Should We Eliminate Eliminative Materialism

Eliminative materialism is a philosophical denial of mind. Paul Churchland in his essay, "Eliminative Materialism and the Propositional Attitudes" specifically seeks to eliminate folk psychology and functionalism (184-97). I believe that Churchland's materialistic vision of the day we achieve completed neuroscience and that this event will replace folk psychology suffers fatal flaws. I intend to demonstrate that eliminative materialism is false and that we can never replace folk psychology with neuroscience.

Eliminative materialism expresses the notion that eventually we will have total understanding of the neurological mechanics of the brain which shall enable us to explain all human behavior without relying on any reference to states of cognition. This is an unattainable goal. Why? First and foremost, one can never achieve a completed map of our neurological landscape. From a deterministic point of view, it seems as though we should, with enough patience and diligence, follow the synaptic connections and extract an algorithm to predict the development of new neurons and new dendritic formations. I have reason to believe this cannot be done. The brain is a complex system and any such system is bound to be subject to chaos. This is to say, that when one considers the multiplicity of variables that go into forming dendritic connections, predictability becomes less and less accurate.

As Ian Stewart points out in Does God Play Dice? the brain is a highly complex and dynamic system. Strange attractors evident in EEG readouts are unusually nonlinear. Neurologists confounded with this chaotic dilemma have resorted to studying a simpler model; algae. Babloyantz points out that the extraordinary dynamics of the brain require chaos in order for the functioning of the brain to switch states (320-28).

How are synaptic connections formed? We are born with very few in place which then grow at an exponential rate. Our brains develop as we are exposed to extrinsic stimuli. The more stimuli we are exposed to, the more new neurons are produced in addition to the dendritic connections between them. If we were to take a single person's brain and map every dendrite on every neuron and measured the mechanics of each connection in relation to the whole, we would have only accomplished mapping how that particular subject behaved. We are each subjected to enormously different experiences (even growing up in the same household) and the unpredictability of our environment create significant deviations between our neural nets. There can still be "general" similarities between brains; some parts of the brain do the same job in all individuals (we think) but when discussing both the formation of memories and what is called our executive functioning we are mostly concerned with the frontal cortex and possibly the hippocampus (which we know is related to both new memory formations and neurogenesis). This elaborate computer "reinvents" itself as it learns (Hockenbury and Hockenbury 41-3). The brain expands not only its hardware base but its ability to perform cognitive operations beyond that of any other animal. This unique capability makes it virtually impossible to predict what the synaptic map will look like from one day to the next in any one individual; never mind in the entire population of human beings!

But let us suppose that eventually quantum mechanics is able to overcome chaos and we are able to predict with certainty the neurological process in all human beings. We are learning to control chaos in dynamic systems already, but there is more than just the dynamics of the brain itself to consider (Stewart 309-10). Chaos control over the dynamics of the entire environment would also be required as those environmental influences play an integral role in the development of memories, which are the data banks of the human neural-processor. While the probability of such a feat decreases as we continue to expand, by necessity, all the unpredictability needed to be controlled for. Assuming we could do so, would quantum neuroscience then replace folk psychology? I would have to conclude not.

What is folk psychology? Folk psychology (hereafter, "FP") is the system of mentally deducing causes from repeated observations. This is to say, it is the development of propositional attitudes. If I see John buy a hot dog at a hot dog stand, I tend to conclude that John is a) hungry and b) enjoys hotdogs. This is the superimposition of categories upon the proposition that if x is purchasing a food product B it is on account that x desires to eat and enjoys the taste experience of B. John, after all, would not probably buy a hotdog if he didn't have the physiological need for sustenance, nor would he do so if he didn't like the taste of the hotdog. However, I could be entirely wrong. For example, he might be buying the hotdog for his boss, who possesses the above traits. This might lead one to conclude that there is a deeper set of functions occurring, leading to more and more mathematical computations. So why does FP exist? Simply, because it is usually accurate. It isn't in our capacity to examine with scrutiny all the possible motives for John buying a hotdog, thus we take the most likely reason. This is based on both personal experience (when I get hungry I tend to procure food, and if in my means to do so, procure food I enjoy the taste of), and on observation (when people I have encountered claim they are hungry, they tend to procure food that they enjoy eating. I can evaluate their comments about their hungriness and the appeal of the food they are eating). We also have to remember that FP is working in John too. John knows he has a distaste for hotdogs as well as that he isn't currently hungry. He thinks he is buying the hotdog for his boss, and he himself might extend the same attributive inferences in that direction. John believes his boss is both hungry and likes the taste of hotdogs.

