Cognitive Darwinism: rational-emotive therapy and the theory of neuronal group selection.

5. A. Ellis. Rejoinder: Elegant and Inelegant RET, in A. Ellis and J. Whiteley, eds. Theoretical and Empirical Foundations of Rational-Emotive Therapy. Belmont, CA: Wadsworth, 1979.

6. G. M. Edelman. The Remembered Present: A Biological Theory of Consciousness. New York: Basic Books, 1989.

7. A. T. Beck. Cognitive Therapy, Behavior Therapy, Psychoanalysis, and Pharmacotherapy: A Cognitive Continuum, in M. Mahoney and A. Freeman, eds. Cognition and Psychotherapy. New York: Plenum Press, 1985. (Originally published in 1983.) Introduction and Philosophy of Science

ALBERT ELLIS'S THEORY of Rational-Emotive Therapy (RET) shares many basic ideas with general semantics theory.(1) Several of Ellis's books have appeared in E-Prime (2), a writing and speaking methodology developed by David Bourland and used by some general semanticists.

This paper explores the relationship of Ellis's theory of psychotherapy and personality to a particular theory of brain function, Gerald Edelman's Theory of Neuronal Group Selection (TNGS). (3) Philosophically, this inquiry is an exercise in "interfield theory," whose relevance to cognitive science has been cogently outlined by William Bechtel. (4) Such an interfield theory aims at a "mechanistic explanation" of emotional disturbance, in addition to purely "functional" analysis. Rather than attempting the "reduction" of psychology to neuroscience, however, interfield theory seeks to "bridge boundaries," rather than "break boundaries," between the two theories. Such an approach seeks to examine the same phenomena at different levels of organization, and also to examine emotional events at the same level from two different points of view.

It should be emphasized that this exercise in "interfield theory" must base itself on the very same assumptions which underly the two parent theories. Every scientific theory must remain underdetermined, to some degree or another, by the available empirical evidence, and RET and TNGS both have plenty of competition within their own respective fields. But since both theories deal at some level with the same phenomena, it should be instructive to demonstrate how these two particular theories could explain the same events. It is likely that different "interfield" programs might also relate alternative theories from each of the two fields.

Specifically, I want to explore more precisely the distinction between Ellis's "elegant" and "inelegant" forms of RET practice. (5) In addition, because RET has frequently claimed a biological basis for the human propensity for "irrational beliefs," I explore this physiology in the present article. Finally, I seek to clarify the relationship between RET and psychopharmacology.

Ever since Egyptian surgeons cracked open the brain 5000 years ago, humans have tried to imagine mechanical models of the brain's functioning. These machine analogs have usually compared the brain's processes to man-made devices of the modeler's own particular historical period -- e.g., pipe organs, steam engines, telephone switchboards, digital computers, computer networks, etc.

Instructionism and Selectionism in Biology

Edelman has chosen as his own analog the humoral, antibody-mediated immune system. This example requires Edelman to remain scrupulously selectionist in all aspects of the theory. At this point, let us look briefly at the differences between instructionism and selectionism in biology. The distinction goes back to the nineteenth century debate between "Lamarckian" and "Darwinian" explanations of evolution. The "instructionist," or "Lamarckian" paradigm suggested that the individual organism gets "instructed," or molded by its environment. In other words, the individual organism adapts to its environment by acquiring more adaptive characteristics, and then passes on those acquired adaptations to its progeny.

On the other hand, the "Darwinian," or "selectionist" paradigm hypothesizes that adaptive characteristics already exist among variant individuals within a biological population. The environment then selects those individuals which exhibit these more adaptive characteristics. In evolution for example, these adaptive characteristics become enhanced within the population over generations as the individuals which bear them gain the ability to reproduce in greater numbers than their less favored brethren.

Edelman has called his approach "Neural Darwinism," because he has based it on previous selectionist theories which have been applied to the evolution of species, and also to the adaptation of an individual organism's immune system to particular antigens in its immediate environment. Edelman's original construction is called the "TNGS Proper," and it deals primarily with perception, memory, and learning. More recently, Edelman has proposed, an "Extended TNGS" which goes much further and attempts a brain-based theory of consciousness itself. (6)

This paper will seek to apply some of Edelman's ideas more specifically to the cognitive realm. In particular, it will explore their relationship to Ellis's (RET) version of cognitive-behavioral psychotherapy (CBT). Aaron Beck, the inventor of "Cognitive Therapy," has pointed out that there are still many holes in the power of cognitive, behavioral, psychodynamic, or pharmacotherapeutic analyses to explain common clinical events. (7) RET itself offers no explanation for the frequency and persistence of irrational beliefs in both disturbed and "normal" humans. (8) I suggest that a combination of TNGS and RET may further explain emotional disturbances as the result of some particular concepts about the world operating on a particular neurological substrate of human consciousness.

