Kevin McMahon
Phil 462: Theory of Knowledge
On Behalf of Nozick
In “Philosophical Explanations” Robert Nozick lays down one of the clearest and best accounts of knowledge that I have ever had the pleasure to read. As with any account of knowledge there are others who do not agree, and find problems within the account. Nozick originally wrote his “Philosophical Explanations” in 1981 and there has been plenty of time for a body of objections to blossom. In this paper I will defend the plausibility of knowledge using Nozick’s account, against the objections of Peter Klein in “On Behalf of the Skeptics”. Klein, who favors more of the reliabilism model of knowledge does not feel that we gain much by the use of Nozick. We will start by first laying down the basic tenets of Nozick’s account, and then we will examine Klein’s replies on behalf of the skeptics. We will then examine a few possibilities to still favor Nozick in the light of what Klein has to say.
First of all, Nozick is an externalist, he believes that whether or not something is knowledge is generally determined by factors outside of the mind. To him it is a matter of relations between our beliefs, and the causes of our beliefs. If there is an appropriate relationship between a belief and its cause, then it is a case of knowledge.
Nozick’s account of knowledge begins much the same way as most of the other accounts of knowledge available for us to examine today. The first tenet of his account is, of course, that person S must hold a certain belief P. Next, he holds that belief P must be true; otherwise there is no chance of actual knowledge, since by definition something we have knowledge of must be true.
In the third and fourth tenets, Nozick makes use of the subjunctive conditional to show us the causal relationship that is needed. He states that “if P were not true, then person S would not have the belief of P.” Roughly stated, if something was not true, then you would not believe it; this shows how the actual truth of P must be what causes us to have the belief in it. To secure the causal relationship even further, the final step uses another subjunctive, “if P were true, then person S would have the belief of P.” At first glance this seems a bit trivial since it is merely the opposite of the third tenet; we will see when we look at examples of each why this is necessary.
There is no large epistemological debate over the first two tenets of Nozick’s theory. They are well accepted, and they are used as the basis for almost all of the accounts of knowledge available to us. Everyone agrees that something more is needed than just true belief for knowledge. It is always at the something more, be it some sort of justification, reliabilism or a causal account like Nozick’s, that the debate begins, and so we will begin our discussion there.
Nozick’s argument using the subjunctive conditional requires us to look at other possible worlds to determine the truth value of the conditional. Since I do not have an available character especially for the use of the subjunctive conditional, from here on assume that ® represents not a standard conditional, but a subjunctive one. The third tenet put into logical notation is ~P ®~the belief of P. To properly utilize the subjunctive conditional we have to imagine, not just the actual world, but possible worlds where P would not be true. The third condition handles the Gettier-type examples against knowledge as true justified belief, rather easily. Let us consider the professor who believes one of his students owns a Ford. He believes this because he has spoken with student A who told him he owns a Ford. He had no reason to doubt the student and since he also saw him driving a Ford, he was justified in believing it. Unbeknownst to him, another student (B) actually does own a Ford, while student A does not. Although he does hold a true belief that a student in his class owns a Ford, it would seem as though he does not have real knowledge here because although he is right in believing someone in his class owns a Ford, it is for the wrong reason. When we examine other possible worlds we can see that Nozick does not allow this example as knowledge either. Let us imagine another world where student B does not own a Ford; the professor would still have the belief that someone in his class did own a Ford because he believes that student A owns one. He now has the belief of something that is false, thus violating Nozick’s third tenet. This violates the third tenet only because the proper causal relationship is not there.
The fourth tenet, P ® the belief in P, although at first sounds redundant is needed to ensure a proper causal relation in some cases. Nozick gives a clear example of why it is necessary. Suppose that person S really is a brain in a vat being stimulated by scientists. Now also suppose that person S can have no outside perception of the world around him save for that which the scientists feed into his brain. If they were to feed into his brain the sensations of being a brain in a vat then he would most certainly believe he was a brain in a vat. We have to look at the possible world, though, where he would be the brain in the vat and yet the scientists would give him no such belief. So here is a case where S does not have the belief that P, even though P. This then fails the fourth condition because there is an improper causal link, and no knowledge is then evident.
