A Review of Understanding Behaviorism and the Philosophy of Psychology

Introduction

Understanding Behaviorism by William M. Baum is a very silly and intellectually lightweight book which should not be taken seriously. However, there are inexplicably to be found a number of favorable reviews online, so I think another look at the book is well worthwhile. Understanding Behaviorism is philosophically confused and extremely dogmatic. Furthermore there are fundamental inconsistencies with the premises that the author advocates, with the conclusions he arrives at, and with what he actually, apparently, believes.

Pragmatism Versus Idealism

Baum begins by contrasting two philosophical stances he calls realism and pragmatism, the later of which he believes to be the correct one. The characteristic of realism is the distinction between sensory experience and the processes causing the sensory impressions which are to be explained. Interestingly enough, Baum calls Bishop Berkeley a realist because of this distinction even though Berkeley usually would be considered the opposite of a realist. A pragmatist, on the other hand, does not speculate upon the existence anything beyond the senses. This is the view Baum favors. According to pragmatism,

Science is the pursuit of economical and comprehensive descriptions of human natural experience (i.e., our experience of the “natural world”). The goal of a science of behavior is to describe behavior in terms that render it familiar and hence “explained.”

Although this is an interesting philosophical distinction, I don’t believe it is all that helpful either for doing science or for the philosophy of science. All science is ultimately pragmatic regardless of the metaphysical stance of the scientist because as a phenomenon becomes better understood, both realists and pragmatists will find it easier to identify and remove unnecessary concepts from it. To a realist, the concept no longer appears to be an implication of the theory, whereas to a pragmatist it simply no longer appears to be useful. When a phenomenon is first being studied the pragmatist has no advantage over the realist because it is not at all easy to judge how useful a concept is at first. Between competing theories there is rarely one that is obviously more pragmatic than the other; instead one will appear to be useful in some contexts but problematic in others and vice versa.

Baum seems to believe that a major problem in developing a theory is the introduction of spurious concepts that a good pragmatist should be able to identify and remove. He discusses several historical theories that could be said to have spurious elements, such as the phlogiston theory of combustion, the luminiferous ether, and the law of horreur vaccui. However, I would say that each of these theories would have seemed perfectly pragmatic to the scientists of the day because they introduced only those causes sufficient to explain an unexplained phenomenon in a way that fit in well with what was believed about the rest of science at the time. Pragmatic objections would have meant little to the scientists of the day because they were unaware of alternative and more pragmatic theories. I suggest that the much more important problem to developing a theory is lack of imagination, not too much imagination.

Let’s look at the realist/pragmatist distinction from another angle. Different branches of natural science cannot be understood wholly independently. Rarely is a phenomenon studied in the light of no contextual knowledge from related or more fundamental branches of science, and in this case the theory is constrained by this contextual knowledge, often to a single objective answer in terms of the more fundamental theory. Here the scientist is not free to choose concepts as needed, so his philosophical stance is effectively a realist. The behavior of a brain whose anatomy is understood all the way down to the cell, and whose chemical and biological processes are understood quite well cannot be investigated from an wholly pragmatic stance.

Thus Baum’s philosophy begins along lines I find misleading. Baum next distinguishes between methodological behaviorism and radical behaviorism. Baum does not define these terms explicitly, but a methodological behaviorist appears to be a scientist who takes behavior to be a means of studying internal mental states, whereas the radical behaviorist attempts to categorize behavior in ways that most predictive connections to be found between them, and who rejects such concepts as the mind. Baum identifies methodological behaviorism with realism and radical behaviorism with pragmatism.

Do you see the problem here? Baum’s philosophical schema is exactly opposite what it should be. Methodological behaviorism should be considered the pragmatic position and radical behaviorism should be considered the realist position, at least according to Baum’s definitions. Why? A pragmatist, by Baum’s definition, should not have any problem in postulating occult concepts (including that of a mind) to explain phenomena—if they work, then they work. A pragmatist therefore, ought to have no trouble entertaining the idea of mental processes, even if he considers them to be no more than a convenient fiction. It is the realist, according to Baum, who takes a metaphysical stance about what is real and what is not.

