November 16, 2008

This Post is Old!

The post you are reading is years old and may not represent my current views. I started blogging around the time I first began to study philosophy, age 17. In my view, the point of philosophy is to expose our beliefs to rational scrutiny so we can revise them and get better beliefs that are more likely to be true. That's what I've been up to all these years, and this blog has been part of that process. For my latest thoughts, please see the front page.

Three Varieties of Certainty

'Certainty,' whatever that is supposed to be, would certainly (!) seem to be important in epistemology. Like a lot of important words, it frequently gets thrown around without definition. I know of at least three totally distinct ways of using this term, and the only thing they all seem to have in common is 'very high epistemic status' - i.e. something is certain if we really know it, in some way that is 'better' (more certain!) than ordinary knowledge. I'm going to outline here these three different varieties of certainty.

Cartesian Certainty (also called 'demon-proof certainty') is attributed to a belief when I could not be wrong about that belief even if all of my sensory input was caused by a malicious entity (the 'Cartesian demon') whose sole purpose was to induce me to believe falsely. In other words, Cartesian certainties are truly infallible. Good candidates for this status are the claim that I exist and simple arithmetic truths. However, if some strong form of semantic externalism is true, it may undermine all Cartesian certainties, since it may actually turn out that I have no beliefs, or that my beliefs have no content, depending on the details of the theory. (On this, see Robert Fogelin, Pyrrhonian Reflections on Knowledge and Justification, pp. 186-188.)

Moorean Certainty is attributed to a belief when I am (rationally) more confident of this belief than I am of the soundness of any argument against it. This, I take it, is the kind of certainty Chisholm means to define when he says: "p is certain for S =Df. For every q, believing p is more justified for S than withholding q, and believing p is at least as justified for S as is believing q." (Roderick Chisholm, Theory of Knowledge (3rd ed.), p. 12, as quoted in Fogelin, Pyrrhonian Reflections, p. 128) If you find this definition rather difficult, you are not alone. If you stare at it for a while, it might start to make sense, but, if not, don't worry. At any rate, Moore's famous claim to Moorean certainty was for the proposition 'I have two hands.' Another claim of certainty he made was for the proposition 'the earth existed for many years before my birth.' One simple way of understanding Moorean certainty is to say that, no matter what premises are involved in a valid argument against the claim, one ought (rationally) always to perform a 'Moorean shift' - that is, if someone gives us an argument of the form 'p, therefore you do not have two hands,' we should respond, 'but I do have two hands, therefore not-p.'

What I shall call Wittgensteinian certainty (I am hesitant to claim that it is a correct interpretation of Wittgenstein's later works taken as a whole), is attributed to a belief when it makes no sense in a given context to question that belief. I call this 'Wittgensteinian' as a result of passages like this: "Imagine a language-game 'When I call you, come in through the door.' In any ordinary case, a doubt whether there really is a door there will be impossible." (On Certainty 391) A certainty in this sense is simply a statement which for some practical reason must be taken for granted.

It seems to me that there are, at least arguably, real instances of all three of these forms of certainty. It also seems to me that all three of them are philosophically interesting. Are there more?

Posted by Kenny at November 16, 2008 6:06 PM
TrackBack URL for this entry:


Chisholm's definition can't be right if it's supposed to be just that you're rationally more confident of the belief than of any argument against it. Unless he further specifies q, then q can be anything, including anything you can have Cartesian certainty about. That means you're not certain of p unless you can be at least as rationally confident about p as you are of any other belief you've got. It's a belief more rationally confident than which you have no other beliefs (although ties are allowed).

I suppose the ancient skeptics could satisfy this criterion pretty easily, since admitting their first belief (since they have none) would automatically qualify. But for the rest of us, it's very hard to satisfy this. Once you perform the cogito, you won't be certain of anything according to this definition unless you're as confident of it as you are of your own existence.

Posted by: Jeremy Pierce at November 17, 2008 2:14 PM

Jeremy - yes, you do seem to be correct. So 'Chisholmian' certainty should probably be its own category. It isn't really Cartesian, since the definition doesn't include infallibility, and I was wrong to classify it as Moorean for the reasons you cite (it is much too strong).

Posted by: Kenny at November 17, 2008 8:01 PM

Post a comment

Return to