Introduction

We discovered in the last chapter that there is a difference between true belief and knowledge. There are many instances in which somebody has a true belief but does not know the proposition that she believes. For example, imagine that Erika read in a comic book that there is alien life on Mars and thus comes to believe that there is alien life on Mars. Furthermore, let’s say that there is a single species of microbe that lives on mars, if this was the case then Erika would be correct; however, despite her truth belief, Erika has not attained knowledge. Why not? The simple answer is that we don’t think that somebody can attain knowledge without reason. You must achieve knowledge.

In this chapter I am going to explain some prominent theories of epistemic justification and warrant. A theory of justification or warrant is a theory regarding what turns mere true belief into knowledge. Alternatively, I could say that a theory of justification or warrant is a specific explanation of what condition removes the possibility of luck from knowledge. By the end of this chapter, you should be able to:

• Explain the difference between justification and warrant
• Explain the difference between epistemic internalism and externalism
• Indicate whether any theory of justification or warrant is an internalist or an externalist theory
• Explain some of the prominent theories of justification and warrant

Internalism, Externalism, Justification, and Warrant

A good chunk of epistemology over the last eighty years has been dedicated to solving the Gettier problem. Epistemologists are often put into two factions: internalists and externalists. Very roughly, internalists believe that the condition which differentiates knowledge from mere truth belief relies solely on something that happens in the mind of the believer. Internalists refer to this differentiating condition as ‘justification’. Meanwhile, externalists believe that the differentiating condition between knowledge and true belief relies (at least partially) on something that happens outside of the believer’s mind. Externalists refer to this differentiating condition as ‘warrant’.[1] This is a confusing distinction, and it will be helpful to go over a few cases to illustrate this difference.

Scooter and the Magic Eightball: Scooter is a sad boy whose only possession is his Magic Eightball. Scooter—for seemingly no reason—always believes what his magic Eightball says. One day, Scooter asks his Magic Eightball whether he will ever become rich and famous. The Magic Eightball replies positively to Scooter, which makes him believe that he will indeed become rich and famous. Unknown to Scooter, his Magic Eightball is controlled by an omniscient God who always tells Scooter the correct answer. Does Scooter know that he will become rich and famous?

On an internalist account, we would consider the information that Scooter has in his head and determine whether he has enough information to be justified in his true belief. If so, then he would know his belief. In this scenario, it is apparent that Scooter does not have enough justification to know that he will become rich and famous. Although his Magic Eightball is correct, Scooter has no reason to believe that it is correct. Meanwhile, if we were externalists then we would consider the factors that happen outside of Scooter’s mind. One very popular externalist theory, reliabilism, essentially states that the believer is warranted (which is the externalist’s equivalent to justification) in their belief if they have formed their belief via a method that generally leads to true belief.[1] Since Scooter is forming a belief based off information given to him by an omniscient god, it follows that Scooter’s belief is warranted. Thus, since it is warranted and true, then it follows that Scooter knows that he will become rich and famous (according to a reliabilist).

Take note that an externalist does not have to claim that the believer knows why their belief is warranted, they just need to state that it is warranted. In fact, according to the typical externalist, the believer’s knowledge is not dependent on anyone being aware that they have achieved knowledge. Furthermore, the believer’s knowledge is not dependent on anyone knowing that the believer is warranted in their belief. So, if Scooter merely believes that he will become rich and famous then he would achieve knowledge, even though nobody thinks that he has attained knowledge and even though nobody thinks that he is warranted in his belief.

I generally notice that students who have just begun studying epistemology find externalism odd and confusing. Why would we want to say that Scooter has attained knowledge in this situation? The answer goes back to the Gettier problem that I discussed in the previous chapter. To summarize the Gettier problem, knowledge has been traditionally conceived of as justified true belief; however, there are describable instances (called Gettier cases) in which a person can have a justified true belief while they intuitively lack knowledge. There is a famous Gettier case attributed to the philosopher and mathematician Bertrand Russell referred to as “Russell’s clock” (1948: 170-1). I’ll use this example as a reference point so that we can analyze how externalists attempt to solve the Gettier problem.