In fact, Baum’s description of radical behaviorist methodology corresponds well with Bishop Berkeley’s philosophy of natural science, whom he quoted earlier as a representative of the realist side. Berkeley did not simply remain agnostic to the reality of scientific explanations, but rejected any theory which postulated any kind of hidden processes. He summarily discounted any idea of causes for our sensory impressions. He criticized Newton’s theory of motion for being too abstract and he wanted all scientific theories to be described as relationships of sensory impressions rather than abstract hidden causes. This is precisely the same attitude Baum takes regarding mental processes.

Berkeley’s idealism (which is what his philosophical stance has traditionally been called) better describes Baum’s position than pragmatism because it is a much better description of the positions that Baum takes throughout most of the book. However, I would propose that all this philosophical stuff is more fundamentally a red herring because it is applied in an inconsistent and ad-hoc way simply to promote a foregone conclusion. If there is some concept Baum likes, he treats it pragmatically and concludes that it is a good idea. If there is some concept Baum doesn’t like, he treats it idealistically/realistically and concludes that it is a phantom. There is no unifying principle to Baum’s philosophy other than behaviorism itself. The philosophy I have discussed here should be seen as nothing more than a superstructure caused by Baum’s behaviorist bias, and does not serve as an underpinning to behaviorism at all.

The Rejection of Mentalism

Baum goes on to define public and private events. Private events are those which one person can report and public events are those which more than one person can report. Consistently with his pragmatic stance, Baum says that there is no special difference between these two kinds of events. However, he then goes on to define fictional events, which he wishes to reject from scientific theories.

Fictional things and events are unobservable, even in principle. No one has observed a mind, urge, impulse, or personality; they are all inferred from behavior. A person who behaves aggressively, for example, is said to have an aggressive personality. No one will ever see the personality, though; one sees only the behavior. (40)

I do not understand how, as a supposed pragmatist, Baum is able to declare that some theorized causes are observable and some are not. A consistent pragmatist would have to agree that the concepts necessary to relate our experiences may be extremely far removed from them, and that no proposed cause is unobservable as long as it has some effect on my experience. There is no way to say that a cause is too far removed from experience to be considered fictional. Hence personalities are just as observable as anything else and perfectly valid as a proposed cause to a person’s behavior, and Baum’s rejection of them is much more consistent with an idealist philosophy than a pragmatic one.

Shortly thereafter, arguing against the concept of intelligence, he says

But where is this intelligence? What is it made of? How could it cause behavior? Its ghostly nature derives from its being the label, rather than an instance, of the category. (45)

Now obviously a pragmatist would not need to ask what anything is made out of unless postulating some kind of composition were helpful in organizing his experience. Only a realist would have such a philosophical requirement.

Let us apply these rhetorical questions to an earlier point where Baum describes radical pragmatism.

The pragmatist (radical behaviorist), having no commitment to any idea of real behavior, asks only which way of describing the man’s behavior is most useful, or in Mach’s terms, most economical – that is, which gives us the best understanding or the most coherent description. That is why radical behaviorists favor definitions of activities that include the man’s reasons for running, like exercising and fleeing the police.

If Baum can ask what intelligence is made out of, then I can certainly ask what a motive is made out of. If Baum wishes to reject intelligence upon his realist objection, then either he must also reject motives from his theory or he must propose some kind of hidden structure that produces motives (ie., a theory of mind). And if I am already developing a theory of mind, then the mind must have actual properties—exactly what Baum wishes to reject when he rejects personality and intelligence.

As a realist, Baum has no principled way to say that a motive is a private event whereas a personality is fictional. If there is such a thing as a personality, then I am observing it constantly when I am observing my own motives. The only difference is that I only observe, through my motives, one part of my personality at a time. To discuss behavior in terms of motive is inherently mentalistic.

As a pragmatist, Baum has no basis to reject fictional events, or even identify an event as fictional. Furthermore, think of how many fictional concepts are a completely uncontroversial part of science. Just think of a force in physics. No one has ever observed a force either; one only observes masses and changes in velocity. Yet all of Newtonian physics is based around the concept of a force. If Baum wishes to reject personality from psychology then he ought to reject forces from physics. Once again, this is just what Berkeley wanted to do.

Baum also rejects the concept of will earlier in the book. I don’t understand how rejecting will but not rejecting motives makes any sense at all. Will is just a kind of feeling that I directly experience just as I do my motives, and therefore should therefore be considered a private event.

Dualism

Pragmatic arguments are incapable of refuting mentalism out of hand on metaphysical grounds, which is why I have analyzed Baum as an idealist. However, as a pragmatist he also can attempt to refute mentalism with the criterion of parsimony. The results here are no better because he still lapses into idealist arguments.