For these last two examples we had to consider other possible worlds to see if the conditional holds. Nozick does not think that we have to consider ALL possible worlds, just those worlds that are nearby. This allows us to rule out weird possibilities; as long as in all relatively similar situations the conditional would hold true. Generally, we know that if someone drops a ball it will fall. Suppose someone is holding a ball five feet in the air. We can say that if he were to drop the ball, then the ball would fall. This is true in the normal sense supposing nothing out of the ordinary happens. If a bird were to fly through a closed room and grab the ball, it would not fall, but that is a remote possibility. Since it is a remote possibility we do not have to use a world like that when we consider the subjunctive conditionals in Nozick’s account of knowledge. Later we will see how this helps us believe in the possibility of knowledge, in the face of skepticism.
Let us look at another example to clarify the nearby possible world theory. Suppose that person S sees his good friend Bob walking around the mall. He is in a hurry so he does not have the opportunity to stop and talk with Bob, but he does get a very good look at him as he walks by. He gets home and tells someone that he saw Bob at the mall. Does he know that he saw Bob? Under normal circumstances the answer is yes. S believes he saw Bob, and Bob was actually at the mall so it is true that he saw Bob. If Bob were not at the mall would S believe that he was? Under normal circumstances the answer is no. If we consider nearby worlds where Bob was not at the mall it is easy to conclude that if S had not seen Bob he would not think he was at the mall. We could imagine, though, a possible world where Bob has a twin brother Tom who was actually at the mall and not Bob. This makes it seem as though maybe S could not have knowledge because it is not true that if Bob were not at the mall then S would not have the belief that Bob was at the mall. It is true that he would hold the belief in the case that he mistook Bob’s brother Tom, for Bob. If we consider the fact that S has known Bob for a long time and had no reason to suspect that Bob had a twin then we can say that this possible world is not nearby and therefore S did know that Bob was at the mall. By adding that all the possible worlds we consider be nearby, we allow ourselves to rule out the weird or tricky situations from confusing our idea of knowledge.
Nozick adds another stipulation to the sufficient conditions for knowledge in his account. He thinks we should be concerned about the method by which we come to these beliefs. Our method used to come to beliefs when considering the conditionals must be consistent. If we use different methods to come to beliefs for different possible worlds then it is possible to confuse actual knowledge with no knowledge. Nozick states :
“1) P is true.
2)S believes, via method or way of coming to believe M, that P.
3)If P weren’t true and S were to use M to arrive at a belief whether (or not) P, then S wouldn’t believe, via M, that P.
4)If P were true and S were to use M to arrive at a belief whether (or not) P, then S would believe, via M, that P.”
Nozick characterizes this with his grandmother example. Grandmother S believes that her grandson is in good condition because she saw him earlier that day and he was not sick. We can imagine, though, a possible world where he is actually sick, and also that her family does not want her to know. Even though he is sick, they tell her that is he is doing fine. Therefore she has a belief in P that even if P were not true, she would still have. Nozick discounts this counterexample because the grandmother is using a different M, or way of coming to know. In the first case she believes that her grandson is not sick because she saw him earlier that day and he was fine. In the second case she holds her belief because of what she is told. If we consider only possible worlds that she still uses the first method in, she has knowledge because if she were to see him and he was very ill she would not think that he was well.
The traditional skeptic believes that we know very little or nothing of what we think we know. One of the standard skeptical arguments is that of the evil demon. The skeptic first puts forth the proposition that it is possible for us to be deceived by an evil demon and not know it. Said evil demon could possibly have such an elaborate deception around us that everything we see is all manufactured by him. Another possibility proposed by the skeptic is that we could be a brain in a vat on some distant planet. Scientists, who know way more about the brain than we do, have found ways to put beliefs into this brain in a vat. They pump the beliefs of an entire existence on earth into that brain. Without even being on the same planet, that brain-person now has the belief in many things around it. Since it is conceivable for us to be deceived in this manner, and not know it, the skeptic claims that we can not know anything about what we perceive.
It is by using the “nearby possible worlds” clause that we can avoid this argument. I have the belief that my computer is sitting in front of me as I type this paper. I am using the method of my eyesight to determine this. If we consider the evil demon possibility then condition 3 is not met because I would still have the belief that my computer is in front of me even if it was not. The nearby possible worlds idea saves us though, because the idea of an evil demon is not relatively close to the actual world. The evil demon hypothesis is one of those weird cases that the use of the subjunctive excludes. In this case we are able to say that I have knowledge of the computer in front of me. In relatively similar possible worlds when there is not a computer in front of me, I do not believe there is. Also, condition 4 is met because in nearby possible worlds, when there is a computer in front of me, I do believe that there is.