Churchland posits that this type of reasoning is in fact a theory (184). He comes to this conclusion, however, by a type of circular reasoning. Theories include laws, and as FP is a theory, it must include laws. Because FP can be expressed with modus ponens deduction[1], it must include laws (rules) and based on this it must be a theory. I would contend that it is not a theory at all. FP is a system of attributive heuristics, which is to say, a series of cognitive "shortcuts" that allow us to attribute possible (and usually probable) causation to any effect of human behavior we encounter; in other language it is a "rule of thumb." The amazing thing is that we not only apply this to fellow humans but we also apply this to animals. As Churchland points out, it was once also used by animistic societies to explain the forces of nature (188). The amazing thing is not this but that we are not the only animals that employ this strategy! Consider a dog that bites a skunk and, as a result, gets sprayed by the skunk. A year later this same dog encounters another skunk. While I do not wish to suggest that the dog "thinks" about the skunk, it reacts nonetheless (hopefully). The dog avoids the skunk entirely and thus avoids being sprayed again. Now we can call this a behavioristic approach, which would be proper, but it operates under the same conditions as FP. In that sense, FP is a learned behavior (those unpredictable synaptic developments at work!). If Janice calls me up to wish me a happy birthday, I believe this is due to her sincerely wanting to make such a wish. I base this on the fact that typically when people have wished me a happy birthday, they displayed no notable ulterior motive.

The most amazing thing about FP in my opinion, is that this type of naturally occurring deductive reasoning has been refined to the application of research science. By learning to systematically (as close to algorithm as possible) eliminate improbable causes for experimental subjects upon which we can control as many variables as possible, we can come to a conclusion that is probable (never certain). This system of advanced folk psychology for lack of a better term, has led us to neuroscience in the first place!

So why can't neuroscience replace FP? They are two different species entirely. The point of FP is to attribute a reasonable cause to observed or experienced behavior. I mean behavior to include emotions and reflexive thoughts as well as somatic behavior. How can even a complete understanding of neurology serve this purpose? It would be entirely maladaptive to this end. If one is walking down the street and sees a disheveled man muttering to himself while holding a bloody cleaver, we are willing to risk being wrong when we apply our FP to determine this man to be dangerous. We assume he may intend us harm, and in cases such as these, we would probably be correct most of the time. Perhaps this man was just fired from his job in a butcher shop for his appearance, and was ruminating to himself about his dismissal. Possible, perhaps, but FP allows us to skip the critical examination of data and jump to a conclusion that has a good chance of being correct. A computer analyzer could conceivably know everything about the operation of her computer. When the analyst opens up her word processing program and types "K" she will be able to identify every physical process that occurred from the moment the plastic key completed a connection in the electronic pathway below the key, until the visual image of the letter "K" appeared on her screen. If this is the proposition "Computer believes that Input A$: "K" then print "K" and goto input B$ etc. this could be expressed simply as a matter of the CPU-science. Does her knowledge of this process in any way benefit her when typing "K" compared to a CPU-science ignoramus? It seems that although both type "K" if the analyst were to stop and consider the physical process rather than the proposition she would be easily outpaced by the far more efficient ignoramus. Now if the complete understanding of CPU-science isn't very helpful to the analyst in the execution of running a computer program, is it possible that a completed neuroscience could fare much better?

