critique-of-pure-reason.com

08 October 2010

Imploding the Demon: Crispin Wright’s Response to Scepticism, Part 2

Click here for Part 1

To recap, the first version of Wright’s sceptical formulation looks like this:

(P1) ¬Rx [¬Dx]

(P2) Rx [Rx[P] → ¬Dx]

¬Rx[P]

There is no available warrant for believing a proposition based on perception such as 'I have a hand in front of me.’

Faced with this argument, we might ask if (P1) really is true. Is it really the case that I am never warranted in believing that I am not dreaming? Maybe there is some way that Descartes can establish that he is not dreaming. Stroud addresses this question:

If he could find some operation or test…that indicated to him that he was not dreaming, perhaps…he could know that he was not dreaming.

Wright also looks at this idea, and similarly finds that it is problematic. The problem is in establishing that any such procedure or test has been executed properly: I have no more warrant for believing that it has, than I have for believing I am not dreaming, because I might be dreaming.

This is what Wright calls the Proper Execution Principle:

If the acquisition of warrant to believe a proposition depends on the proper execution of some procedure, then executing the procedure cannot give you any stronger warrant to believe the proposition in question than you have independently for believing that you have executed the procedure properly.

I have no more warrant for believing that I am not dreaming than I do for the belief that I have properly executed the procedure to acquire that warrant – that I did not just dream that I executed it properly. If I am warranted in believing that I properly executed the procedure, then I must already have the warrant to believe I am not dreaming, which is exactly what I am trying to get with the procedure. Therefore, there can be no procedure that establishes that I am not dreaming, so (P1) looks good.

The Ultimate Sceptical Paradox: Second Version

At this stage Wright has set out the paradox in terms of perception, but now he goes on to apply exactly the same reasoning to intellection, which is reasonable to suppose is in the same boat as perception, in that it is excluded by dreaming. This move is no doubt motivated by the suspicion that the sceptical position presupposes reasoning ability that would be unavailable while dreaming. Perhaps we have already found the answer to the sceptic! Let’s see…

Q = a proposition that requires a warrant from competent intellection if it is to be warranted at all

The argument goes as before, but with Q substituted for P:

(P1) ¬Rx [¬Dx]

(P2*) Rx [Rx[Q] → ¬Dx]

And, by the same reasoning as before, the conclusion is

¬Rx[Q]

That is, you are not warranted in believing anything that can be warranted only through intellection. If the warrant required to believe a proposition depends on intellection, then you cannot get it.

Where does that leave us? Was the suspicion of implosion well-founded? Does the reasoning of the argument itself make use of a Q-type proposition? Yes: the thought expressed by the implication proposition in (P2*). Thus, substituting it for Q in the conclusion we get

¬Rx[Rx[Q] → ¬Dx]

which contradicts (P2*).

This is effectively a reductio ad absurdum from (P1) and (P2*), showing that these premisses cannot be true together. Since this new paradox is exactly parallel to the original paradox – that, in fact, (P2) is true if and only if (P2*) is true – then we can also conclude that (P1) and (P2) cannot be true together.

Remembering the second constraint, we have to ask what is actually wrong with the premisses. (P2) looks too strong to doubt: I know that if I know P then I know I’m not dreaming. It is in the nature of the concepts involved that this be the case, virtually by definition. So (P1) must be false, and I can be warranted in believing that I am not dreaming. But we have already shown, with the Proper Execution Principle, that (P1) is true: we cannot get the warrant we need.

This is where Wright suggests that we can have a warrant for a belief even if there is no possibility of carrying out a process to acquire that warrant. Such beliefs are Wittgenstein’s hinge-propositions, the basic propositions that it makes no sense to doubt, and on which all of our other beliefs hinge. They might also be seen as the bedrock or ultimate foundation of our knowledge.

Thus, the hinge-proposition here is

Rx[¬Dx]

or perhaps simply

¬Dx

I am not dreaming.

But wait. Remember that we began to use intellection in place of perception because we assumed it was similarly excluded by dreaming. If this is not the case, we cannot infer the inconsistency of (P1) and (P2) from that of (P1) and (P2*), and the dream argument is safe. Intuitively, it seems quite reasonable to suppose that dreaming does not exclude intellection. In fact, to take the hypothesis seriously, that for all I know I might be dreaming now, I have to actually presuppose that I can reason competently when I’m dreaming.

But I think that the above result against the sceptic is the crucial form of the answer to the sceptical paradox. Wright wants to be able to make it stand, so he is compelled to press on to find a formulation from which he can more legitimately arrive at the inconsistency of (P1) and (P2) together.

The Ultimate Sceptical Paradox: Third Version

We can see that even though the reasoning ability that the sceptic presupposes might be possible while dreaming, he still presupposes being able to distinguish between two dream-states: competent intellection, and a state in which he is not reasoning competently. If the sceptic cannot do this, then his argument looks vulnerable.

With these things considered Wright reconstructs the sceptical argument with a view to finishing it off, roughly in the way that looked so promising above, by exposing the incompatibility of the sceptical argument with its assumptions. The next step seems clear. It seemed reasonable above to think that intellection could happen during dreams, so all we need to do now, to trap the sceptic in his corner, is define a psychological state in which we cannot perceive or competently reason. This will show that he is depending on knowing the difference. Wright calls it maundering.

x is maundering at t just in case x is then in a phenomenologically smooth state which, like dreaming, necessarily precludes the causal conditions for perception but, in addition, likewise precludes the causal conditions of competent intellection.

