Philosophy Things

Saul Kripke: A Criticism of Nozick's Reliabilism

In response to a traditional view of what criteria constitutes knowledge; Edmund Gettier presented two counter-examples to the widely accepted theory, sparking a revolution of thought for epistemologists. These two examples presented by Gettier, however, were by no means flawless. In an attempt to “patch” these issues that the cases entailed, Robert Nozick presented his own theory of knowledge by using reliabilism in place of the justification condition. In response to Nozick’s reliabilism theory, philosopher Saul Kripke presented counter-examples that appear to have uncovered flaws with Nozick’s theory. While Nozick did modify his tracking theory to include belief formulation methods, the task involved with relating a belief to truth still remains an issue, primarily due to the difficulty involved in defining how, on what basis, reliability between the two can be measured.

Traditionally, the conceptual analysis of propositional knowledge entails three “individually and jointly sufficient” (Bernecker 5) conditions: truth, belief and justification. The truth condition of this analysis asserts that if one knows a certain proposition, then that proposition must be true. The belief condition of this traditional account states that if one knows a certain proposition, then such knowledge entails the belief of that proposition as well. The final condition, the justification condition, requires that if one has knowledge of a proposition, then that proposition must be evidentially supported, which serves to prevent lucky guesses from counting as knowledge. While this analysis may, on the surface, appear to be an absolute criterion for knowledge, there are actually several problems associated with employing this method. One of the biggest issues, perhaps, is the difficulty in determining a complete set of necessary and sufficient conditions that can be effectively used across different languages, societies and generations.

Edmund Gettier was the first philosopher to publish his concerns regarding this “justified-true-belief” analysis of knowledge. While he agreed that all three conditions were necessary, he disagreed that those three conditions alone were jointly sufficient for knowledge. The two examples that Gettier presented exposed the possibility that someone could have a justified, true belief without having knowledge. Both examples shared the same commonality; that what justifies a person in believing a proposition can be different from what actually makes the same proposition true. In other words, “There [can be] a gap between the satisfaction of the justification condition and the satisfaction of the truth condition” (Bernecker 7).

In an attempt to bridge the gap between the justification and truth conditions exposed in the Gettier cases, philosopher Robert Nozick proposed two counterfactuals to function in place of the justification condition. They are the “variation” condition and the “adherence” condition. The variation condition states that if any proposition were not true, then the person would not believe it to be true. The adherence condition states the opposite; that if a proposition were true, then the person would believe it to be true. The key notion behind Nozick’s theory is that for any belief to count as knowledge, it must be reliably related to the facts that make it true. This reliability approach is Nozick’s attempt at blocking the Gettier cases. Or, to put it another way, it is Nozick’s attempt at preventing accidentally true beliefs. Once a belief satisfies all four conditions for knowledge (truth, belief, variation and adherence), Nozick claims that the belief “tracks”, “hooks up with”, or “indicates” the facts that make it true, and in the process, eliminates the possibility of arriving at a true belief by lucky coincidence.

As a counter-argument to Nozick’s reliabilism, a “grandmother” case was presented in an attempt to show the importance of belief formation methods. The example goes as follows. Imagine a grandmother that only knows of her grandson’s health when he comes to visit her. But even if he were sick or dead, the relatives of the grandmother would still tell her that grandson is well, in order to spare her the emotional strain. This example violates the variation condition; on the basis that just because the grandmother doesn’t know that her grandson is not healthy, that doesn’t mean that she doesn’t believe that he is not healthy (on the contrary, she believes that he is healthy).

In response to this criticism, Nozick’s theory was revised to include a clause about belief formation methods. His new theory of knowledge includes the following, updated criteria for the belief, variation and adherence conditions: 1) the person believes that a proposition is true based on some method of coming to believe it, 2) if the proposition weren’t true and the person were to use the same method to arrive at that belief, then the person would not believe the proposition via that particular method, and 3) if the proposition were true and the person were to use the method to arrive at the belief, then the person would believe the proposition via the same method.