Baum’s tactic is to ignore his previous (pragmatic) admission that private events are valid as an object of study, then use idealist arguments to dismiss any contrary evidence, and then claim that the concept in question is useless. We can see this in his discussion of what he calls copy theory, which is the theory that the brain produces a copy of its sensory impressions in order to analyze them.

This notion may be useful in understanding some things about the eyes, but it in no way explains seeing. The problem of how the tree is seen is now replaced by the problem of how the tree’s copy is seen. Copy theory has all the defects of mentalism. The appearance of an explanation – you see the tree because you have a copy of it in your eye or in your brain – distracts us from our attempt to understand what seeing is. The copy is superfluous, because the question remains the same whether we ask about seeing the tree or seeing the copy. (p. 52)

It is perfectly obvious that it would be impossible to think about something if it were not represented in some way in the brain. The explanation for seeing is simply that the representation is in some format that the brain can directly manipulate. What this format actually is I don’t know, but whereas I would consider it an actual problem for psychology to figure out, Baum apparently doesn’t see that there is a problem.

Here is a good example of my earlier point that science is often constrained to engage in realist thinking so as to produce a result consistent with what is already known. We’re not talking about some black box that exists out of the universe here—we’re talking about a physical object, a brain, that must function Through the transformation of its physical state according to known laws of geometry and information processing and there is no philosophical problem with relying on contextual knowledge of these laws in order to theorize how it works. This is not a realist metaphysical stance, it is just a necessary conclusion if I am to take all the rest of science seriously.

Furthermore, to say that my brain has a copy of the things I see is not just pie—in—the—sky theorizing. It’s a private event that I can report on. The copy is exactly what I am seeing now. It is right there. I can remember or imagine an image rather than see it directly and the copy is directly experienced. There is no longer an instance in front of me, but there is a copy, or representation in my mind’s eye. If a motive counts as a private event, I don’t see why an act of vision or of imagination should not.

Here is another inconsistency in Baum. It would certainly seem that if the mind represents information in some way, then this should certainly be observable in principle as some configuration of neurons or electrical signals in the brain. A pragmatist should therefore should have no particular problem with copy theory.

Soon thereafter, Baum drops this bombshell:

Many psychologists cling to the idea that if an activity repeats, it must somehow be represented inside the person, presumably in the brain. When faced with the defects of representations as copies, they often insist that the representation is just the workings of the brain. By such reasoning, when I start my car each morning, the running engine must be represented in the resting engine.

In this astonishingly ignorant, breathtakingly obtuse quote, Baum shows us just what lengths he will go to prevent himself from understanding opposing arguments. How he could respond with such an inappropriate analogy is beyond me. There is no reason to analyze a car as an information processor so of course nothing about it is a representation. Apparently he is only able to think of the brain as a machine that transforms energy from one form to another and is incapable of viewing it as a kind of computer. I suppose this is because if he were to analyze the brain as a computer he would have to start imagining that there are complex processes going on IT.

Just take a look at this:

Copy theory attempts to explain dreaming and imagining by the idea that copies are stored in and retrieved from memory. Questions about recollection become questions about ghostly mental processes of encoding, storage, and retrieval. If when I imagine my childhood home I see my father there, supposedly that is because the two copies are somehow linked together in memory. If when someone says, “Think of birds,” I think of sparrows, finches, and ostriches, supposedly that is because copies of those things are linked in some way in memory.

In contrast, the behavior-analytic view points to facts of life. When I was a child, seeing my childhood home, I saw my father, too. When I heard about birds, I often heard about sparrows, finches, and ostriches. If these things are linked, it is not in memory, but in time and place. Recollection is repetition. (p. 54)

Huh??? Are you kidding me? This has to be among the dumbest things I’ve ever read. Not only is it no explanation whatsoever, it’s not even clear. It’s slogans and poetic language. When confronted with the problems of vision and memory and having nothing remotely useful to say about them, Baum lapses into nonsense.

As further evidence of how foreign any kind of information processing view is this view is to Baum, take the following quote:

The radical behaviorist’s objection to mentalism is really an objection to dualism, the idea that two sorts of existence, material and nonmaterial, or two sorts of terms, referring to the material and the nonmaterial, are necessary to understand behavior fully. (p. 43)

However, dualism is a pretty obviously useful philosophical and perfectly not controversial in the related field of computer science. In computer science, we can discuss computers, which are concrete machines which represent information and manipulate that representation in various ways to do computations, and programs, which are abstract processes that act on information, and which can run on any computer regardless of its specific architecture. These are two different metaphysical things: matter and information. That’s dualism, and there would be no computer science without it.

