Okay! Brace yourselves for a tricky piece of work. (Well, it was tricky for me…) I’ll do my best to give an accurate summary of Robert Nozick’s counterfactual account of knowledge, but there may be some slight errors. Basically, Nozick is suggesting that knowledge is belief that tracks the truth. Recall our JTB account of knowledge:

Subject (S) knows something (p) iff:

(P1) p is true
(P2) S believes that p
(P3) S is justified in believing that p.

Note: ‘iff’ means ‘if and only if’

Now, Nozick proposes to replace the troublesome justification condition (P3) with two premises that enable us to ‘track the truth’. So even if things had been otherwise, our subject would still have had a true belief. Firstly then, our new argument:

Subject (S) knows something (p) iff:

(P1) p is true
(P2) S believes that p
(P3) if p were true, S (using method M) would believe that p
(P4) if p weren’t true, S (using method M) wouldn’t believe that p

(Method M is added to ensure the same type of method is used when determiningwhat we come to believe. This avoids small counterexamples that can crop up).

What does this mean? Well, try thinking of it like this. (P3) could be rewritten as:

(P3) If p were true, S would believe it was true even in a close possible world.

Similarly for (P4):

(P4) If p were not true, S wouldn’t believe it was true even in a close possible world.

Did you read about Possible Worlds? This is a case where that controversial account of Lewis’s helps us to understand tricky ideas in Philosophy such as these counterfactuals. Here, we are thinking about what would be the case in the closest possible world to our own. Why does this help us? Well, let’s consider those troublesome Gettier Cases again shall we.

Stopped Clock

Here we have the case where I look at the clock and see 11.56. I form a justified true belief that the time is 11.56. However, the clock is broken. But by a stroke of luck, it is showing the correct time (the clock stopped exactly 12 hours prior to me looking at it, thus it is showing the correct time). But in this case, most of us would say we do not know what the time is because our true belief was based on luck.

In comes Nozick with his truth-tracking theory. (You will notice that, since we are talking about close possible worlds, we only change one small thing in them). Consider (P3) and ask yourself:

If p (the time is 11.56) were true, would I believe it were true even in a close possible world? (In our example the closest possible world is where p is true. So that would be the world where the clock is working correctly and shows the real time [11.56]).

Yes, I would believe p is true (and my belief would be correct). This would be like our normal world. When the clock is working correctly and I read the time (11.56), I form a justified true belief from it and gain knowledge. No problems here then! Onto (P4), ask yourself:

If p (the time is 11.56) were not true, would I believe it were true even in a close possible world? (In our example the closest possible world is where p is false. For example, a world where I arrive one minute later and look at the broken clock, so the real time would be 11.57).

Yes, I would believe that p is true (but my belief would be wrong!!) The time would read 11.56 because the clock is broken showing that time. But in the closest possible world, the real time is 11.57, so my belief is false and I do not have knowledge. This fits with our original intuition that we do not have knowledge in the Stopped Clock example. Hence, Nozick has tracked the truth and avoided the Gettier Case.

**********

Phew! Still confused? Try the second Gettier Case:

Charlie’s Book

Remember Charlie? Charlie has a copy of Epistemology: An Anthology. I have seen this book many times. But one day, unbeknownst to the both of us, Charlie’s book is stolen from his bag. Luckily, he has a second, hidden copy of the book in his room at home that I’ve never seen. I have a justified true belief that Charlie has a copy of the book, but it seems wrong.

Again, since this seems to be lucky knowledge, most people agree I don’t know that Charlie has a copy of the book. But by the JTB account, I do know. How to solve it? Try tracking again:

p is (Charlie has a copy of the book Epistemology: An Anthology)

(P3) If p were true, S would believe it was true even in a close possible world.

The closest possible world where p is true is where Charlie has one copy of the book (it is not stolen). So my belief is true. This is the normal outcome.

(P4) If p were not true, S wouldn’t believe it was true even in a close possible world.

