is propositional knowing? Or: What is it to know a proposition? (To see the difference between these two questions, consider the difference between the questions: What is a gift? What is it to give?)
We are looking for an answer of the following form:
To know a proposition is to ...
Such an answer (whether or not it is true) we call an account, or a theory, or an analysis of (propositional) knowing.
Note that we are not trying to give a definition. To give a definition is to take a linguistic expression, e, something else, o, and stipulate that e is to refer to o. We are not trying to do that. In particular, we are not trying to stipulate what the word 'know' is to refer to. Rather, we are assuming that the word 'know' already refers to something, and we are trying to work out the nature of that thing.
If it turns out that 'know' refers to nothing, or to different things for different people, or to something that it really shouldn't, then perhaps we should try to give a definition instead. But that's not how we will start.
The JTB account
There is an account of knowing that was thought to be correct for thousands of years, up until about 1963. (See, for example, Plato's Theatetus (201c - 202d) and Meno (97e - 98a), Kant's Critique of Pure Reason, and Ayer's The Problem of Knowledge.)
We shall call it the JTB account:
JTB: To know a proposition is to have a justified true belief in it.
JTB is to be understood as an identity claim. It claims that one thing, the relation of knowing, is identical to another thing, the relation of having a justified true belief in. The expression 'to know a proposition' refers to the first relation; the expression 'to have a justified true belief in it' refers to the second relation; 'is' expresses identity. JTB has the form: relation1 = relation2.
(In an account of this form, relation1 is called the analysandum, and relation2 is called the analysans.)
A slightly weaker claim
There is a different kind of answer that is often given:
JTB': Necessarily: for all s and propositions p: s knows p if and only if s has a justified true belief in p.
What this says is that in every possible situation, the pairs of things that stand in the relation of knowing are the pairs of things that stand in the relation of having a justified true belief in. That is, in every possible situation, these two relations are co-extensive. That is, these two relations are necessarily co-extensive. JTB' is not an identity claim.
JTB' is weaker than JTB. That is, JTB' is entailed by JTB, but JTB' does not entail JTB.
(An aside on the relative strength of propositions. Suppose that p and q are propositions. Then:
- p is stronger than q iff p entails q but q does not entail p.
- p is weaker than q iff q entails p but p does not entail q.
- p and q are equally strong (or equivalent) iff p entails q and q entails p
- p and q are incomparable in strength iff p does not entail q and q does not entail p.)
We can think of JTB' as being the conjunction of the following four claims:
- Necessarily: for all s and propositions p: if s knows p then p is true.
- Necessarily: for all s and propositions p: if s knows p then s believes p.
- Necessarily: for all s and propositions p: if s knows p then s is justified in believing p.
- Necessarily: for all s and propositions p: if (p is true and s believes p and s is justified in believing p) then s knows p.
These claims are often abbreviated as follows:
- Truth is (individually) necessary for knowing. Or: Knowing is factive. (The truth condition.)
- Belief is (individually) necessary for knowing. (The belief condition.)
- Justification is (individually) necessary for knowing. (The justification condition.)
- Truth, belief, and justification are (jointly) sufficient for knowing.
Note that these four claims are consequences of both JTB and JTB', so each account is committed to them being true. Whether or not the first two are true is something we will consider soon; whether or not the last one is true is something we will consider next week.
A much weaker claim
Here is a third claim that is related to JTB and JTB':
JTB": For all s and propositions p: s knows p if and only if s has a justified true belief in p
What this says is that in the actual situation, the pairs of things that stand in the relation of knowing are also the pairs of things that stand in the relation of having a justified true belief in. That is, in the actual situation, these two relations are co-extensive. That is, these two relations are actually co-extensive. JTB" is not an identity claim.
JTB" is even weaker than JTB': JTB" is entailed by JTB', but JTB' is not entailed by JTB".
A claim like JTB" is not of much interest to us in our endeavour to answer the question of what knowing is - it need not tell us anything about the nature of knowing. For here is a true claim of a similar form, which does not tell us much about the nature of having a heart:
For all x: x has a heart if and only if x has a kidney.
Sometimes you will find people putting forth their theory of knowing in what seems to be the same form as JTB" (i.e. a claim about actual co-extensivity); what they probably mean is something of the form of JTB' (i.e. a claim about necessary co-extensivity)(they are taking the modal operator to be understood).
The justification condition
Here is a question that we should consider: Does the justification condition of JTB' render the truth condition superfluous? For suppose someone can only be justified in believing p if p is true - if p is false, then she does not count as justified in believing p, no matter how good her justification seems to be. Then we could just drop off the truth condition from JTB'.
But most people who endorse JTB' (and JTB) allow that one can be justified in believing propositions that are false, in which case the truth condition is not superfluous.
Is truth necessary for knowing?
Is truth really necessary for knowing? Here is an argument that it is not:
People used to know that the Earth is flat. But the Earth is not flat, and was not flat back then either. So people to used to know something that was not true. So it is possible to know things that are not true.
What should we make of this? The obvious response is to deny the first premise: people did not know that the Earth is flat, they just thought that they knew.
But this response faces a challenge. We seem to convey something true when we say, 'People used to know that the Earth is flat'. If what we say is strictly-speaking not true, then what are we conveying that is true, and how do we do it? If you can't explain this, then perhaps what we say is strictly-speaking true. (See the exercise for this week.)
Another natural example: "You shouldn't trust what John says, even if he claims to know what he's talking about. After all, he used to know that the world would end at 12am on 1st January 2000, and he knows that smoking is not bad for you."
Here is a different kind of case, one that I've recently been wondering about. John asks me the time; I look at my watch and it says 1:58 pm; I tell him "2pm"; concerned about the accuracy of my watch, he asks me if I know that it's 2pm; I tell him that yes, I do know (knowing full-well that it's 1:58).
What should we make of this? One natural response is to say that what I mean by 'it is 2pm' is that it is roughly 2pm, or near enough 2pm. But there is evidence that this is not so - consider how i would react if an eavesdropper were to challenge me. So what is going on here?
Is belief necessary for knowing?
Is belief really necessary for knowing? Here is an argument that it is not:
John knows that his house is on fire, but he doesn't believe it. Therefore, it is possible to know something without believing it.
What should we make of this?
Here is another kind of case:
Albert gives the correct answers to all the capitals of the US. He claims to not know know any of them. In fact, he claims to not have any beliefs about them either - he claims to be just guessing. But he seems to know them. He seems to know, for example, that Casper is the capital of Wyoming, even though he does not believe that Casper is the capital of Wyoming. Is this is a case of knowing without believing?
Consider a case in which a speaker communicates something true using, 'People used to know that the Earth is flat.' If this is not literally true, then what is being communicated, and how? Read through the following blog: http://el-prod.baylor.edu/certain_doubts/?p=610. Pick out at least three different explanations that are given there, and briefly describe them. Which of these do you favour? (Some entries are quite technical - just ignore them.)