The analogy to psychology is obvious and the concept of a personality or mind has very obvious uses. Just as we can speak of an algorithm in abstract terms without reference to the particular computer architecture that runs it, we should be quite able to talk about a person’s internal neurological functions abstractly without reference to neurology. The fact that Baum can’t understand the value of these concepts just supports my claim above that pragmatism is not easy to apply in practice and not terribly helpful for either doing science or investigating the method of science.

Hidden Realities

The only way I can see to make sense of Baum’s philosophy is to suppose that he has an irrational prejudice against abstract concepts in psychology (but not in other sciences), and that there is some arbitrary level of abstraction that he refuses to entertain. A motive is relatively concrete. Motives are ok for Baum. A personality, on the other hand, would be a cause that explains many motives. Too abstract! Since his arguments don’t make any sense together and are applied completely inconsistently, my best hypotheses is that it is all just motivated reasoning based upon the overall effect that Baum seems to believe results from them.

This explains perfectly why Baum continually lapses into idealist arguments but is never a consistent idealist. Baum likes a concept if he can observe it. It’s ok if the concept can only be observed a little indirectly. But if observations must be too indirect, then he doesn’t like it. He doesn’t like it even if it could be observable in principle with superior technology or if it denying it would lead to flagrant absurdities. This is a very idealist prejudice.

Let’s see where this prejudice would lead if it were applied more generally throughout the natural sciences. One of the hallmarks of a great theory is that it may predict phenomenon that has never been imagined or observed before. This has happened many times in the history of science—to name just a few examples from 20th century theoretical physics: black holes, antimatter, general relativity, the Higgs boson, the W and Z bosons, the top quark—but under Baum’s philosophy it is impossible. To create such a theory would require creating an abstraction and applying it to untested circumstances. Since Baum hates abstractions, this could never happen by his method.

Baum’s skewed history of science discusses only events in which theoretical ideas had to be abandoned, but none in which they turned out to be exactly right. For probably the best such example, consider the concept of the atomic bond, which was not directly observed until over a century after it was proposed. Despite this, 19th—century scientists used the proposed geometric relations of bonded atoms to explain the thermodynamic properties of substances, the nature of the chemical reaction, the properties of crystals, and the reflection and refraction of light within matter.

This example is a good case study of pragmatism versus idealism. The pragmatist position, when applied to the theory of molecules, leads in a theory with greater depth and greater abstraction. This is as it should be; a pragmatist would have no qualms about using invisible geometry to make sense of the properties of matter. Whereas Baum consistently and inappropriately attempts to apply pragmatism so as to reduce the level of abstraction in the theory. If Baum had lived in the 19th century, he would have insisted that the thermodynamic properties of matter should be the fundamental qualities to be related to one another and that molecules are a needless complication.

There have been movements similar to that which Baum advocates in other fields and these have, to my knowledge, miserably failed. An example would be the 18th century German naturalist movement which produced Goethe’s theory of colors. Another would be S-matrix theory in particle physics, as advocated by Fritjof Capra in his popular book Tao of Physics in the 70s. Around this time in physics there was this idea that quantum field theory added a lot of layers of nonsense into the computations that told how particles interacted, until people realized that field theory can allow for all kinds of objects that are not particles at all, called solitons. These, too, have never been observed, but play an important role in ideas about the history of the very early universe. (Examples of such objects would be cosmic strings and magnetic monopoles)

Finally, think about the theory of evolution in biology. This example just shows how inconsistent Baum is because here he is perfectly content to be a pragmatist. He believes in historical explanations, so he is fine with evolution. Yet, from a Berkeleyian perspective, isn’t there something extremely far removed from our senses to postulate billions of years of history as an explanation? This history cannot be observed because it happened long before the human race existed, and the evidence for it is quite oblique. I a consistent idealist would have to reject theories about ancient history as much as he would reject mental processes. In fact, Baum’s arguments against mentalism I find to be very similar to some arguments that creationists use against the ancient Earth.

