Black Boxes I

There are several reasons for the theoretical morass we find today. The most important of course is the lack of understanding of brain function in general and in particular with respect to psychological issues. Another is that adherents of many theoretical perspectives are able to help people. This makes for the tendency on therapists' part to take an Alfred E. Newman "Why me worry?" attitude when it comes to serious theoretical comparisons. They have a theoretical perspective they are comfortable with personally and which generally helps the clients who come to them, so they are happy, and generally don't see any point to attempts at developing a psychological theory that seeks roots in fundaments of human nature.

There are several reasons why therapists of many theoretical persuasions are able to achieve successes. One is that a self-selective process is at work in bringing client and therapist together. That is, a client stays with a particular therapist who is providing the kind of help that the client at this point in his life wants. What I am saying is that the word "help" is a very relative term. The most fundamental reason why therapists of various theoretical viewpoints are able to help people, though, is that the most important thing that a therapist brings to the clinical setting is his or her sense of empathetic intuition. A clinician who can relate elements of his own life to what his patient is experiencing adds an appreciation of humanity to the clinical equation that far surpasses the knowledge contained in any current theory, and probably any future theory. So in a sense, then, maybe you shouldn't be agonizing over the theoretical correctness of this theory or that. Be happy! Make a theoretical choice from the BOL list and run with it.

The rest of this note is for those who for some unexplainable reason are interested in attempts at psychological theories that seek rootedness in fundaments of human nature. How can such a theory possibly be constructed, given our ignorance of neuronal function and connectivity? I think that a model can be found in physics.

There was a time when scientists knew nothing about the fundamental constituents of matter. They had no clue about atoms, and so had no fundamental way to conceptualize how liquids, gases, and solids interacted with one another. They could have thrown up their hands in resignation and advocate that no theory be constructed until more fundamental scientific facts were discovered--as some suggest we should do with psychological theory--but that's not in fact what they did. Instead of waiting for a microscopic theory--one expressed in terms of atoms and their interactions--they developed a macroscopic theory--one expressed in terms of properties that were measurable with their crude instruments. This was the theory of thermodynamics.

The striking feature of thermodynamics is that everything in the theory is based on only three ad hoc assumptions. Thermodynamics is a physical counterpart to geometry. It is a logical structure in which everything is derived from a small number of axioms. And because those axioms were wisely chosen, thermodynamics for many years seemed to explain everything about how gases, liquids, and solids interacted with themselves and each other in considerable detail.

Eventually thermodynamics was found to be inadequate in several respects, but that did not detract much from its usefulness, because as its replacement theory--statistical mechanics--was being developed in terms of the interactions of atoms and molecules, thermodynamics served as a measuring stick for the new theory. To be accepted, the new theory had to explain everything thermodynamics did and what it couldn't.

Any psychological theory developed today would not be a final theory, but could be useful in itself for a time and as a stepping stone in the future, just as thermodynamics was. Actually such a theory could have a head start on a thermodynamic-type theory, because it could be based on a mix of macroscopic and microscopic axioms. We, after all, do know about neurons and something about how they interact. True, much of this knowledge relates to relatively simple physiological accomplishments--such as how we see--however, at least some of what has been learned about general neuronal functioning in this context may be applicable to the psychological realm.

So let's begin. Our goal is to indicate how human behavior can be described in terms of neural networks.

There is a good deal of research into how neurons function and are interconnected. Little of this is conclusive and virtually none of it pertains directly to psychology. For this first-order theory, however, we need little more than the fact that the nervous system has many interconnected networks, which for our purposes are so many interconnected black boxes. One of the few things we need to know about these black boxes is their general function. Happily, this much is known about them. I refer to the fact that these networks are apparently all input-output devices. Signals come in, are processed in some way, and then resultant signals are sent to other neuronal black boxes. The net effect of all of this communication and signal processing is often overt behavior.

Although we have apparently said very little, what little we did say has enormous implications. Our memories are in those black boxes, and if those boxes are input-output devices, then we cannot have inside us a stream-of-consciousness, personal record of every moment of our life, as neurosurgeon Wilder Penfield suggested. What do we have? That answer depends on the point of all this inputting and outputting.

The quick answer is that the point is to provide us with appropriately adaptive behaviors, so that we can meet the day-to-day requirements of our lives. This again is one of those deceptively simple statements, deceptive because adapting does NOT mean simply acquiescing to the requirements imposed from outside. These black boxes, taken together, have imperatives, too; we refer to these by saying that people have needs. So as far as these black boxes are concerned, life is an ongoing attempt at problem solving whose aim is to effect a reconciliation between what the outside world seems to allow and what the black boxes taken as a networked whole "want." The latter defines our needs, and presumably happiness is to be equated with the fulfillment of these needs.

