The problem of the criterion

I recently read "The Problem of the Criterion" by Roderick M. Chisholm. I first learned about this problem from J.P. Moreland a couple of decades ago. J.P. learned about it from Roderick Chisholm, though, and this was the first time I went to the source. I like the fact that I've now had it explained to me by two different people in two different ways. I want to write about it today because I've been thinking about it a lot and feel the need to get it off my chest. Also, writing about stuff you learn helps to make it stick. That is unless you have a misunderstanding about it, in which case the wrong thing gets stuck.

The problem of the criterion is a really general problem in epistemology. On the one hand, it seems like before we can sort the true things from the false things, we first need some method or criteria by which to test or determine whether something is true or false. But on the other hand, it seems like before we can distinguish between the good methods and criteria and the bad methods and criteria, we have to know which ones reliably give us truth and which ones don't. This creates what Chisolm calls a wheel. It seems to result in circular reasoning. We know such and such is true because it meets our criteria, and we know we're using the right criteria because it always delivers the truth.

It seems, on the surface, to be an inescapable problem, and there are only three possible ways out of it. According to Chisolm, none of the escapes are particularly satisfying, so it really just comes down to picking the lesser of three evils.

One way out of it is just to throw up your hands and say we don't have knowledge. The wheel is inescapable, and we might as well give up knowing anything. Let's call this way skepticism. J.P. calls it global skepticism--the position that we don't have any knowledge about anything.

The problem with global skepticism is that it can't be justified. It undermines itself. A person presumably reaches the conclusion that we have no knowledge because they've recognized the problem of the criterion. Skepticism turns out to be the conclusion to a line of reasoning consisting of premises and inferences. The reason this kind of skepticism is self-defeating is because one can't justifiably conclude that skepticism is true unless they could know that their premises were true and that their inferences were valid. Even if we allowed them their premises and inferences, they couldn't claim to know skepticism was true because that would be self-refuting. It would essentially be claiming to know that there's no knowledge.

Besides that, if we're just reasonable people, and if we're honest with ourselves, we're going to admit that we know at least a few things. I know I exist, that my cat is a picky eater, that the sun will rise tomorrow, and that drinking soft drinks is bad for your health.

The other two escapes to the problem of the criterion involve breaking the circle. You can break the circle either by beginning with a method or criteria, or you can break it by beginning with what seem to be clear case items of knowledge. Either you know something is true, and then use that to figure out some good methods and criteria, or you start with methods and criteria in order to find out what's true.

Let's call people who begin with methods and criteria methodists, and let's try not to confuse them with the Christian denomination by the same name. The problem with methodism is that it leads to an infinite regress. If you can't know anything unless you first apply some criteria or use some method to discover it, then you can't know the criteria or method either unless you first apply some criteria or method. Methodists think that before you can know anything, you first have to be able to account for how you know it. If a methodist claims to know something, one can ask, "Well, how do you know that?" To be consistent with their position, they're going to have to say they know it because it meets some criteria or because it used some method to test it or whatever. They'll have to give a reason to account for their knowledge claim. But then a person can ask them, "But how do you know that?" And the methodist will have to account for it in the same way. You should be able to see how this leads to an infinite regress. You can just keep asking, "How do you know?" forever, and the methodist will be obliged by their point of view to offer reasons. There's no starting point, so methodism leads inevitably to global skepticism, which we've already talked about.

The third option is to begin with what seem like clear case items of knowledge. Let's call these people particularists since they begin with particular cases of knowledge. While this position seems troublesome, too, it's the least troublesome of the three. As I said, it does seem to pretty much all of us, that we know at least some things. This is the position I hold because not only is it the most reasonable of the three positions, but I think it's the position almost everybody uses anyway. It's how knowledge actually works in practice.

Consider somebody who sees a cat sitting on their bed. They immediately form the belief that there's a cat on their bed. While this belief may be the result of an underlying prior belief in the reliability of their sensory perceptions, people don't consciously reasoning from that prior belief to the belief in the cat. Nobody explicitly thinks,

If I see something, then it must be there.
I see a cat.
Therefore, there must be a cat there.

If we begin with modest claims of knowledge about mundane things, we can use that knowledge to work out methods and criteria by which we can expand on what we know. Most people assume that the future will resemble the past, for example, and they apply this principle automatically without even being consciously aware of it. But if you take the time to think really hard about what justifies your belief that fire is hot, or whatever regularity in nature you've observed or learned about, you'll eventually discover that you were relying on the principle of induction or the uniformity of nature. Induction can then be used as a tool (i.e. a method) by which to learn about other things.

I think that is enough of a solution to the problem of the criterion to work for most people. Induction itself (or whatever other criteria or method you've discovered) can be considered an item of knowledge without necessarily having to know what justifies it. If you're a particularist, you've rejected the methodist principle that you have to be able to prove everything before it counts as an item of knowledge.

But that doesn't mean there isn't a justification for it. It only means you can know it without knowing how you know it. You can, if you like epistemology, think harder and try to discover what justifies belief in the uniformity of nature (or a host of other things like the reliability of your sensory organs, etc.). In doing so, you're going to be faced with another problem. It's a dilemma. Either you're going to get into an infinite regress or else you're going to hit the foundation.