The skeptic does have a little bit more of a role to play before we are done with him for now. His ugly head rears again when we consider knowledge by inference. This is when a belief P entails the belief in some other proposition Q. It is said that if person S knows that P and also that P entails Q then person S also knows Q. If it can be later said that person S does not know Q then it should also be said that they do not know P since it entails Q. This is the idea that knowledge is closed under entailment. It has been used before to derail other accounts of knowledge. Nozick believes though that he can defeat this principle, and he can.
Nozick characterizes the skeptics arguments about entailment in the following way. Let’s take the example again of my belief that there is a computer in front of me. This is belief P. The skeptic claims that if I were to know that there is a computer in front of me, then that entails that I also know that I am not being deceived by an evil demon. The fact that I am not being deceived we will say is belief Q. Since, even Nozick acknowledges that I cannot know that the evil demon is not deceiving me, then I cannot know that the computer is in front of me. Nozick claims that knowledge is not closed under entailment and that it is possible for me to know that P and know that P entails Q and yet not know Q. This is because of truth tracking. When a belief meets all four of Nozick’s criteria it is said to track the truth. My belief P that the computer is in front of me, as discussed earlier, tracks the truth by meeting all of the criteria. The belief Q, though entailed by P, does not track. It fails to meet number 3 because even if I were to be deceived by the evil demon I would not know it. Nozick says that although some knowledge is available by inference, all inference does not lead to knowledge. Both P and the Q that it entails must track the truth for it to be considered knowledge. This is how Nozick escapes the skeptical argument of the closure principle.
In a much shortened version, his is over 110 pages long, that is a characterization of Nozick’s account of knowledge. He neatly met with many of the classic arguments against knowledge. He defeated Gettier, skeptics generally, as well as the closure principle of entailment. Peter Klein in “On Behalf of the Skeptic” did seem to find a few flaws in Nozick’s work; I will try to answer on behalf of Nozick.
After also doing a brief overview of Nozick, Klein goes right into a counterexample that he feels proves Nozick’s account is inaccurate because it is too strong. He felt that it denied us the ability to say that we have knowledge in cases when we really do. If this were true, it would be easy to agree that Nozick’s account was inaccurate and should not be followed.
He first uses the “hit rock case” to test Nozick’s account. Take an example where someone is driving down the road in their car and a rock flies up and hits the gas tank gauge under the car, thereby dismantling it. This causes the gas gauge to read empty. The person driving the car looks at the gas gauge and comes to the belief that the tank is out of gas. Coincidentally the tank is actually out of gas. It is clear that person S has no knowledge of the fact that the tank is empty because it is clearly just by chance that he came to the correct belief. Nozick’s system would agree since it fails to meet number 3 because even if the tank were not empty, he would still believe that it was, since the gauge is broke.
Klein then goes onto the “missed rock case.” This is an almost identical case as the first, but instead of the gas gauge being hit by the rock, the rock just misses and does not strike the gauge. The gas gauge is still working fine and reads that there is no gas in the tank. Since the gauge is working fine we should want to say that person S has knowledge that the tank is empty. Klein tries to prove the inadequacy of Nozick’s account by attempting to show that it wrongly denies S his knowledge, in this case, of the fact that the tank is empty. Klein imagines a possible world where the tank was originally half full instead of empty. In this possible world, the extra weight made the gas tank hang just a little lower and, therefore, it was struck by the flying rock. The gauge is no longer working properly and now reads empty. Therefore person S would still have the belief that the tank was empty even in the case that it was not. Klein says that, employing Nozick’s rules, this is not a case of knowledge because there is a possible world where S would believe P even in the case that ~P. For Klein then, Nozick’s whole system is out the window because it denies us knowledge of something that it seems clear to say we should have knowledge of.
There are two possible responses to this counterexample that show that Nozick is still plausible. The first is to assume that we are using a different M or method of coming to know P, in each case. In the empty gas tank world, we are using the working gas gauge to know that the tank is empty. In the nearby possible world, we are using a broken gas gauge to come to know that the tank, which is not empty, is empty. Klein already tried to answer this possibility by quoting from Nozick about the identity of different methods.
“The notion of knowledge holds the method fixed...fixed enough only to exclude differences the person would detect, believing it to constitute a difference. The method used must be specified as having a certain generality if it is to play the appropriate role in subjunctives. This generality is set by the differences the person would notice...”