TNGS uses a specialized vocabulary, so please refer to the following glossary.

GLOSSARY OF TNGS TERMS

Concept: A categorization by the brain of its own previous global mappings. In this way, concepts relate different perceptions to each other, even in the absence of those environmental stimuli which engendered those perceptions in the first place. Conceptual ability gives an animal the ability to control its own behavior in a general way, based on conscious generalizations about the environment.

Degeneracy: The characteristic of a biological recognition system which allows several structurally different recognition groups to perform the same function and "recognize" the same given input or stimulus. Degeneracy is distinguished from "redundancy," in which several structurally identical groups perform the same recognition function. Degeneracy leads to greater range than redundancy, but less specificity of recognition.

Epigenesis: A process of development which is bounded by genetic constraints, but in which morphology is also affected by non-genetic, environmental factors which are present during development. Epigenetic development in the brain leads to a uniquely different brain physiology for every single individual, even for identical twins with the same genetic program.

Gesture: One of the degenerate sets of all those coordinated movements that produce a particular pattern of action in an individual. This includes all types of voluntary movement patterns, and thus constitutes a somewhat broader definition than the normal meaning (in which only those movements which signal "body language" are normally considered to be "gestures").

Global mapping: An action of the whole brain which ties together many simultaneous motor and sensory maps of the environment with the behavior of the animal. In this way, the brain correlates various representations of individual features of the environment, in order to form a continuous representation of objects or events.

Local map: A translation by the brain of certain physical properties of real-world objects and events into a neural representation at well-defined locations within the cortex. Local maps constitute representations within the brain of selected features of the environment which are recognized by pre-existing neuronal groups.

Neuronal group: A group of neurons consisting of hundreds or thousands of individual neurons which are more closely connected internally to each other than they are externally to neurons outside the group. Neuronal groups are collected in "repertoires," which are polled to determine which groups best "map" an environmental event.

Polymorphous set: A set which can be disjunctively characterized by several different, alternative collections of various features. A polymorphous set differs from the classical view of a "category," in that it is not defined by a collection of features which are singly necessary and jointly sufficient.

Primary consciousness: The state of having mental images in the present time, which provide a continuous scene of ongoing events. Primary consciousness differs from "higher-order," or human consciousness, which involves becoming aware of one's own consciousness, and being aware of a personal past and future.

Primary repertoire: The collection of neuronal groups which are present in an individual's brain at birth. According to TNGS, this internal "wiring" will not change during the individual's life.

Re-entry: Communication between brain regions, in which the "efferent" output of one region is "re-entered" as "afferent" input to another region. In TNGS, different neuronal groups simultaneously recognize different categories of the same environmental stimuli. Association between different classifications of the same stimulus occurs when an internally generated signal from the output of one "recognizer" is re-entered as input to another "recognizer," as if it were an external signal.

Secondary repertoire: The 'brain's collection of neuronal groups which' have already been selected by experience. Some of them will have been internally strengthened by use, and others internally weakened by lack of use, during life experience.

Synergy: A class of related gestures which produce the same adaptive pattern of action. Synergies represent the units of control employed by the brain in controlling movement.

Value-category memory: Located in the frontal, temporal, and parietal cortex, this module stores internal evaluations of previous perceptual categories, in order to evaluate ongoing, real-time perceptions, before these perceptions can enter conscious awareness.

Epistemological Background

TNGS mainly concerns itself with how the brain generalizes; that is, how it forms categories of different but related stimuli. Edelman's theory is based on the philosophy that the stimuli which confront the human brain in life become grouped by the brain into "polymorphous sets." (9) These are defined as sets which can be variously categorized by alternative, disjunctive collections of different features. Individual features are thus neither singly necessary nor jointly sufficient attributes of a polymorphous set.

It may be useful to compare polymorphous sets to sets which are essentially defined, and to those which are arbitrarily defined. Let us suppose that there are n possible identifying features of a set. A polymorphous set would be characterized by any combination which includes a minimum number m (which is more than 1, but less than n) out of n features. If all n features were singly necessary and jointly sufficient for membership, this would constitute an essential set. If only 1 feature defined membership, this would constitute an arbitrary set.

For example, a bird may be identified polymorphously as a bird, by any number of different, disjunctive combinations of identifying features, e.g.:

a. Feathers, wings, small size, animal; or

b. Sings, lives in trees, flies, lays eggs; or

c. Kept in a cage, can mimic human voice, etc., etc.