One way that Churchland believes neuroscience would succeed where FP fails is through the suggestion that with such a completed neurology we would be able to adapt our technology to allow our brains to interface with another by means of an implanted device. The information exchange that occurs in the corpus callosum would essentially operate between all brains (195-6). Working from his prediction, we can easily see that all human beings would have to be hardwired into the system, for if any were left out, we would not be able to rely on the direct transmission of synaptic information for that person and we are back to using FP again to compensate for our lack of information about their behaviors. This seems to lead to the conclusion that Churchland is onto something; such developments would override the chaos problem as we would all be sharing one giant brain, and that collective brain would have no need for FP at all as we would have the neurological ability to know their behavior. After all, your behavior would be my behavior as well. The inherent problem is that it still would require us to use FP on animal behavior (nor is it realistically feasible to consider adding every other animal on earth into the collective brain). Consider that one of our bodies, Body Steve, has just encountered a bear. Body Steve's sensory organs send data to the collective brain which indicate that the bear is staring at Body Steve, and growling. All of our neuro-technology will not help us come to a certain conclusion as to the bear's intentional state. We then would still drum out good old handy FP and determine that from what we have learned of bear behavior, this bear means business! The collective would then order Steve to beat feet. Thus, we cannot escape the use of FP, nor do we have to give up our identities in order to function in the world fairly effectively. Fodor claims that FP has "predictive adequacy," (see "Introduction: The Persistence of the Attitudes" 252) and while a completed neuroscience would be entirely more accurate, it would not be anywhere near as adequate.

If we can accept that FP is a heuristic, not a theory, what are we to make of the origins of propositional attitudes? Where do beliefs, feelings, and such, come from? What are they in the first place? To be as succinct as possible, they emanate from not just the brain, but the complex network of neural connections in the brain. Digital computers run on bits, ones and zeros, that form a code that interact according to the way both their hardware and programming syntax dictates they should act. Our brains are also digital computers, although far more advanced. Neurons, the basic "mind" cells of the body, are likewise one's and zero's. In neurology this is known as the all-or-none principle. Either the neuron is firing or it isn't (Hockenbury and Hockenbury 44). It is either a one or it is a zero. A belief arises when the neurons fire in the right syndrome of sequences generating a state of probability. The mental program receives some input that causes a probability state to arise, based on the data (memories) in the hard-drive (neural pathways), and selects the most salient data (similar to your "most recent" file). The program then assembles a propositional attitude. I believe it will rain today. I calculate informally that there is a fair probability of rain today (based on my schema of the conditions necessary to produce rain). This sounds like a cold way to describe ourselves but I would note that it is far less cold than eliminative materialism, retains the utility of FP, and allows for that "voice" inside our heads to exist.

In conclusion I would like to add that I believe neuroscience, as a set of theories, to be a noble pursuit from which we can glean a lot of information that will help scientific psychology and the human condition in divers ways, but when all is said and done, it is a different creature entire from the non-theoretical heuristic FP and thus serves a king of a different kingdom.

Churchland, Paul. "Eliminative Materialism and the Propositional Attitudes." Problems In Mind:

Readings in Contemporary Philosophy of Mind. Ed. Jack Crumley. Mountain View, CA:

Fodor, Jerry. "Introduction: The Persistence of the Attitudes." Problems In Mind:

Readings in Contemporary Philosophy of Mind. Ed. Jack Crumley. Mountain View, CA:

Hockenbury, Don H., and Sandra E. Hockenbury. Discovering Psychology. 2^nd ed. New York: Worth,

Stewart, Ian. Does God Play Dice? The New Mathematics of Chaos. 2^nd ed. Malden, MA: Blackwell,

[1] Modus ponens is a type of proof used in philosophy and other disciplines to arrive at a logical conclusion by means of following premises. For example, if p then q. P, therefore, q.