Let’s look at biology from the standpoint of Baum’s philosophy from another angle. I would suggest that the concept of life has a very similar metaphysical status to that of mind, and that any argument which can be used to reject one can apply to the other. Nothing is objectively alive; all we see is a physical processes and everything we normally consider to be alive can be understood in physical terms. We impose life upon our experience in order to separate parts of the world from the rest and understand them differently. Thus, life is a “vague and useless” concept like that of consciousness or mind and therefore much of biology should simply be seen as a phantom with a lot of theoretical nations that only complicate matters. More collateral damage from Baum’s rampage.

I’ll go one step further. The concept of behavior itself is as subject to these criticisms as are those of life and mind. For how do I know that something is behaving? I never actually see anything but a physical process obeying physical laws. If I declare that some part of this process is separate from the rest and furthermore say that it is an organism with its own behavior, this is no less of a judgment than that which I would use to declare that it is conscious and able to think rationally. I can understand the motion of an animal purely in terms of physics, so there is no need to understand it in terms of behavior. Furthermore, I could just as easily argue that to do so implies a dualist philosophy, insofar as I postulate physical law and behavior as metaphysically different kinds of explanations. Thus, all the arguments that Baum uses to reject minds and consciousness could just as easily be used to reject behaviorism as well.

A typical animal completely replaces every single atom of its mass every few years, and it is constantly expelling and incorporating matter to and from its environment. Thus, to even identify an animal as having its own identity is, metaphysically, quite a strong statement, one which I cannot see as being any less questionable than that of explaining its behavior in terms of a mind.

Baum could get around this by saying that behavior is just another word for motion, but then he would also have to say that an avalanche or tornado behaviors as much as an animal does. If so, Baum would not be able to identify psychology as a separate science from physics. Baum could also say that behavior is an emergent process that is made out of complicated physical processes, but if he did I could just counter that minds are emergent processes made out of the complex behavior of neurons in the brain, and if he can study emergent processes on their own terms then so can I. I suppose Baum could also try to deny physics, but that’s a topic I get into in the next section.

Behaviorist Explanations

Understanding Behaviorism contains some of the most absurdly confused philosophy I have ever read. Baum first describes a philosophical stance which he claims to hold, but he actually tends more toward a stance which he previously argued against and was not able to name properly. Neither of the stances he inconsistently vacillates between is capable of defending his actual stance, so the philosophy he ultimately arrives at has nothing to do with the philosophy he uses to get there. On his way to it, he misinterprets the history of science and fails to notice that his philosophy, if applied consistently, would destroy virtually all science, including the one he wants to save.

Now I want to talk about the kinds of explanations Baum’s philosophy offers. Let’s be frank. As long as hypothesizing about the internal state of an organism is not allowed, a behaviorist experiment is incapable of explaining any behavior whatsoever. It is only capable of showing that a change in the environment produces a certain change in behavior of the subject. It is not capable of explaining the behavior in an absolute sense.

Say I train my cat to perform a trick, such as to run in a circle, by giving him treats each time he more and more closely approximates the desired behavior. How is he able to run without falling over? How is he able to navigate a three—dimensional space so as to move in a circle? Behaviorism cannot answer these questions because all of them require speculating on the internal mechanisms that produce behavior.

And what about the nature of reward and punishment? These concepts seem to show up in Baum without any explanation at all. How does the cat recognize a treat as a reward? What is the mechanism which makes him like some things and not others? It should be impossible to learn that one thing is a reward and another is a punishment because learning only occurs as a result of rewards and punishments.

For that matter, how is the cat able to make the connection between behavior and reward? If the cat performs a trick that I like because he expects a reward, he must have been able to observe a coincidence between the reward and punishment and to treat it as a causal connection. But why? There is no inherent reason to think that his behavior and his reward are causally connection, and no logical reason to take notice of the coincidence. How is he able to experiment and recognize that some behaviors produce treats and others don’t? There must be some explanation for that as well.

Furthermore, there are some things my cat will never learn. I could encourage him to listen to me talking for the rest of his life and he’ll never understand what I’m saying, whereas human children who are exposed to language generally learn a language. Behaviorism can’t explain that. I could try to teach a rock to perform a trick too, and the rock will never learn it. What is the reason the rock cannot learn but the cat can? Behaviorism cannot explain that.