Klein claims that the person would not notice a difference at all in each case. I believe though that person S would notice a difference. Were the tank actually half full and the rock were to break the gauge then there would be a sudden and dramatic change in the perception of Person S as to the contents of the tank. We can only truly believe that the gas gauge is right if we have experience that it is generally a reliable method of knowing how much gas is in the tank (I myself have driven cars that I knew not to trust what the gauge said, because it was unreliable). Thus we can only be said to have knowledge based on what the gas gauge says, if we believe it to be reliable. The only way to know that the gas gauge is reliable is to keep a general idea of how much gas is in the tank. I think that in the half full tank example, if the tank suddenly dropped to empty that the person would notice a difference in M. Since the person would notice a difference in M then this counterexample cannot be used to show that the belief does not track the truth in Nozick’s third condition.
This idea of the reliability of the gas gauge also brings up another response to Klein’s argument. If we know that in the normal sense, the gas gauge is correct and reliable, then we can look at it and know what is in the gas tank by the use of the gauge. We know this because barring any weird circumstances it is reliable. Take for example the man holding the ball up in the air. We know that if he drops it, the ball will fall; unless a weird circumstance like a bird flying through the room prevents it from happening in the normal sense. It is the same with the gas gauge; a rock coming up and hitting the gauge is a weird circumstance, so it is not a nearby possible world, even though it is a possible world. Again, since this possible world is not a normal case being relatively similar, it is not to be counted among the possible worlds we consider for the use of the subjunctive conditional. Since we cannot count this possible world, Nozick’s account holds up in the face of Klein’s counterexample.
Klein then attempts to negate Nozick’s answer to the closure principle by employing an example similar to the hit rock and miss rock examples already given. He uses the hit rock and miss rock in conjunction with the “taped gauge case.” Suppose that the gas gauge in the car was partially taped over so that from 1/4 tank to 3/4 of a tank was hidden. There would then be no way for person S to directly know that there is half a tank of gas in the car. It would be possible however, for S to know that there is no gas in the tank of the car, and then by inference to know that there is not a half tank of gas. Let P be the belief in no gas, and Q be the belief in not a half tank of gas. Belief in P logically entails belief in Q, and then as long as the truth tracking holds up, knowledge of P entails knowledge of Q. This inference certainly seems of the type that would be acceptable since, in the normal sense, both the belief P and the entailed belief Q track the truth through 3 and 4 of Nozick’s account. Since Klein felt that he proved in the miss rock case that it was possible to deny knowledge when it was there he felt that employing the same method here would also deny knowledge of the entailment. We have seen though that knowledge of P is allowed using Nozick’s account. In all nearby possible worlds we believe the gas tank is empty only when it is in fact empty. Also then, in all nearby possible worlds when we know that the gas tank is empty, we also know that it is not half full. We know this because we would not believe the tank to be not half full if it were half full. Hence, both the knowledge of P and the knowledge of Q track the truth of P and Q respectively. Since they both track the truth according to Nozick’s system, it is a case of knowledge by inference to Nozick. It does not violate any principles so again Klein is refuted in his attempt to show that Nozick denies us knowledge of things we actually do have knowledge of.
Klein also attempts to employ an argument for skepticism based on the simple idea that we can be deceived. He uses an idea from the ancient text of Cicero to illustrate how one could be fooled by twins into an incorrect belief. Nozick never states that it is impossible to be deceived. He does hold, though, that in the normal case one is not deceived. Klein’s argument for deception is not strong enough to defeat Nozick for reasons stated earlier in the paragraphs about nearby possible worlds. It may give us some doubt as to what we can know, but it will always end in a tug of war between Nozickians and skeptics about where to draw the line of nearby possible worlds. For example, if we allow a world full of identical twins to be considered nearby, then knowledge about people is nearly impossible.
I believe that we cannot function without knowing something; our daily lives would be pathetic and meaningless if we knew absolutely nothing. Therefore, I favor drawing the line of nearby worlds at a point where I can have knowledge of those things that we generally feel we know, and yet ruling out skeptical hypotheses. We can always admit the possibility that we could be wrong, and still have knowledge. In light of these denials of Klein’s counterexamples, and pleas to set the nearby limit in a reasonable place, I believe that Nozick clearly survives. There is no reason why we cannot continue to believe that Nozick had the right idea in “Philosophical Explanations.”
Bibliography
Klein, Peter. “On Behalf of the Skeptic” Luper-Foy pp 267-281.
Luper-Foy, Steven, ed. The Possibility of Knowledge:
Nozick and His Critics. Rowman and Littlefield.
Nozick, Robert. Philosophical Explanations. Belknap
Press.
Home Outdoors
Various School Projects Rat Babies Wacky
Test Results Contact Kevin McMahon