This epistemology implies that the world must appear a completely unlabeled place. The brain can never know in advance which features to search for to identify similarity. Therefore it must sample many different features at the same time. By correlating these many different samplings simultaneously the brain can create its own conditional, polymorphous sets out of the unlabeled environment. Each brain retains only those polymorphous sets which turn out to identify an adaptive similarity. (10)

Selectionism in the Immune System

The immune system contains another biological "memory" system, and Edelman's approach to the brain parallels his own previous study of the immune system. The key features of his Darwinian paradigm are: variation within a biological population, selection by the environment of the more adaptive variant individuals, and enhancement of the adaptive characteristics within the population. This paradigm has been successfully applied to the following major biological theories:

1. Evolution: Individuals within a species demonstrate great variability. The environment "selects" the more adaptive variant individuals by allowing them to reproduce in enhanced numbers. The adaptive characteristics survive and are inherited by the progeny via genetics.

2. Immune System: Edelman won the 1972 Nobel Prize for confirming a Darwinian theory of the immune system previously suggested by MacFarlane Burnet and Niels Jerne. Each individual is born with a certain finite "repertoire" of antibodies. Environmental antigens "select" the more adaptive antibodies via selective binding. These adaptive antibodies then clone to produce replicas of themselves. The adaptive characteristics of the selected antibodies become strengthened within the repertoire by cloning, or asexual reproduction.

Selection Within the Brain

Edelman's essential insight is that the brain itself does not receive "instruction" from the environment, in that it does not mold itself to the environment or "trace" a mental representation of the environment. Further, the brain differs from a digital computer in that it uses no fixed categories, and the brain has no "program" of symbol manipulation. Table 1 illustrates some of the empirical differences between a brain and a digital computer.

Edelman hypothesizes that the brain adapts in a Darwinian way to its environment. He considers the brain as composed of biological "populations" of individual "Neuronal Groups," each of which consists of hundreds or thousands of neurons which are more closely connected to each other, than they are to neurons outside the neuronal group. Recognition "repertoires" within the brain contain similar, but variant, "Neuronal Groups." The variant Neuronal Groups within a particular brain "repertoire" compete among themselves to respond to incoming stimuli. Each member of the neuronal group repertoire has its own unique pattern of connections, so it responds to the same outside stimulus in its own unique way. That is, the neuronal groups within a particular repertoire are "isofunctional," but not "isomorphic." None of the Neuronal Groups can be an exact "map" for any particular stimulus. However, for any particular stimulus, some of the groups match better than others.

Darwinian selection within the brain takes place when the more adaptive Neuronal Groups become strengthened by enhancement of synapses within the group, making that group more likely than other members of its repertoire to respond to similar stimuli in the future. The brain's wiring itself does not change; rather the relative responsiveness of individual neuronal groups varies within each repertoire. For adaptive Neuronal Groups within a repertoire (i.e., those that have successfully "recognized" stimuli) the synapses are strengthened; for non-adaptive groups the synapses are weakened.

TABULAR DATA OMITTED

TNGS Proper

TNGS proper is based on three basic hypotheses:

1. Morphogenesis: During development prior to birth, each individual brain forms a completely unique "primary repertoire" of Neuronal Groups. Even identical twins, with the same genetic program, will develop different primary repertoires which possess unique anatomical variation at the finest neuronal branches. Development of the brain is never completely determined by genetics, but instead is guided by a process called "epigenesis," in which the environment dynamically interacts with the genetic code during development to produce anatomical variation.

2. Synaptic Selection: After birth, anatomical connections between neurons remain fixed. However, during behavior as a result of neural signaling, the more adaptive Neuronal Groups become selected, by strengthening of the synaptic connections within them. The synaptic connections within less adaptive Neuronal Groups become weakened. Those neuronal groups whose internal connections become "strengthened" will then become more responsive to similar stimuli in the future.

Thus, without alteration of anatomy, a "secondary repertoire" is created which consists of selectively strengthened and weakened Neuronal Groups.

3. Re-Entrant Signaling: Through independent sensory and motor channels, the brain continually samples independent features from the environment. This sampling creates parallel "local maps" of different feature collections. The brain correlates these parallel mappings by a communications process called "re-entry". Re-entry is a form of communication which occurs only in the nervous system, in which the "efferent" output of one region is "re-entered" as "afferent" input in another region, as if it were an external signal. Re-entry enables the brain to generalize and form its own polymorphous categories from the environment.

The brain's self-created "polymorphous sets" become its basis for generalization, by which it responds adaptively to future stimuli.

