With eliminative materialism, we saw a reaction to dualism. Since folk psychological theories of mind got it so wrong (i.e., dualism), eliminative materialists argued we need to abandon everything related to folk psychology--including the concepts of belief and desire--and develop entirely new concepts that arise organically from modern scientific observations about the brain. It simply doesn't make sense to try to force millennia old pre-scientific concepts onto the findings of modern brain science. We should let the science develop from scratch--not impose old concepts onto its observations.
Responses to Eliminative Materialism
There have been several responses to eliminative materialism: The main one we're going to look at comes from Jerry Fodor, considered to be one of the most important living philosophers of science (and he doesn't even wear a labcoat!).
He uses 3 main arguments to make his case:
(A) Beliefs and brain-states are independent levels of analysis;
(B) Mental states should not be defined in terms of underlying physical structures instead they should be functionally defined. Defining mental states functionally means
(i) they are not reducible to physical structures
(ii) they can be instantiated by many different types of physical structures.
(C) There is no theory that can explain and predict behavior better than one that uses the concepts of belief and desire.
Ok, lets check out Fodor's arguments...
Different Levels of Analysis
His argument goes something like this: folk psychology and neuroscience/cognitive science are giving an account of the phenomenon (behaviour) at different levels of analysis: Just as we wouldn't study what characters are on a computer screen from the level of electrical impulses in computer's circuits, we shouldn't try to discover what beliefs are by looking at electro-chemical signals in the brain. Doing so is too fine grained for the purpose of a behavior-of-organisms level of analysis.
Another example would be geology. We wouldn't study tectonic plate movements at the atomic level. Furthermore, the concept of tectonic plates can't be reduced to one defining atomic structure. Many non-fundamental sciences require independent (non-reducible) levels of analysis. It follows that the concepts used in these sciences won't necessarily be reducible to particular atomic structures.
Folk concepts are at a much higher level of abstraction than neuroscience or some other fine-grained brain science. So, just because we can't find one to one "matches" between the concept of beliefs and what's happening at the neurological level doesn't mean there's no such thing as beliefs, it only means that the distance between the levels of abstraction is too great for any meaningful matching to occur.
Functionalism: How to Define Mental States
The Eliminativist Attack: Non-Reducibility
One of the arguments eliminative materialists (EM) use against the concepts of folk psychology such as beliefs and desires, is that they can't be reduced to basic physics--or at least physical states--and therefore should be eliminated from a scientific theory of mind. The idea is that, in science, things only exist if they can be reduced to physical phenomena. But it seems that this can't be done with folk psychological concepts.
Before we see why reduction might not work for folk psychological concepts, lets revisit an example of reductionism that does work in science. The classic example is that the concept "water" reduces to the physical structure "H20". There is a one to one correspondence with the concept "water" and the physical structure "H20". Any time someone refers to water, we can reduce that concept down to an underlying physical structure and that physical structure will always be the same. If you use the concept "water" for something that doesn't have the physical structure "H20", then you are misusing the concept.
Now lets think about a belief. Does a belief reduce down to a uniform physical structure that is the same across all brains? Very unlikely given that everyone's brain is slightly different and the vast number of types of beliefs one can have. What is it that every belief has in common in terms of the physical brain state that underlies it?
When you have the beliefs "it's raining", "apples are crunchy", and "judo is a great sport" what could the underlying structure of all these beliefs have in common? Probably nothing.
Given what we know about neuro-plasticity, there are many different ways the same belief, behavioral state, or cognitive process can be underpinned by a brain structure. Water, on the other hand, will always reduce down to two hydrogen atoms and one oxygen atom.
Consider another example: pain. Is there something common to the underlying physical structures of all types of pain that we can reduce the concept down to? Given that there are so many different types of pain (stubbing your toe, biting your tongue, burning your finger, heartbreak, etc...) it's hard to say that they can all be reduced down to one common element in the physical state of the brain--again, especially when you consider that everyone's brain is structurally a bit different and that we also attribute the concept "pain" to animals whose brain structures are different from out own.