The basic purpose of changing behavior rather than explaining it can be seen in the distinctions between positive reinforcement and extinction and between negative reinforcement and punishment. These terms are all defined on how a stimulus relates to a specific behavior. If the emphasis was on simply explaining behavior then there would be no need to make these distinctions. The focus on specific behavior rather than on all behavior is because it is not typically possible to observe all behavior. To describe all behavior would require a detailed description of the behavior of the organism within its own brain.

According to Baum, pragmatism is about explaining mysterious concepts in terms of more familiar ones, but the effect of his philosophy is the opposite. He rejects a vast array of interesting and meaningful questions entirely from consideration, namely all those purporting to explain behavior by way of some internal process. Science is easy if you can just ignore all mysteries! Furthermore, the terms by which Baum wishes to explain things do not seem much less mysterious than those he wishes to reject. Concepts such as reward and punishment are taken to be axiomatic and not requiring explanation. The explanations Baum gives sometimes seem downright mystical. He says that “recollection is repetition” without even giving a thought to how the brain is even capable of repetition. This makes his theory far less enlightening or useful than the copy theory one he was attempting to refute, and his explanations far less familiar than the phenomena he wishes to explain with them.

Now on to one of the weirdest aspects of Baum’s philosophy. It appears that Baum rejects the ordinary view of causality in which the present is caused by the immediate past. That may sound absurd, but it is the only interpretation I can make when he says, “When I heard about birds, I often heard about sparrows, finches, and ostriches. If these things are linked, it is not in memory, but in time and place.”

Baum appears to believe that a being is the product of his experiences. How is it possible for experiences to produce that being? Baum doesn’t want to say that these experiences are processed by the brain or that we are active interpreter of them. Instead, those experience reach straight ahead through time to produce behavior years later.

This is stupid. A historical explanation is not sufficient—there must be some immediate cause to a given behavior. History produces the mind—the mind produces the behavior. If Baum wishes to deny this—and he certainly appears to—then he is denying the ordinary understanding of causality.

If I set up an empty terrarium and paint the walls black and throw in a bunch of eggs and seeds once in a while, do I know what is inside the terrarium? Of course not! Some of the seeds will have germinated and others will have died. Some of the eggs will have hatched and produced viable populations and others will have been eaten. I have no idea how everything inside has interacted and what the result will be. Baum would deny this.

According to Baum we are not processes at all—we are a pile of prior rewards and punishments that magically time—travel into the future and cause us to act. I know this may sound like I’m talking about a strawman, but literally every single time Baum addresses any speculation about the internal mechanism for behavior he brusquely dismisses it as mentalism. I can find no evidence that he ever allows for any speculation upon the mechanisms of any internal process to cause behavior.

Baum wishes to rely on historical explanations in psychology, but a historical explanation is clearly not possible without reference to internal states. A historical explanation will take stimuli from an animal’s history and attempt to use what is known about how these stimuli individually affect behavior to say what their sum means about the animal’s present behavior.

However, without having done an experiment testing both stimuli together, the behaviorist has no logical basis for making predictions about what should happen. If he has performed separate experiments with the two stimuli he cannot postulate that the two stimuli should interact in some way without there being some medium by which they interact. Thus, any historical explanation implicitly makes reference to an internal state; one may remain totally agnostic as to the nature of that internal state but in that case one is still a methodolgical behaviorist and has no philosophical grounds to argue that someone else should not be so agnostic.

Towards a Philosophy of Psychology

Baum opens his book with two assertions: a science of behavior is possible; and that the whole of psychology should be the identical with behaviorism. He does a terrible job with the second of these assertions and spends very little time on the first, apparently because he considers the concepts of behaviorism to be unproblematic. However, as I have argued, notions such as ‘behavior’ are similarly problematic than that of ‘mind’, and the arguments which Baum uses against mentalism can just as easily be applied to behaviorism.

If Baum wants to be a behaviorist, he therefore has no consistent grounds to refute mentalism. Is his second assertion therefore incorrect? Not necessarily, because he is not required to affirm mentalism either. There is an analogy here with religion and the debate between the theists, whom I define as those believing that the universe has an ultimate meaning, and the atheists, who deny any ultimate meaning. There is of course no objective way to decide between these views because nothing is objectively meaningful. It is always philosophically valid to say that a thing has meaning or that it does not as long as the judgment is logically consistent with ones general philosophical premises.

