Sean (a) claims that strong atheism is incoherent, because strong atheists are materialists, and intentional states are immaterial. Therefore, strong atheism is self-refuting.
My suspicion is that Sean is paraphrasing a similar argument made by Edward Feser, the Catholic philosopher of religion (b). Let me propose a more formal, and less easily rejected expression of Sean's argument:
Eliminative materialists hold the view that we can only accept as real that which can be shown to be ultimately determined by physics.
Strong atheism leads to eliminative materialism.
Beliefs, desires and other propositional attitudes have non-material existence.
Because of 1, eliminative materialists do not consider beliefs, desires and other propositional attitudes to have real existence.
But, eliminative materialists have beliefs (for example, as in 1).
Eliminative materialism, and hence strong atheism, is self-refuting on the grounds that in order for it to be true, you have to believe it, but they reject the reality of beliefs.
Most atheist commentators on his article have pointed out the problem with step 2. It is of course not true that strong atheism leads to eliminative materialism. But as it turns out, I hold the view that strong atheism together with scientific realism do ultimately lead to some form of eliminative materialism. This view can be rightfully challenged, but it is not my intention with this article to elaborate on that specific point. So let's move on.
Point 3 is probably true. This claim was first made by Franz Brentano (c), who claimed that propositional attitudes cannot be reduced to material processes. I am not aware of any serious challenge to Brentano's irreducibility claim.
Point 4 is true, according to the eliminative materialist - they deny the real existence of propositional attitudes. They claim that our experience of beliefs and desires are illusory.
Initially, this sounds shocking. How can anyone claim that a common sense, everyday concept such as having a belief is not real? To follow this argument, we must look at the ontology of belief – does holding a belief have some kind of real meaning? If we say somebody holds a belief, is there some deeper fact of the matter that we can obtain knowledge of?
Putnam (d) came up with an interesting thought experiment, of which I will help myself with a simplified version from Dennet's “The Intentional Stance” (e) to illustrate the point. Imagine a chain of bars called Sharky's. Every Sharky's bar looks exactly the same inside, as specified by strict franchise rules – it has a bar area, a few pool tables and some bathrooms at the back. There are tens of these Sharky's bars all over Gauteng.
Three friends (John, Jack and Pete) go out to the Sharky's in Randburg. While busy consuming their first beer, John goes to the rest rooms. In a moment of delinquency, he decides to scratch his initials on the back of the toilet door.
While he is busy vandalising the rest rooms, Jack and Pete, equally delinquent, decide to play a trick. They add a sleeping drug to John's unfinished beer. John comes back, takes a big swig and immediately falls into a deep sleep. Jack and Pete quickly settles the bill, drags the sleeping John to their car, and drive off to the Sharky's in Midrand. There they put the sleeping John back in the same seat, with an unfinished beer of the same level as it was when they left Randburg.
John wakes up, and still thinking that he is Randburg, tells Pete and Jack that he has scratched his initials on the back of the toilet door. Pete and Jack shake their heads deviously, not believing that John is telling the truth. To settle the issue, they decide to make a wager – they will go with John back into the rest rooms, and if John cannot produce the initials, then all drinks are on him.
Now let's pause everything at this point. Let's evaluate John's belief at this point in time. Can we say his belief, that he scratched his initials on to the toilet door at Sharky's, is true or false? That depends. If John's belief is that he scratched his initials on the door at Sharky's in Randburg, it is true. If John's belief is that he scratched his initials on the door of the Sharky's he is in right now, then it is false. But in John's head, that is exactly the very same belief. He has no inclination or clue that he is not where he thinks he is. Even though we know all the facts of the situation, we cannot determine the value of the belief in John's head - it has no deeper fact of the matter.
A similar problem with propositional attitudes is called the opacity problem. You cannot simply replace the objects of which the proposition is about, with different terms, even if they have the same numerical identity, and still maintain the value of the original attitude. Sherlock Holmes thinks Catherine's mother is the murderer. Sherlock Holmes do not think the countess of Devon is the murder. But Sherlock Holmes is unaware that Catherine's mother is the countess of Devon. Once again, there is no deeper factual meaning to Sherlock's Holmes' belief as to who the murderer is.
That is why beliefs, desires and propositional attitudes cannot be merely physical brain states. They do not exist independently of the bigger environment. They require contextual information, to which the holder of the belief is not necessarily privy to, to be evaluated.
Does this mean that it is impossible to provide an account for how we experience propositional attitudes using our brains? No, not at all. One way to overcome this is to adopt notional attitudes. Simply put, notional attitudes are those part of belief states that do exist in the brain. John, in Sharky's, have a notional attitude that he both scratched his initials on the bathroom in Randburg and in the bathroom of the Sharky's he is in right now. There is no psychological paradox here – we can easily put our shoes in John's and understand what he is thinking, even if we understand the evaluation of that thought to have different meanings in different contexts.
So now that we have this technical matter out of the way, we can re-evaluate Sean's claim. Yes, eliminative materialists acknowledge that beliefs have non-material existence – but they don't need beliefs to exist in this fashion in order for their view to be coherent.
It is also now obvious that 4 and 5 begs the question. These statements presuppose the reality of belief, without providing an account for how eliminative materialists consider belief ontology. It presupposes that our common sense attributions of beliefs are real and correct.
Paul Churchland addresses this same argument (f). In the Middle Ages, alchemists believed that living matter was separate from non-living matter due to the presence of vital forces. Now, if a modern chemist were to go back in time and stumble on an alchemist and claim that there is no such thing as vital forces, the alchemist could easily respond:
“Well, that's just silly. You can't claim that you don't have vital forces, because then you would be dead.”
Hopefully, I have shown here that eliminative materialism is not at all refuted by this line of attack. I also hoped to have shown that the ontology of belief is a more complex topic than what our common sense tells us it is.
f) Eliminative Materialism and the propositional attitudes: http://www.jstor.org/discover/10.2307/2025900?uid=3738032&uid=2&uid=4&sid=21102209935877