Sextus Empiricus presents us with a certain problem. This is the problem of criterion.

We find that we have two typical questions, which seem to be related, yet lead us to problems.

(1.) What do we know?

(2.) How do we know?

For example, I can say that I know that ” there is a computer screen in front of me”. This answers question number one. However, now I will be asked question number two, ” How do I know that there is a computer screen in front of you?” For, knowing that there is computer screen does not answer how I know it. However, it does not stop here. In order to know something, I supposedly have a criterion for how I know it. But now there is even more trouble. I can say that I know this is the criterion.

However, now I will be asked, “How do you know that criterion?”. Thus, I will, eventually, have to have a criterion, and I will need a criterion to support that criterion. I need this other criterion because I cannot use that criterion to support that criterion in question. For to use the criterion in question to support the criterion in question is to beg the question. It is to argue in a circle. But if I need a criterion for the criterion, then I will need a criterion to support the criterion I invoked to help the other criterion, or I will again be arguing in a circle. Thus, we can keep doing this, and we would have to argue ad infinitum. We would have an infinite regress. Thus, we either have to beg the question or go with an infinite regress.

Either we know something or do not know something. If we know something, then we need a criterion. If we do not know something, then we do not need a criterion. Thus, either we have a criterion to know something or we do not have a criterion to know something.

Either we have a criterion to know something or we do not have a criterion to know something. If we do not have a criterion to know something, then we beg the question. If we do have a criterion to know something, then we need a criterion for our criterion and ad infinitum. In short, either we beg the question or we are left with an infinite regression.

So we find that we come into trouble when we make a claim about what we know. For once we make this claim, we either just take what is said or ask someone to back it up with some criterion for how they know it. Once they make the claim for how they know it, then we ask them to back up that claim with a criterion, since it, in short, becomes another “what we know” and we want to know “how they know” what they know.

A short explication of this problem is given by Roderick Milton Chisholm, and goes as follows: “To know whether things really are as they seem to be, we must have a procedure for distinguishing appearances that are true from appearances that are false. But to know whether our procedure is a good procedure, we have to know whether it really succeeds in distinguishing appearances that are true from appearances that are false. And we cannot know whether it does really succeed unless we already know which appearances are true and which ones are false. And so we are caught in a circle.”

The argument is also given by Sextus Empiricus in two places. The first follows like this: “This dispute, then, they will declare to be either capable or incapable of decision; and if they shall say it is incapable of decision, they will be granting on the spot the propriety of suspension of judgment, while if they say it admits of decision, let them tell us whereby it is to be decided, since we no accepted criterion, and do not even know, but are still inquiring, whether any criterion exists. Besides, in order to decide the dispute which has arisen about the criterion, we must possess an accepted criterion by which we shall be able to judge the dispute; and in order to possess an accepted criterion, the dispute about the criterion must first be decided. And when the argument thus reduces itself to a form of circular reasoning the discovery of the criterion becomes impracticable, since we do not allow them to adopt a criterion by assumption, while if they offer to judge the criterion by a criterion we force them to a regress ad infinitum. And furthermore, since the demonstration requires a demonstrated criterion, while the criterion requires an approved demonstration, they are forced into circular reasoning.”

And here is the second explication of the problem by Sextus Empiricus: “In another way, too, the disagreement of such impressions is incapable of settlement. For he who prefers one impression to another, or one “circumstance” to another, does so either uncritically and without proof or critically and with proof; but he can do this neither without these means (for then he would be discredited) nor with them. For if he is to pass judgment on the impressions he must certainly judge them by a criterion; this criterion, then, he will declare to be true, or else false. But if false, he will be discredited; whereas, if he shall declare it to be true, he will be stating that the criterion is true either without proof or with proof. But if without proof, he will be discredited; and if with proof, it will certainly be necessary for the proof also to be true, to avoid being discredited. Shall he, then, affirm the truth of the proof adopted to establish the criterion after having judged it or without judging it? If without judging it, he will be discredited; but if after judging, plainly he will say that he has judged it by a criterion; and of that criterion we shall ask for a proof, and of that proof again a criterion. For the proof always requires a criterion to confirm it, and the criterion also a proof to demonstrate its truth; and neither can a proof be sound without the previous existence of a true criterion nor can the criterion be true without the previous confirmation of the proof. so in this way both the criterion and the proof are involved in the circular process of reasoning…”