Jonathan Dancy is a 20th century British philosopher who has written on ethics and epistemology. In his text, An Introduction to Contemporary Epistemology, Dancy offers us three arguments that the skeptics offer us to prove the uncertainty of knowledge. These are ‘Brains in Vats Argument’, ‘The Argument from Error’ & ‘The Arguments from Experience’.Skepticism is a philosophical school of thought which believes that we cannot know anything for certain and thus, tries to de-ground the philosophical branch of epistemology. Knowledge does require a justification and it is skeptics who believe that it is not possible for us to have an adequate justification to ascertain any sort of knowledge.

Dancy begins ‘The Argument from Error’ by stating how there have been instances when I have felt confident about my judgement and I have ended up making an erroneous one, like in the example of an arithmetic problem. In case of the present situation, Dancy believes that there can be nothing that I can point out that makes it possible for me to not be mistaken. For all we know, the present situation might be similar to the ones where I’ve made mistakes. 

This particular argument relies on an epistemological version of the principle of universalizability, popular in the field of ethical theory. This argument states that a particular judgement about an action being morally acceptable is universalizable only when one automatically becomes committed to hold the position that any relevantly similar action is also morally acceptable. When a new and relevantly similar situation arises, one must either call it morally acceptable or take back the judgement that the first action was morally acceptable. An action is said to be relevantly similar if it too has the properties when led someone to make the judgement they made in a previous case. Therefore, this principle states that unless and until there is a difference, we must make the same judgement again.

Suppose, I claimed yesterday that my mother would bring me chocolates from the market in the afternoon on the basis of her mood, her checking the slot in the refrigerator where the chocolates are usually kept, etc. But my judgement turned out to be incorrect. At the time when I made the judgement (in the morning), the fact that my mother won’t bring me chocolates was evidence transcendent to me (I had all the evidence for me to claim that she would but she did not; there are properties that might be absent even though one might have the best possible evidence for them to be present), like all future claims are. If on similar grounds, I claim to know that she shall bring me chocolates today afternoon, I can make this claim only when (a) I continue to assert that I knew that she would bring them yesterday which is absurd because she did not or (b) I say that I did not know that she would not bring them yesterday. Then how can I say that I know that she would, today? (c ) I offer a justification that the only difference  in my stand between the two cases, that on the basis of evidence 1 & 2 I did not know yesterday but I know today, is if my mother actually brings me chocolates today afternoon when she did not yesterday. But that she will bring me chocolates in the afternoon is evidence transcendent is to me, in the morning. Since I cannot use any of the above mentioned justifications, I have to accept that I did not know yesterday and this prevents me from making a claim about having knowledge today.

We can defend the argument from another person’s point of view by saying that if not me then another person might be justified in claiming that while I did not know yesterday, I do know today despite there being no difference between the two cases. The skeptics reply would be simple- if I myself am not certain about knowing something to be the case then how can an outsider claim to know that something on my behalf.

The conclusion, for Dancy, seems to be that if I have once wrongly claimed to know p then I cannot ever claim to know p again unless and until there is a relevant difference between the two cases. And no one else can claim to know on my behalf for all we know, I might not know p in the second case too.