Russell’s Clock: Bertrand walks into a room and looks up at the clock on the wall. The clock is set to 12:30, and it is daylight so Bertrand infers that it must be 12:30 p.m.. Bertrand is right, it is 12:30 p.m.; however, it turns out that the clock has been broken for exactly 12 hours and has thus been stuck at the 12:30 position for some time. So, in a sense, Bertrand was simply lucky that he got the time right. Despite that, Bertrand has a justified true belief that it is 12:30 p.m., but he intuitively doesn’t know that it is 12:30 p.m.

How can externalists solve this Gettier case? Let’s consider the reliabilist theory of warrant that I discussed above (just to give an example of how some externalists would respond). According to a reliabilist, Bertrand knows the time if his belief that it was 12:30 p.m. is true and he came to that belief via a reliable belief forming process. Reliabilism is an externalist theory of knowledge, so we should consider factors that are unknown to Bertrand if we are attempting to analyze if he has attained knowledge or not. One thing which we know (but not Bertrand) is that the clock is broken. Furthermore, we know that broken clocks are unreliable at telling the time. Thus, the reliabilist can state that Bertrand does not know the time since he is basing his belief off an unreliable belief forming method (the method of believing a broken clock). All externalists will attempt to pull a move like this. Externalists essentially say that if we focus on some aspect of the world—rather than the mind of the believer—then we can easily filter out the problematic features of Gettier cases and explain why people in such cases do not attain knowledge. Furthermore, externalists generally hold that we can only avoid the Gettier problem if we include a warrant condition in our theory of knowledge.

Internalists have made a variety of attempts to avoid the Gettier problem while adopting justification (rather than warrant). Remember, justification is essentially dependent on what the believer is aware of in their own head. So, Bertrand—according to the internalist—is justified in believing that it is 12:30 p.m. since the clock says that it is 12:30 and it is light outside. Philosophers such as Lehrer and Thomas (1969) have attempted to avoid the Gettier problem by adopting the no defeater strategy. Essentially, internalists attempt to say that knowledge is equivalent to what they call undefeated justified true belief. This—according to the internalist—is close enough to the original justified true belief condition of knowledge. According to the no defeater strategy, if there is a fact that—if learned—would cause the believer to disregard their belief, then they do not know the initial belief. For example, if Bertrand learned that the clock was broken, then he would not believe that it was 12:30 p.m.. This fact that—if learned—would cause the believer to give up their belief is referred to as the “defeater” (hence the definition that states that knowledge is undefeated justified true belief).

Of course, epistemologists such as Levy (1977) have fired back against the no defeater strategy and pointed out counterexamples in which a believer might give up their belief after having learned a fact, however, the defeater—in these counterexamples—is epistemically irrelevant. For example, suppose that Bertrand walks into a room which is lit up by daylight and looks at a clock which reads 12:30. After seeing this, Bertrand reasonably infers that it must be 12:30 p.m.. Suppose that Bertrand is right, and the clock is actually working; however, Bertrand is picky about clocks and has an irrational distrust for one specific clock company called Cloque. Let’s suppose that this clock was made by Cloque, and that Bertrand would not believe that it was 12:30 p.m. if he learned this fact. The no defeater strategy entails that Bertrand would lack knowledge in this situation since he would disregard his belief if he learned that the clock was made by Cloque; however, I think that Bertrand intuitively knows the time. Internalists have come up with responses to this type of argument; however, the debate about whether these responses work is still ongoing.

Many of you may be wondering, how do you decide between internalism and externalism? What reasons do internalists and externalists generally give in favor of their respective positions? Put simply, externalists can easily avoid Gettier cases which means that they can avoid stating that somebody can achieve knowledge due to luck from outside circumstances. For example, Scooter was lucky that the Magic Eightball was controlled by an omniscient god. Internalists on the other hand have a harder time avoiding Gettier cases, but internalism surely seems much simpler and more intuitive than externalism. At the end of the day, there is not a clear answer as to which you should believe.