The closest possible world where p is not true is where Charlie does not have a second copy of the book (i.e. In order for p to be not true, Charlie must not possess any extra copies of the book – his first one is stolen anyway, so it doesn’t count). So Charlie has his book stolen without either of us knowing, and he doesn’t have a second copy in his room. However, I would still believe that Charlie had a copy, but my belief would be false (as Charlie has no books at all in this possible world). As we know, a false belief is not knowledge, so Nozick has managed to track the truth.

So, the problem for us was the modified Gettier Case ‘Fake Barn Country’. Let’s see how tracking can hold up.

Fake Barn Country

Henry is in Fake Barn Country (FCB) where he is only seeing papier-mâché façades of barns (i.e. they aren’t real barns). But at one point Henry looks at the one barn that is actually real. His true belief is justified, so it should count as knowledge. But it doesn’t seem to because Henry is merely lucky that he is looking at the one real barn.

In this case Henry has a JTB and is not gettiered in the original sense because there are no tacit assumptions that turn out to be false. In the ‘Stopped Clock’ for instance, our tacit assumption was that the clock was working. But is there really a tacit assumption that Henry is not in a country filled with fake barns to deceive him? It seems unlikely. For this reason, cases like ‘Fake Barn Country’ have sometimes been labelled as ‘Dangerous Gettier Cases’.

With tracking:

p is (Henry is seeing a barn)

(P3) If p were true, S would believe it was true even in a close possible world.

(P3) The closest possible world where p is true is where Henry is seeing a real barn. This is the normal case where Henry passes through a place with lots of real barns around him instead of fake ones. No problem.

(P4) If p were not true, S wouldn’t believe it was true even in a close possible world.

(P4) The closest possible world where p is not true is where Henry is not seeing a barn. That is, a world where all the barns in Fake Barn Country are fake. Then, when Henry sees a papier-mâché façade of a barn in this possible world, his belief will be false. He will still believe that he is seeing a real barn and be justified (because he isn’t assuming he is in FCB), but the barn that he is looking at is fake, as are all the rest, and so falsifies his belief. Hence, Henry doesn’t have knowledge and Nozick appears to have tracked the truth.

But can we get away like this? Many philosophers have argued no; that Nozick’s account can still fail. An example is given by Kripke who tells us to consider that the only real barn in FCB is painted red. A section on Wikipedia clearly explains the problem:

(Example from Wikipedia:)

“…the Fake Barn Country example, which explains that in a certain locality are a number of fake barns or facades of barns. In the midst of these fake barns is one real barn, which is painted red. There is one piece of crucial information for this example: the fake barns cannot be painted red.

[Henry] is driving along the highway, looks up and happens to see the real barn, and so forms the belief:

* I see a barn

Though [Henry] has gotten lucky, he could have just as easily been deceived and not have known it. Therefore it doesn’t fulfill premise [P4], for if Jones saw a fake barn he wouldn’t have any idea it was a fake barn. So this is not knowledge.

An alternate example is if Jones looks up and forms the belief:

* I see a red barn.

According to Nozick’s view this fulfills all four premises. Therefore this is knowledge, since Jones couldn’t have been wrong, since the fake barns cannot be painted red. This is a troubling account however, since it seems the first statement I see a barn can be inferred from I see a red barn, however by Nozick’s view the first belief is not knowledge and the second is knowledge.”

So we see that there are yet more problems facing us when trying to sort out our account of knowledge. Furthermore, Nozick’s account doesn’t seem to be able to answer our ‘brain in a vat’ problem either, lapsing into our all familiar circle of needing to know that I’m not a brain in a vat in order to know I have hands and so on…

But, if one does accept Nozick’s tracking account, then one ends up denying Closure (Closure fails because of the tracking condition for knowledge). In which case, the skeptic cannot reply on the Closure ‘brain in a vat’ argument. Anyway. For now, we leave tracking here.

**********

Sources: Lecture handouts written by my lecturer Daniel Elstein, The Internet Encyclopedia of Philosophy and The Stanford Encyclopedia of Philosophy