In criticism of Nozick’s reliabilist analysis of knowledge, philosopher Saul Kripke illustrates “The Red Barn” scenario in which he argues that a belief can track the facts without leading to knowledge, deeming Nozick’s knowledge theory unreliable. The example goes as follows. Peg lives in an environment where both real barns and fake barns exist. All real barns are painted red, and all barn facades are painted colors other than red. Peg does not know that some barns are real and some are fakes; she believes that all barns are real and has no reason to think otherwise. While Peg is observing a red barn, she believes it to be a real red barn. According to Nozick, Peg satisfies all four of the conditions for knowledge. If the barn were real, she would believe it to be real (which it is, and she does). If the barn were not a real barn, and say, a red house instead, Peg would not believe it to be a barn and thus would also satisfy the variation condition. (Adams 215)

According to Kripke, however, because Peg believes all barns to be real, she could very well have been looking at a white barn, which is a fake barn, and she would still think that it was a real barn. It is, only by lucky coincidence, that Peg is looking at a real red barn claiming to have knowledge that it is real, rather than looking at a fake white barn and claiming to have knowledge that it is real. Kripke claims that because the variation condition is not satisfied in this possible world of fake barns (because if Peg saw a fake barn, she wouldn’t believe it to be fake), Peg does not have knowledge of the red barn. Furthermore, Nozick’s tracking theory seems to have proved insufficient in determining knowledge.

Finding this “Red Barn” counter-example presented by Kripke is a damaging one for Nozick because for Nozick, all four conditions for knowledge must be satisfied in order to qualify as knowledge: truth, belief, variation and adherence. If Peg only meets three of the four criteria, she fails to meet the standards of Nozick’s theory, and as a result, she fails to qualify as a candidate for having knowledge.

A possible Nozickian response to Kripke’s criticism could be one that focuses on the relativised method of belief formation discussed in the grandma case. For example, the belief formation that leads Peg to her knowledge that there is a barn is the same in the adherence condition as it is in the variation condition: the “red-barn-look” method (Adams 215). Peg detects that there is something red and something barn-ish, and this belief formation method is what leads Peg to satisfy both of Nozick’s conditions. Peg believes the red barn to be a red barn because of its property of redness and its property of barn-ish-ness (which affirms the adherence condition), and Peg would not believe the red barn to be a red barn if it didn’t have these two properties. For example, if Peg were observing a red house or a blue car, etc. she would not claim to be observing a red barn (affirming the variation condition). Given this type of response, it seems that Kripke’s argument is no longer successful, or at least it would only be successful if he violated Nozick’s relative, belief formation method. (Adams 215)

Another response that Nozick might give to Kripke’s criticism is that in regards to “possible worlds”. Sven Bernecker defines a possible world as one that is “a fictitious situation which differs from the actual world in some respect” and goes on to say that these worlds are “varying in their degree of resemblance to the actual world” and that “some are more similar to our world than others” (Bernecker 215). Nozick makes clear that the variance and the adherence conditions do not need to hold in all possible worlds, and that these conditions claim no guarantee to truth or the infallibility that is explicated from them (Bernecker 23). These could prove to be successful counter-arguments to Kripke’s criticisms. After all, the possible world that Kripke proposes is one where all red barns are real and all other barns are facades, and due to the fact that such a strange world is not a close replica of the real world in which we live, it is likely that Nozick would discredit any tests or scenarios in those worlds as bearing any significant impact on the outcomes of his two reliabilism conditions.

In conclusion, while several attempts have been made to define a strict, foolproof guide to knowledge, it seems that nobody has been able to accomplish that task yet. The attempts at defining the criteria have all resulted in being either too broad, where too many beliefs are justified as being knowledge, or too narrow, where too many cases of knowledge are excluded from the outcome. Additionally, although Nozick was able to revise his reliabilistic account of knowledge to include the addition of belief formation methods, the difficulties with reliabilism still remain in that it is not yet determined what criteria must be met in order to meet the requirements of reliability.

Works Cited

Adams, Fred, and Murray Clarke. "Resurrecting the Tracking Theories." Australasian Journal of Philosophy 83.2 (2005): 207-21. Print.

Bernecker, Sven. "Defining Knowledge." Reading Epistemology: Selected Texts with Interactive Commentary. Malden, MA: Blackwell Pub., 2006. 15-26. Print.