Luckily for those of you who are indecisive, you may not have to pick between internalism and externalism. After all, knowledge could be defined as “justified true warranted belief.” This combo approach seems intuitive to many epistemologists; however, strict internalists and externalists don’t want to compromise their position. Many epistemologists—both on the internalist and the externalist front—hold that such an approach is simply too complicated. After all, parsimony (or simplicity) is often considered a virtue in epistemology.

Internalist Theories of Justification

Epistemology is a fundamentally normative study. In other words, epistemology is the study of how people ought to reason to attain knowledge. The internalist holds that justification is a necessary condition of knowledge. Furthermore, the internalist holds that a believer’s justification is dependent on some mental process in her mind. Thus, if we are going to take the internalist position seriously, it follows that we ought to discover what justification is in a specific sense.

It is worth noting that some beliefs seem intuitively justified while others seem intuitively unjustified. For example, most people will agree that Scooter is unjustified when he believes the word of his Magic Eightball and that Bertrand is justified to believe that it is 12:30 p.m. after reading the time on the clock. Some of you may recognize these intuitions and begin to wonder why it is important to uncover the underlying causes for our justification when we can simply intuit if we are justified in believing some belief. I understand the concern; however, it is important to note that our intuitions differ in many cases. For example, consider how some religious people view religions that they do not believe in. It is not uncommon for a religious person (who believes religion 1) to look at another religion (religion 2) and say something like, “how could people believe those wacky stories?” However, it’s likely the case that somebody who believes in religion 2 would look at religion 1 and say the exact same thing. Thus, it seems like humans often have inconsistent intuitions, which is obviously problematic for epistemologists. It follows that it is critical for the internalist to study justification.

There are two major theories of justification that internalists usually choose between. The first, evidentialism, is the view that a belief is justified if it has sufficient evidential support. For example, Bertrand was justified to believe that it was 12:30 pm because the clock reading 12:30 during daylight offered him good evidence to believe that it was 12:30 pm. Thus, unlike coherentism, the evidentialist considers every belief individually and then considers all the evidence relevant to that belief. If there is enough evidence, then the belief is justified.

Evidentialism sounds plausible; however, it perhaps seems a bit unsubstantial. Some people may think it is obvious that the epistemic justification of a belief is tied to the evidence that one has for that belief. It seems like the important question should be, “what makes one piece of evidence better than another piece of evidence?” Thus, the evidentialist may be trying to push the question back a step without giving much of an answer. However, it’s important to note that we are not yet at the point where we can formulate exact methodologies that can determine what specific beliefs to hold. Instead, we are currently doing epistemology at a bird-eye or general level. Afterall, it is important to figure out what makes a belief justified in principle before we can determine why some specific beliefs are more justified than others. Thus, the evidentialist will likely state that the claim that evidentialism is unsubstantial is just evidence for the claim that evidentialism is incredibly intuitive. Furthermore, the evidentialist will say that we should determine exact methodologies to help us form specific beliefs once we all agree that evidentialism is correct.

The second major theory of justification, coherentism, is the view that a belief is justified if and only if it coheres with your other beliefs. To elaborate, coherentists do not believe that you can determine a singular belief by analyzing that belief. Instead, you need to think of everything that you believe and then you need to consider whether the new belief contradicts any of your previous beliefs. If the belief contradicts any of your previous beliefs, then that belief is unjustified. If the belief does not contradict any of your previous beliefs, then it coheres with your beliefs and is justified.