Thus, a behaviorist may take the view that ‘mind’ is meaningless whereas ‘behavior’ is not but he may not argue that these two beliefs are logically related to one another. Someone else may, without contradiction, believe in both, either, or neither. Admittedly it would be quite alien to believe in minds but not behavior, but this is beside the point.

All science begins with a set of axiomatic concepts which, crucially, conceive of the world such that a means of investigating it through experimentation is possible. The concepts are not objective, but the results, given the concepts, are objective. Odd, but true. Recall my earlier arguments about the concept of life: although nothing is objectively alive or dead, once it is agreed that something is alive, then its functioning can be objectively studied through its anatomy and behavior, and once a group of organisms are identified as being in competition with one another, than the evolutionary forces shaping their history can also be analyzed.

This is the basic reason why Baum’s arguments against mentalism all fail. He continually tries to show that mentalism is either unnecessary or unreal, depending on whether he feels more like a pragmatist or an idealist at that moment; but because the concepts of science are not objective, the same could be said of the basic concepts in any science. All basic scientific concepts exist because of a human purpose, not because they have an objective basis. Thus, they are all both unreal and unnecessary to anyone who has no interest in analyzing the world in terms of them.

Had Baum been a little more shrewed, by the way, he might have produced a much more effective attack on mentalism by attempting to show that it does not admit experimentalism. This is the case with theism: although, philosophically, one might affirm an ultimate meaning to the universe, this affirmation comes with no necessary method of interpreting that meaning. Thus although theism may be a valid philosophical stance, there is no valid theological method.

Now let’s discuss Baum’s prejudice against theorizing about internal processes. Philosophically, this prejudice against internal processes is equivalent to a belief that the only important phenomena to study are those which are directly observable. There is no way this belief can be justified as a part of science. Furthermore, it is subjective whether a phenomena is directly or indirectly observable, since every phenomena is ultimately observed through a chain of causes that ultimately produce some change in the observational instrument, which ultimately produces a change in the scientist’s senses, which produces a change in his brain.

Furthermore, since the concepts of behaviorism are as abstract as those of mentalism, a behaviorist who attempts to describe the internal state of an organism in behaviorist terms is immediately capable of discussing theories which are abstruse as those of any mentalist. He may conceive of an organism as a set of motivators whose combined effect is some behavior, some of which may not be visible. The only way this invisible behavior could have a real effect is if it is a self-motivator, ie, if it motivates some other behavior in the organism.

Once this sort of model is allowed, then it is easily to construct a behaviorist model which is absolutely equivalent to any model which conceives of the organism’s mind as a kind of program composed of interacting parts. A behaviorist could conceive of a set of trillions of self-motivators, one for each neural connection in the brain. He could also conceive of set of a set of motivators which reproduces the theorized interactions of the id, ego, and superego.

Thus, there is no science of behaviorism as Baum conceives it. Any attempt to resolve his contradictions results in something which, depending on how ‘mentalism’ is defined, would be either equivalent to it or would encompass it.

The Science of Mind

The question of whether there is a mentalist science is a question of whether mentalism comes with an experimental method and has nothing to do with either idealism or pragmatism. Pragmatism does not tell us whether a science exists; the pragmatic question is whether the investigation of that science actually satisfies human curiosity.

So is there a science of mind? I will argue that there is such a valid method, but I will first have to give a good definition of mentalism. The problem with science today is that it has been too heavily influenced by physics, a science characterized by a set of mechanical laws that describe the evolution of a system. I suggest that the best analogy for a science of psychology is not physics but rather reverse engineering.

The ways in which purposeful machine, whether deliberately designed by humans or a result of natural selections, is different from a physical process are that machine is not characterized by laws but by a set of costs and benefits, each of which has a greater or lesser role is the form of the machine, and that the design of a machine requires several different schematic theories in order to be understood.

A machine must be described as a set of interacting parts but the number and nature of those parts will differ depending on the purpose of the description. A manufacturer will divide the machine into parts according to the geometry and material of the machine, and will describe them in terms of how the physical parts are put together into the final product, without regard to the role these parts play in its functioning. A designer will prefer to describe the machine as a set of subsystems whose functions can be relatively self-contained. The final user of the machine will want to describe it in terms of how he will get the most use out of it.

In the science of reverse engineering there is the added challenge of deciding which levels of abstraction are appropriate. There is also the problem that no schema is objectively true or false; rather the test of a good schema is how usefully the machine can be described according to the given schema. In physics, it is presumed that there is a single theory, or several equivalent theories, which are the object of discovery, and there will be a partially ordered set of approximate theories appropriate to describe the world under different circumstances. Whereas in a science of reverse engineering there will rather be a set of theories, none of which on its own is entirely sufficient.