This has rather profound consequences for the concept of human freedom, because it means that the concept has no meaningful application except in relation to the fulfillment of these imperatives. We are all slaves to our needs; we have the freedom only to screw up by making choices that separate us from their fulfillment.

Of course, the matter is a bit more complicated than that. That our decision-making processes can lead us away from our needs is testament to how easily these processes are influenced. One influence, of course, is the developmental process whereby our decision-making capacities were shaped in childhood. During this time, a person can be led to mistakenly believe that fulfillment of certain needs is either dangerous or forbidden. The various psychological theories provide rationales whereby therapists are led to help clients recognize and rectify their mistakes.

The object of therapy is to help clients become "normal," which means helping the client to attain roughly the same decision-making capacity as other people in society. Societal values and practices are another great influence on a person's decision-making capacity. Unfortunately, no one has developed a theory about how societies should be constructed so as to maximize the chances that their citizens will become wise enough to satisfy their needs. As a result, the best notion that humans have come up with so far is to frame societies so as to permit citizens with the maximum amount of external freedom, in the hope that the citizens will somehow learn how to make correct choices for themselves on their own.

We are getting ahead of ourselves. Returning to the network of back boxes that is our nervous system. Generally speaking, experiences have two sorts of impacts on us. One is to help fill each black box with neurons, determine how each neuron is hard-wired with other neurons in the box, and how each box as a whole is hard-wired to other boxes. Such hard-wiring changes mostly occur during infancy and early childhood. Another type of change as a result of experience is more liable. It seems to operate by making some interconnections--both inside and outside the boxes--temporarily more important than others. This rerouting can lead to the same inputs triggering different outputs, and thereby causing changes in behavior. We tend to label these potentially temporary experience-borne adjustments as learning.

As a child matures, hard-wiring changes diminish and learning takes over as the predominant cause of behavioral change. The "talking-cure" therapies assume that clients arrive hard-wired normally, and that any symptoms the client has relate to learning problems. This is probably a good assumption for most clients, but may not be true of those having a brutal infancy and/or early childhood.

Assuming normal hard-wiring, how should we conceptualize the relationship between the hard-wiring and the learning that gets impressed upon it? Should we assume that the hard-wiring is merely a passive recipient? That really doesn't make a whole lot of sense, if you think about it, because it implies that a child is a passive agent in its own learning, which we know is not true. This brings us to the conclusion that the hard-wiring itself can operate as source of developmental motives whereby the child seeks the experiences that will teach it what it needs. Learning theory speaks not only of motives, but of readiness, too, which implies that there must be some structural aspects to this developmental process.

One component of this structure comes from one of the types of black boxes in our nervous system. These boxes are devoted to associative memory whereby inputs from certain black boxes are paired with outputs to other black boxes. What's important about this type of memory is that associations are not made in a willy-nilly fashion, as Freud supposed. Rather, linkages are established on a categorical basis. This type of memory also links and groups the categories.

Once one realizes that interrelated categorization accounts for much of a child's learning, it becomes clear that the child requires a tremendous amount of innate guidance, especially initially. This is handled by what I call developmental instincts.

There is evidence that an infant senses early on that mommy and daddy are different and that it quickly comes to see itself as belonging to either a mommy or daddy category. The imperative to make this distinction and relate it to itself is one developmental instinct. Another is the so-called egotism of the child, the tendency to relate every experience to itself and to see itself as the cause of much that happens to it and its family. If you think about it, without this egotism, a child couldn't internalize many lessons about itself.

A third developmental instinct leads the child to identify with and imitate both parents, with an emphasis on the parent of the same sex.

A fourth instinct was discovered by Freud, who conceptualized it in terms of oedipal strivings. This is the imperative to compete with the parent of the same sex for the other parent's attention and affection.

Play is yet another developmental imperative.

Following these procedures, the child learns lessons about itself that another instinct causes it to view as representing truth in the absolute. That is, the criticisms it receives from its parents are not accepted merely as the opinions of particular human beings. Rather, they are endowed with "awesome authority"--to use Joseph Weiss's phrase. The child acquires a worldview that it will thereafter take for granted as representing what is to be expected from itself, others, and life in general. In short, the child is led to "overgeneralize," to again quote Weiss. This worldview becomes the person's touchstone to reality; it forms the person's basis for judgement. It's this fact, more than any other, that accounts for the continuing participation of childhood lessons in adult behavior. This worldview may be thought of as consisting of interrelated categories of thought. What a person tends to do later in life is live within the context of this general categorical framework.