Coherentism is simple which is a positive for the viewpoint. Afterall, coherentists don’t have to do much elaboration to define what a justified belief is since we all know what it means for one belief to contradict another belief. Contrast this viewpoint with evidentialism which is (arguably) hindered by the fact that evidentialists need to do a lot of work to determine what makes one piece of evidence stronger than another. Despite the simplicity of coherentism, some will likely say that it is implausible. To give a counterexample, remember Scooter with his Magic Eightball. Scooter did not have any contradictory beliefs; however, his belief that he will become rich and famous is clearly unjustified if he believed it because he listened to his Magic Eightball. Thus, it follows that we have counterexamples that indicate that coherentism is flawed. Furthermore, it is easy to come up with more ludicrous counterexamples. For example, imagine that Carla only has two beliefs. Her first belief is that the color blue is the one true deity, and her second belief is that she only exists in blue’s dreams. The coherentist would have to claim that Carla’s beliefs are justified despite how ridiculous they seem.

Before I move on, it is worth noting that both the evidence for a belief and that beliefs coherence with one’s other beliefs are intuitively important for justification. For example, it would be irrational to say, “I believe that Napoleon was 5’4, but I have no evidence for it.” Furthermore, it would be irrational to say, “I believe that Napoleon was 5’4 and I also believe that Napoleon was 5’10.” Thus, it seems like evidence and coherence are both necessary conditions for justification and it thus follows that we may not be able to define justification by merely appealing to one of them.

Externalist Theories of Warrant

Unlike internalist theories of justification, externalist theories of warrant are not directly related to how one ought to reason. Externalists are instead more concerned about defining knowledge in a manner that corresponds to our intuitions.[3] Externalists attempt to find a definition of knowledge that explains why we intuitively don’t ascribe knowledge to people in Getttier cases (such as Bertrand in the Russel’s Clock case). As said earlier, externalists contend that the differentiating condition between knowledge and true belief relies (at least partially) on something that happens outside of the believer’s mind. In this section I will cover two different externalist theories that each claim that warrant is determined by some unique differentiating condition.

The first option we have for a warrant condition is reliabilism (which I talked about earlier). The reliabilist claims that a belief is warranted if the believer formed the belief via a method that is reliable. It is important to emphasize that the reliabilist is concerned with the reliability of the method as defined in the actual world rather than the method as described by the believer. What do I mean by this? Remember the Russell’s Alarm Clock case. In that case the (unreliable) method can be described as believing an unreliable clock (rather than just believing what is written on the clock). Of course, Bertrand does not know that the clock is broken, but his knowledge of that is irrelevant to whether or not the belief is reliable. In other words, the reliabilist defines the method from the point of view of the external world (how the clock actually is) rather than the point of view of the believer (how the clock looks). Reliabilists are thus able to avoid the Gettier problem since the reliability of the method is necessary for warrant (and thus knowledge), and since the method is defined from the view of the external world. If we view a Gettier case from this external vantage point, then we will see that the believer came to justified true belief in a lucky (and thus unreliable) manner.

Reliabilism is a simple externalist theory (at least relative to other options) which easily avoids Gettier cases. This makes it an attractive option for many externalists. Unfortunately, reliabilism has two major issues that are difficult to address. The first issue is that there are many possible cases where a belief is formed by a reliable method from an external point of view but also formed irrationally from an internal point of view. Remember the Scooter and the Magic Eightball example from earlier. The Eightball itself is reliable (since it is the voice of God); however, it is apparent that Scooter formed his belief in an irrational manner which makes it intuitive to say that he did not achieve knowledge.

The second major issue for reliabilism has been written about by philosophers such as Feldman and Conee (1998) and is referred to as the generality problem.[1] The generality problem occurs since there are often multiple ways in which a belief-forming method can be accurately described. For example, when we are talking about Scooter and the Magic Eightball, we could reasonably say that Scooter formed his belief via the method of believing God’s Magic Eightball. However, we could also reasonably say that Scooter formed his belief via the method of believing the Magic Eightball and not considering whether this was a good method. Both of these descriptions are accurate. If we describe the method in the first manner, then it is apparent that the belief has to be warranted (according to the reliabilist). However, the second manner is likely unreliable. Thus, it is apparent that you can describe the belief-forming method in multiple different true ways and get different results as to whether or not the relevant belief is warranted. This seems like a problem.