So, replacing dreaming with maundering, the argument ends up looking like this:

(P1**) ¬Rx [¬Mx]

(P2**) Rx [Rx[Q] → ¬Mx]

And, by the same reasoning as before, the conclusion is

¬Rx[Q]

Once again, you are not warranted in believing anything that can be warranted only through intellection.

This is stronger than the previous version, because we are certain that maundering excludes the possibility of intellection. As before, we find that (P1**) and (P2**) are inconsistent with each other, and we feel we ought to judge that (P1) and (P2) are correspondingly inconsistent. But does this really follow? By analogy we can say that based on parallel reasoning it cannot be established that (P1) and (P2) can be true together. This is a weaker claim than that they cannot be true together.

So here is what we have established: There is no warrant for the conclusion of the dream argument.

According to the third constraint, the second-order sceptic will reply that we still have no warrant to deny that we have no warrant to believe we are reasoning properly. He has not established it properly, but he doesn’t have to: we are still left with doubt.

However, it is still a good result. We have exposed the sceptic’s argument as unwarranted. His reasoning entails that he cannot get warrant for a belief based on reasoning, and thus his own reasoned conclusion is unwarranted. He is assuming what he is attempting to doubt: that one can have warrant for a belief based on reasoning; that one can tell that one is reasoning properly; that one is not dreaming.

The Second-Order Dreaming Argument

We are looking now for warrant to deny the sceptical argument, in particular to deny the conjunction of the two premises of the dream argument, (P1) and (P2).

With

Ax[P] = x has no warrant to deny P

We are faced with this premise:

Ax[(P1), (P2)]

That is, x has no warrant to deny that (P1) and (P2) are both true. Therefore:

Ax[¬Rx[P]]

That is, x has no warrant to deny that there is no available warrant for a belief based on perception.

To counter this, we need to show that we can get the inconsistency of (P1) and (P2) from the one we found in (P1**) and (P2**). Considering that (P2**) seems untouchable, we need to focus on (P1**). Here is how it could be done:

The material implication

Ax[(P1)] → Ax[(P1**)]

translated as

¬Rx[¬(P1)] → ¬Rx[¬(P1**)],

with a negated consequent

Rx[¬(P1**)]

by modus tollens gives us

Rx[¬(P1)]

and

¬Ax[(P1), (P2)]

x has warrant to deny that (P1) and (P2) are both true, which negates the premise above.

In this line of reasoning we can take the liberty of negating ¬Rx[¬(P1**)] on the assumption that (P2**) is true, because we previously discovered that (P1**) and (P2**) cannot both be true. We can deny the premise that we are not warranted in believing we are not maundering,

This is just what we want, so we need to support Ax[(P1)] → Ax[(P1**)]. This is saying that the sceptic, in making the dream argument, commits himself to making the equivalent argument for maundering; and since we can deny the maundering argument, we feel we should be able to deny the dream argument as well.

What would it mean to say that ¬Rx[¬(P1)] → ¬Rx[¬(P1**)] is false? It would mean that ¬Rx[¬(P1)] is true and ¬Rx[¬(P1**)] is false; x has no warrant to deny (P1) but does have warrant to deny (P1**); or, x has no warrant to deny that he is dreaming but does have warrant to deny he is maundering. How could this be? There is nothing in the way we are understanding these terms to make such a difference between the two states.

Remember the Proper Execution Principle:

If the acquisition of warrant to believe a proposition depends on the proper execution of some procedure, then executing the procedure cannot give you any stronger warrant to believe the proposition in question than you have independently for believing that you have executed the procedure properly.

To get a warrant to deny you are maundering you would have to get around this somehow. It is not plausible that the principle could apply to dreaming but not to maundering.

We previously established that

There is no warrant for the conclusion of the dream argument.

And now we have established positive warrant:

There is warrant to deny the conclusion of the dream argument.

We win!

This boils down to a problem with (P1): it is either unwarranted, or else we are warranted in denying it. Either way, we might wonder exactly how we can be justified in believing that we are not now dreaming. We have already seen a possible answer, which is that it is a hinge-proposition that gives us justification. Wright notes that if only we were more aware of the unquestionable fastness of this bedrock belief, we would never be troubled by the sceptical argument in the first place, because its crucial premise would be obviously false. In fact, the sceptical argument would never have occurred to anyone at all.

To me, this indicates a mistake on the part of philosophers such as Descartes, in taking scepticism too seriously. In his quest for absolute certainty, Descartes affected to doubt things that in fact were deeply indubitable. I am not dreaming, and there really is an external world.

Of course, as Wright points out, we have not actually established definite warrants for these propositions – and we might suspect that the task is impossible in principle – but we’ve done pretty well.

In summary then, the sceptical paradox “implodes” because it depends on abilities that it calls into question. The possibility that we are in fact dreaming is still left open, but there is no reason at all to believe it or any of the other sceptical paradoxes, because the arguments in their favour are not justified and we are entitled, logically, to deny them.

Wright, C. 1991: “Scepticism and Dreaming: Imploding the Demon”, Mind, vol. 100, pp87-116
Pritchard, D. 2001: “Scepticism and Dreaming”, Philosophia 28
Stroud, B. 1984: The Significance of Philosophical Scepticism, Oxford University Press
Descartes, R. 1996: Meditations on First Philosophy, Cambridge University Press

comments powered by Disqus