Eliminative materialism

II. Why Propositional Attitudes Can’t Be Eliminated

Eliminative materialism is a frequently misunderstood position according to which there are no propositional attititudes. Its primary advocates have been Paul and Patricia Churchland. If would be a mistake to say, as some commentators have, that eliminative materialism is the view that there are no mental states. Nor, at least in some significant sense, can it be said that eliminative materialists deny the existence of intentionality. What I have described earlier as simple representation will certainly not be denied by eliminative materialists. What the eliminative materialist denies is the existence of propositional attitudes. These would include believing a proposition, doubting a proposition, fearing that a proposition is true, desiring that a proposition be true. So it is true that eliminative materialist claims that there are no beliefs.

To be fair, the eliminativist position is somewhat more complex than that. Eliminativism maintains that “belief” and “desire” are not mental states we are directly aware of, as “seeing red” or “feeling sick” would be, but are posits of a theory called “folk psychology.” In the history of science, “folk” theories have been succeeded by scientific theories. Sometimes the scientific theories absorb the “folk” theories in such a way that the “folk” theory is taken to be fundamentally right; just standing in need of some development by the scietific theory. In other cases, such as the move from Ptolemaic astronomy to Copernican, the succeeding theory showed the previous theory to be dead wrong, and the posits of the theory to be nonexistent. The Churchlands maintain that when neuroscience “looks under the hood” of the brain it will not find objects in it corresponding to “belief” and “desire.” Hence the right thing for science to do given this state of affairs is to deny the existence of beliefs and desires in much the way present-day science denies the existence of phogiston and ether.

The self-referential rebuttal is pretty obvious. “Come on Paul, you expect me to believe that, Paul?” Or, we could even present an argument that if eliminative materialism were true, no one could possibly know that it was true.

1. Knowledge is justified, true, belief (plus maybe a fourth condition).

2. If eliminativism is true, then no one believes that eliminative materialism is true, since there are no beliefs.

3. Hence, if eliminativism is true, no one knows that eliminativism is true (consequence of 1 and 2).

Here the Churchlands would reply that our standard definitions of knowledge are, of course, laden with folk-psychological assumptions, and when those are overthrown and a new theory based on neuroscience is developed, a fully adequate conception of knowledge will emerge.

Now the promise of successor concepts seems to many people to be, at best, a huge promissory note drawn on future science, and we are told very little about that the successors are actually going to look like. The successor concepts are going to have to do everything for us that we thought propositional attitudes did, except that these will be a more neurophysiologically accurate way of talking about human behavior and will not be propositional states.

Now propositional attitude psychology does a lot of work for us, in everyday life, and in science as well. Lynne Baker makes this point:

Suppose I dialed your phone number and said “Would you join us for dinner at our house on Saturday at 7:00?” You replied “yes.” On Saturday, I act in the way I should act if I believed that you were coming to dinner. But if neither of us had any beliefs, intentions, or other states attributed by “that”-clauses, it would me amazing if I actually prepared dinner for you and if you actually showed up.

Consider the whole practice of political polling which is very often able to predict the outcome of elections before they occur. Pollsters ask respondents who they intend to vote for, or who they believe is best equipped to deal with health care or terrorism.

What is most critical, however, is that if science is what every naturalist I know says that it is, a rational method for discovering the truth, then it we have to be able to know the precise content of the terms and concepts we are using. This is especially true in the area of mathematical reasoning, which is at the heart of physics. We have to be adding, not quadding. The definite integral has to be definite if it is to do the job assigned to it. There has to be some state of the person that recognizes the mathematical content of, say, Maxwell’s Equations (which to me is the propositional attitude of understanding that p), and if there has to be such a state, why should we not call this a propositional attitude.

It seems to me that there is an introspectively accessible state of knowing what one means when one says something. Now it may be that the full and complete content of what we know when we say it is not known to us. For example, I can say “I want a glass of water” without having any idea of the exact chemical composition of water. But there has to be an internally accessible content of the term “water” which will allow me to recognize whether I have been given a glass of water or a glass of coke. Of course there can be errors here, if it turns out that “What he thought was H₂O was H₂SO₄.” But one might be tempted to think that sulfuric acid was water, but it would be unlikely to be tempted by the likelihood that Coca-Cola is water, because Coke doesn’t look at all like water, but sulfuric acid sort of does. All of which suggests to me that we do have internally understood concepts of what we mean by words, and if we didn’t we wouldn’t be able to get through life. I don’t see how you can accept the existence of internally understood concepts of what we mean by words without also accepting propositional attitudes. I also fail to see the possibility that further brain-mapping is going to change this situation. This seems to me to be an insuperable difficulty for eliminative materialism.