An infinite regress might happen if you assume that for anything you know, it has to be inferred from something that's logically or epistemologically prior. This leads to an infinite regress because it means everything you know has to be preceded by something else you know. If that's the way things are, then knowledge is impossible. That leads again to global skepticism which is self-defeating.

The only way knowledge is possible is if there are at least some things we know a priori. In other words, there are some things we know that we do not infer from something that's logically or epistemologically prior. These items of knowledge form the foundation for everything else we know.

But it raises a thorny question. If these foundational items of knowledge are not arrived at by inference from prior items of knowledge, then what justifies them? Don't we typically think of justification in terms of some kind of inference? Well, yes, I think we usually do, but it can't be the case that we always do.

The only way I know how to explain what justifies our a priori knowledge is by pointing to examples of it and having you think through it with me. So let me use some examples.

Let's start with the cat on the desk. How do I know there's a cat on my desk? There are two reasons. First, because I can see it, and second, because I know my seeing is reliable. To make this illustration simple, let's talk about the first. I can see the cat. Now, how do I know I can see the cat, or at least what I take to be the cat? Notice that I don't base this on anything more than the immediate experience of seeing the cat. I just see it, and that's all. I'm immediately aware of my own sensory perceptions. I experience them immediately in the sense that I don't infer them. This is about as direct and basic as you can get.

Now let's look at something a little different. Let's consider a hypothetical. If I knew that Jim was taller than Bob, and Bob is taller than Dan, what could I infer about the height difference between Jim and Dan? Who is taller? Well, if you think about it, you should be able to see that Jim is taller than Dan. What I'm interested in here is not in how tall Jim, Bob, and Dan are or even in whether they exist. What I'm interested in is what justifies the inference. How do I know that if Jim, Bob, and Dan were related in the ways we're imagining that Jim would be taller than Dan? How do we know that it follows that Jim is taller than Dan just on the basis that Jim is taller than Bob and Bob is taller than Dan? Well, again, there isn't anything more foundational that we base this on. We base this on nothing more than our own careful reflection on it. As long as we understand the relation of "taller than," we can just "see" that it's true. We have a rational intuition about it.

Here's another similar thing. Suppose somebody says, "My cat is pregnant," and then turns right around in the next breath and says, "My cat is not pregnant." You might at first think he has two cats--one is pregannt, and the other isn't. Or maybe you think that between the first statement and the second statement, his cat must've given birth. Or maybe you think pregnancy is a metaphore being applied in two different ways. Or mabye one statement is literal, and the other is a metaphore. Notice that in all these cases, what you're trying to do is reconcile the two claims. Rather than jump to the uncharitable conclusion that the person is lying, you look for a way for it to be possible for both claims to be true.

Suppose you can't, though. Suppose after talking to the person, you realize he's talking about the same cat, at the same time, and in the same sense. At that point, you'd know he was lying. But what tells you he's lying? It's the fact that you know these two claims can't both be true at the same time and in the same sense. That's the law of non-contradiction. A contradiction (or at least an explicit one) is when one claim is the negation of another claim. The claim that my cat is pregant directly contradicts the claim that my cat is not pregnant as long as we're talking about the same cat being pregnant in the same sense at the same time. And we know that can't be. If one claim is true, than its negation must be false.

But how do we know the law of non-contradiction is true? Again, it isn't something you can infer from something prior. In fact, if you think about it, nothing you say can make any coherent sense at all unless the law of non-contradiction is already true. Unless your statement excludes its negation, it's a meaningless statement. So everything you say, if you're trying to be coherent, presupposes the law of non-contradiction. With that being the case, nothing can be logically priori to the law of non-contradiction. So how do we know it's true since any effort to justify it would seem to presuppose it and therefore beg the question?

Well, again, this is just something you have a rational insight about it. You can simply reflect on it, and once you understand what it is saying, you can immediately recognize that it's true. This is called knowledge by rational intuition.

I'm not going to go into it right now, but there's a handful of other things you can know in the same way. You can know some basic math, geometry, and logic simply by reflecting on it. The knowledge is immediate. You just "see" these things.

These are the sorts of items of knowledge that sit at the foundation of everything else we know.

That's about all I have to say. To wrap it all up, we don't have to necessarily know how we know something before we can justifiably consider it an item of knowledge. Most people aren't so reflective that they plumb the depths of what justifies every little thing they claim to know. We just walk around claiming to know basic things because we observe them, experience them, learn about them, or whatever. But if we do plumb the depths, then we discover everything we know rests on a foundation of knowledge that cannot itself be demonstrated to be true. The foundational items of knowledge can only be intuitively recognized to be true. And it's this rational intuition--this "seeing"--that justifies our foundational items of knowledge.

Since nothing can be more certain than the premises upon which it is based, and everything is ultimtely based on unprovable premises, it follows that the things we can know with the greatest degree of certainty turn out to be things which cannot be demonstrated to be true. This refutes any kind of evidentialism, empiricism, or scientism.