Memories fade, so why don't elements of this framework fade in time? The answer has several components. Involved is the fact is that the input/output mappings computed by individual networks inside the black boxes are not precise. That's because neurons themselves are not precise devices; they fire probabilistically and cannot fire during "rebound" periods after having just fired. This lack of precision underlies the fact that these networks are able to generalize, in the sense that similar inputs can produce the same output. Most of the childhood lessons that are the subject of psychological inquiry during therapy involve mommy and daddy, the representatives of maleness and femaleness. Everyone else the person meets later in life falls into one or other of these two categories, and as such evokes aspects of what the person learned as a child. This happens every day. That's the quick and dirty answer to why these memories are kept alive.

What I am saying is that the highfalutin concept of transference is an everyday occurrence. It's just more obvious in therapy because the clinical setting more closely mimics the person's childhood situation. The person is in therapy to learn lessons about life in much the same way that contrary lessons were established in childhood.

Apparently because of the lack of precision in neuronal activity, much of the brain's operation appears to be governed by constraint satisfaction. I mean that the brain often seems to use many different sources of information at the same time in an effort to arrive at behavioral responses that satisfy all constraints simultaneously. We respond to life situations in terms of multiple clues, each of which acts as a weak constraint on our behavior. The requirement that all of the weak constraints must be satisfied at the same time makes for a very strong constraint, which helps ensure that our behavior will be appropriate to the situation we face, taken as a whole.

As an example of this sort of thing, consider the case of someone who received a very disappointing raise, someone who thereafter conjures up images of himself storming into the boss's office and telling the skinflint off. On the basis of this very limited and self-generated sensory input, the person is able to vent his anger and get everything off his chest. But when he arrives at work the next day with the intention of repeating his performance, he finds he has second thoughts simply as a consequence of physically being in the office. Then when he sees that the boss is in a bad mood, he decides to say nothing at all. Or perhaps he does complain, but finds himself complaining meekly. The point is that there are many components to our behavior, and these components are shaped by what we experience, taken as a whole.

Another example of that has to do with transferences. A bit of how we responded to our parents is to be found in our later responses to men and women generally, with the extent of the participation being determined by how closely our perception of a present situation mirrors that of a situation we faced as a child. Our responses as a whole, however, are based on past experiences from many different periods in our life.

Let's now look at how we respond on the basis of past experience in some detail. Experiencing a life situation has the power to automatically elicit appropriate responses we learned in the past. This is not to say, however, that what is elicited are necessarily those responses in their entirety. We need to distinguish three cases:

1. Familiar tasks we perform everyday. There is evidence that we do not learn from performing such tasks in a rote manner, which would suggest that they are evoked in their entirety.

2. Familiar tasks we haven't performed recently. We are a bit clumsy is doing something we haven't done for a while. If we start performing this task everyday, however, we quickly become skillful again, which shows that the old skill was still in memory. Our problem was mainly that some recall connections needed to be reestablished. The fact that we quickly become skillful again shows also that subsequent to our clumsy performance some learning must have taken place whereby the "forgotten" components to our performance were integrated into our responses once again.

3. Tasks that are new to us. This time, the experience of attempting to perform the task evokes behaviors that are very rudimentary along with problem-solving routines we used when facing new tasks before. Subsequent to performing these tasks, a lot of learning takes place.

Let's now restate everything in terms of the black boxes. Sensory input from a life situation impinges on our nervous system, sending inputs and outputs zooming around from black box to black box, with the result that we behave in some appropriate way. But that can't be the whole story, because if it was, there would be no explanation for the learning that accrues from that life situation. Learning implies that the black boxes must make adjustments to themselves after the situation has been brought to an end.

Of course, we don’t experience just one situation in a day, but a continuous series of situations, each one of which must trigger a subsequent learning response. Either we must assume that our brains can accomplish this subsequent learning in a flash or that they run the risk of being overloaded with attempts at learning as our waking intervals lengthen.

Some of our learning does take place in a flash, while we are dealing with the situation at hand. We try one response and then another to learn about the outcomes of each, and from this data decide how best to proceed. But the evidence is that more learning takes place afterward and that this learning proceeds over a period of time. So while our black boxes are providing us with appropriate behaviors here and now, they are also engaged in learning from the waking interval's previous experiences. It would seem, then, that there is danger of becoming overloaded with new information as a waking interval lengthens. What is needed to prevent an overload is a respite from experience. The only respite of that sort that we have--other than coma and death--is sleep. It must be, therefore, that sleep is at least partially devoted to giving our black boxes a chance to make the learning adjustments they need in response to the experiences of the previous waking interval.