Mental concepts such as beliefs and desires aren't defined in terms of their physical structures, they are defined in terms of their function. Check it: We don't say of mouse traps that they are defined in terms of a particular physical structure and anything that doesn't have this particular structure isn't a mouse trap. A mouse trap can be made of plastic, wood, metal and it can have many different shapes. The technical term for this is, mouse traps are multiply realizable; that is, mouse traps can take many forms yet perform the same function. What makes a mouse trap a mouse trap is not it's particular structure or what it's made of, what makes a mouse trap a mouse trap is that it traps mice! (now say that 5 times fast).
The same goes for beliefs and desires. What makes a belief that ice cream is a delicious treat a belief is its functional role in behavior. Someone who has this belief is very likely to eat ice cream when it's presented to them. The underlying physical structure of having this belief isn't important.
Consider another example: The belief that the assignment is due on Tuesday isn't a matter of having a particular brain state, it is a matter of how that belief functions. Having that belief functions to make people complete their assignment by Tuesday.
Just as the concept "mouse trap" can't be reduced down to a particular physical structure (a la water to H20), the same goes for the psychological concepts of belief and desire--they are non-reducible concepts.
To summarize so far, Fodor argues that the folk psychological concepts of belief and desire are independent, non-reducible, functionally defined, and multiply-realizable. (If your brain just exploded, that's ok.)
The Rumors of the Death of Folk Psychological Concepts Has Been Greatly Exaggerated
Fodor's third argument is that not only does eliminative materialism greatly exaggerates the folk psychology's short comings, but it ignores the fact that there is no theory that doesn't use the concepts of belief and desire that explains and predicts behavior as well as a theory that uses those concepts.
Suppose I call my friend and tell him to meet me at 8pm at a bar. He shows up at 8. Now, how the heck would you explain his behavior without appealing to the fact that he had a belief that I'd also be showing up at 8 to meet him.
Talking about neurological states, or electro-chemical reactions in certain parts of the brain cannot explain why he showed up at 8. You need the concept of belief to explain behavior.
Not only does the concept of belief explain behavior but it allows you to make very accurate predictions about behavior. I also could have predicted in advance (after I sent the invitation) that he'd show up at 8. Magic! We successfully predict the behavior of others all the time using (implicitly) the concept of belief. This is how everyone operates and they do so because the results are often correct.
So, the concepts of belief and desire have the hallmark properties of any good scientific theory: they allow tremendous explanatory and predictive power. So, there eliminative materialists! In your face!
One Last Thing: Ontology and a Dualism 2.0
The next thing Fodor argues is that since the concepts of belief and desire allow tremendous explanatory and predictive power, they must exist; that is, they must refer to something real in the world.
So, it seems like in a way, we have a kind of property dualism: there are mental objects called beliefs and desires that aren't reducible to physical stuff. Of course, Fodor isn't saying that there are dis-embodied souls that have these properties, only that these mental properties exist...as properties of brains.
There's a philosophy of science/epistemology objection to Fodor's ascription of ontological status to beliefs and desires. Simply because a concept gives explanatory and predictive power isn't enough evidence for ontological status. There is a wasteland of failed scientific concepts that at one time had good explanatory and predictive power given the empirical observations at that point in history.
One of the most famous examples is phlogiston theory. Scientists of the day postulated an invisible substance called phlogiston to explain heat transfer. It had excellent explanatory and predictive power until someone came up with an experiment that falsified the hypothesis. So, the story goes, explanatory and predictive power aren't enough to justify ontological existence.
Objection 2: Vs Non-Reductionism
Property dualists want to say that mental properties arise out of the complexity of physical structures. That is to say, over evolutionary time, when physical structures reached a certain level of complexity, mental properties arose.
At the same time property dualists want to say that these properties aren't reducible to any particular structure. How can you say on the one hand that mental properties arise out of particular physical structures yet mental properties aren't reducible to particular structures. It seems like the property dualist is having their cake and eating it too.