The second option we have for a warrant condition can be referred to as sensitivity and it is derived from (part of) a theory of knowledge created by the philosopher Robert Nozick in his book Philosophical Explanations (179). The sensitivity theorist claims that a person cannot achieve knowledge if they would have the same belief in the most likely scenario in which that belief is false. To reference an earlier example, Bertrand walks into the room and believes that it is 12:30 pm since the (broken) clock says that it is 12:30 and the sun is out. However, let’s imagine the most likely scenario to the one at hand where it is not 12:30 pm. In this similar scenario, Bertrand walks into the room at 12:29 pm or 12:31 pm and then reads the clock and believes that it is 12:30 pm (since the clock reads 12:30 still). Thus, it follows that Bertrand would hold the same belief if that belief were false which makes it insensitive and thus unwarranted and unknown (which thus avoids the Gettier problem). Before I move on it is important to note that the most likely scenario where the relevant belief is false is not determined by our understanding of probability, instead it is derived from the scenario which would actually be most likely to happen as judged from some external vantage point.[6]

Sensitivity provides us with a clever and somewhat intuitive response to the Gettier problem which avoids the issues associated with reliabilism; however, there are unfortunately prevalent counterexamples to sensitivity that dissuade most externalists from adopting it. Consider the following example:

Dave and the Matrix: Dave is a professor and—like all good professors—believes that his students are real people who are actually in the room with him while he lectures. Because Dave believes that his students are actually in the room with him, he thus believes that his students must be real physical beings and not in the Matrix (the computer simulation in which people are trapped in the movie).

If Dave’s beliefs were sensitive, then it follows that Dave would know that his students were in the Matrix in the scenario where they are. However, if you’ve seen the movie, then you know that is not the case (people don’t realize they are not in the Matrix). Thus, if Dave wants to be logically consistent, then he would have to say that he does not know his students are in the room with him since he doesn’t know whether his students are actually in the Matrix or not (since you clearly can’t be in the room and in the Matrix). Thus, it follows that either Dave doesn’t know that his students are in the room (which he probably should) or the laws of logic seem to break. This is problematic for the sensitivity theorist.[2]

I often find that students find externalism rather hollow. After all, epistemology is a normative field, and it seems like a warrant condition which is inaccessible in a person’s mind cannot be used as a guiding condition to determine what one should believe. For example, if I contend that the actual causal history of a belief—which is hidden to me—determines whether I know that belief, then it remains unclear what I should believe. Afterall, I would be uncertain about the causal history of that belief. I understand this concern; however, it is important to note that externalists can still hold that there are norms that help us achieve knowledge. For example, I may hold that the causal theory is true, and I should thus reason in a manner that generally leads me to knowledge (as defined by the causal theory). I may not know that I know the causal history of every belief, but I should still reason in a manner that generally allows me to uncover the causal history. Thus, it follows that externalists are not completely shunning normativity from epistemology, they are merely distancing the connection between normativity and the relevant differentiating condition (between knowledge and true belief).

Final Thoughts

There are advantages and disadvantages to internalism and externalism, which makes it difficult to decide which theory you should adopt. The same holds for specific justification and warrant conditions. It is however important to note (again) that internalism and externalism are not necessarily mutually exclusive. You could reasonably believe that knowledge requires a belief that is true, justified, and warranted. Many epistemologists hold that this combo-approach is overly complicated; however, many may say that it is intuitive to some degree. Some may say that externalist theories of warrant help us avoid lucky knowledge that is caused by factors from the external world (clocks that are broken, etc.) while internalist theories of justification help us avoid unjustified knowledge caused by intuitively bad reasoning (Scooter believing his divine Eight Ball, etc.).

Comprehension Questions

Reading Suggestions

Primary Sources

Goldman, Alvin, “What Is Justified Belief?”
Levy, Steven, “Defeasibility Theories of Knowledge”
Feldman, R. and Conee E., “The Generality Problem for Reliabilism”

Secondary Sources

Feldman, Richard, Epistemology, chapters four and five
Poston, Ted, “Internalism and Externalism in Epistemology”