Mentalism, therefore, I define as the theory of the reverse engineering of behavior. This definition is capable of encompassing all aspects of psychology. Behavior may be seen as having been produced by an abstract program without worrying about how this program is physically embodied. This is the approach taken by psychoanalysis. A machine may be studied by examining the physical interactions taking place within it. This is the approach taken by neurology. A machine may also be examined by looking at its behavior and examining a priori the kinds of processes which would be capable of producing that behavior. A machine may be examined in terms of the known purposes for which it was constructed. This is the approach taken by evolutionary psychology. This is the approach taken by cognitive science. A machine can also be legitimately viewed as a “black box” upon whom to experiment. This would be behaviorism.

A science of mind would be preoccupied with the following questions: what kinds of programs can be run on an organism of a given species?; in what ways can these programs be described most usefully as a set of interacting modules?; what can be observed in an individual to learn about the specific program he is running?

This is the science of mind, and to me it appears that psychology is preoccupied with these very questions. It is a science of the techniques for reverse engineering a kind abstract process. It is about learning how to model the mind as a set of interacting parts in several different ways, evaluating the utility of each kind of model, and determining the correspondence of each of these models to features in the functioning brain.

This approach has been very successful for studying functions for which natural selection was heavily constrained by objective features of the environment. Take, for example, vision. The problem for a brain with vision is to reconstruct a three-dimensional representation of the environment which is categorized into a set of objects, given the constraints imposed by the physics of the eye, the physical needs of humans, and the typical environments in which humans live. All of these constraints will have implications for what sorts objects the vision programs will specialize in recognizing.

I have yet to read anything that attempts to describe in much detail how an image would be represented as a structure in the brain. Nevertheless, human vision has been explained in terms of a very detailed set of modules which can be described both as a computational process, and as a hierarchy of modules with broader and broader functions. The computational aspects of this question have yet to be fully explained but the theory of human vision is still broadly understood.

Similarly, many other modules have been identified in the human program whose existence and functioning are quite well explained by evolutionary considerations. People have specialized modules for physics, biology, economics, ethics, and general science (that is, identifying causal relationships). Much of what is fallacious in human reasoning and self—destructive in human society can be explained extremely well according to the ways that these modules are adapted to the specific conditions to be expected in human hunter—gatherer societies that do not apply to modern society or to human traditions (such as natural science) which have developed in civilizations. For example, people are very prone to post hoc ergo propter hoc errors because in the wild the risk of failing to see a causal connection (the sign of a predator or poison perhaps) tends to be much greater than that mistakenly identifying a causal connection. Thus, the human science module is adapted to identify causality everywhere. Unfortunately it is still not clear how each of these modules work from a computational standpoint, but nevertheless the theory appears to be fairly clear and general.

On the other hand psychology has been less successful to date regarding psychological processes whose purpose is less easily defined. Humans are also capable of reasoning more generally without relying on specialized modules. They can adopt and apply rules according to a goal rather than by instinct. A human can learn to be a good scientist even though his natural instinct is to be a bad scientist. A human, for that matter, is also capable of evaluating goals and of monitoring his own internal processes (though not terribly accurately) to evaluate them in light of his goals.

These abilities would have evolved largely to deal with other humans, according to what Darwin called sexual selection. Its purpose cannot be understood without understanding how it works, which makes it a much more difficult engineering problem than that of understanding a machine whose purpose is determined by largely independent features of the environment. How is it possible to model a program that exists to model other versions of itself?

Perhaps I am mistaken but the science of human conscious thought and social psychology seems characterized less my overarching theory and more by a set of tantalizing observations. The possibility that the conscious process exists for the purpose of interacting with other conscious processes may pose a difficulty for my theory of reverse engineering and it may be that the correct approach to studying the conscious process has not been described in this essay.

Conclusion

Understanding Behaviorism has many obvious problems but its worst do not appear on the surface, but rather are so thickly woven into the author’s thought that he does not state them explicitly. From a psychological perspective, the only reason to read this book is as a case study in motivated reasoning. I am really pretty surprised that this book is taken seriously by other psychologists. It is suggestive of a severe philosophical problem within the psychological profession.