Re-entry in TNGS

Edelman's concept of "re-entry" appears somewhat difficult to grasp intuitively, so let's take a moment to look at it. (11) The idea originally derived from cybernetics (12), and was first proposed on theoretical grounds as a method for the CNS to perform simultaneous "feed-forward" and "feed-back" functions to control voluntary movement. Subsequently, anatomical evidence of re-entrant connections was discovered throughout the brain.

In Edelman's theory, when the brain perceives a stimulus, it must coordinate many "recognitions" of different features from different channels simultaneously, in order to come up with a composite image. Let us take a look at one specific path of such a recognition. Say that an outside stimulus is transduced to one particular feature detector of the brain. At the first stage, this signal "polls" the repertoire of candidate feature "recognizers" ("R" groups) to find approximate matches. Since these "R" neuronal groups within the isofunctional repertoire exhibit degeneracy, more than one group will recognize the same stimulus approximately, and none will recognize it exactly.

Then, at the next level of abstraction, the output from a successful "R" group will poll the repertoire of candidate "recognizers of recognizers" ("R of R" groups). This process proceeds on to higher and higher levels of abstraction. Now, in order for these "recognitions" to become associated with recognitions of other features from the same stimulus, the output of one successful "polling" is "re-entered" to the input of another repertoire of different feature "recognizers," ("|R.sub.2~" groups) as if it were an external signal.

In this manner, re-entry is able to perform the following functions in TNGS:

1) It guarantees continuity within the neural construct by continually comparing signals from adjacent times and locations with each other;

2) It allows the system to recognize novelty in the environment, by comparing new and reentrant signals with each other;

3) It accounts for the fact that memory becomes distributed throughout large portions of the brain;

4) It accounts for the associative and content-addressable nature of the brain's memory storage.

Perception, Memory, and Learning

TNGS proper has the following general implications for mental functions:

1. "Degeneracy": The "maps" within the brain cannot precisely trace or mold environmental stimuli, nor can they depend on fixed "categories" of the environment. Rather they match environmental stimuli with pre-existing internal Neuronal Groups which more or less represent certain environmental features. In general, more than one Neuronal Group within a repertoire can match approximately to the same feature and none will match exactly. In Edelman's terminology, all mappings are "degenerate." Degeneracy ensures both reliability in case of error, and the range to adapt to unpredictable events.

2. Global Mappings: The act of perception really involves recognizing and correlating shifting, self-created categories. These categories are formed into a perceptual image by a process called "Global Mapping," which combines parallel sensory inputs from local maps, spatio-temporal coordination of all the local maps with each other, and associated behavioral output of the animal. All aspects of the global mapping perception communicate with each other by re-entry.

3. Memory in the brain: Memory does not involve any trace or replica of a previous perception. Instead the process of memory recall is itself a dynamic act, and it really requires an ability to re-categorize past inputs. As a result, each recall becomes a unique creation of the brain, based on past behavior and current context.

Extended TNGS and Consciousness

The above ability to categorize environmental stimuli into correllated polymorphous classifications leads to being able to generalize from experience, in order to benefit from similarities between past experience and current situations. This ability -- executed by "global mappings" within the brain -- lies at the basis of animal perception and action. Even pigeons possess the ability to discriminate extremely complicated classifications of visual stimuli, such as pictures of different human faces.

However, perceptual categorization alone may be insufficient to cope with a complex environmental scene, particularly for an animal with a diverse behavioral repertoire. "Consciousness" requires the capacity to decide which signals from the environment are most important to the individual animal at that particular time, and should therefore be given salience (or "awareness") over simultaneous, competing stimuli. This latter ability constitutes what Edelman calls "primary consciousness." Primary consciousness involves "developing a composite 'image,' relieved to some extent of the immediate flux and variation of (individual) signals." (13)

Edelman proposes in his "Extended TNGS" that this primary consciousness results from a constant comparison between the neural recognitions of two separate modules within the brain:

1. The "interoceptive" module of primary consciousness involves the brain stem, hypothalamus, and autonomic nervous centers. It is concerned primarily with registering internal states of the animal. These concerns include "adaptive, homeostatic, and endocrine functions of the individual animal.... These functions of the individual animal relate to its immediate needs of survival and reflect evolutionarily selected values that have contributed to fitness." (14)

2. The "exteroceptive" module of primary consciousness includes the cerebral cortex. It is dedicated to perceiving (categorizing) signals, mainly those received from the external world.

Edelman's "Extended TNGS" proposes that "primary consciousness" results from the ongoing comparison of the workings of the interoceptive and exteroceptive brain modules.

The mechanism for this comparison depends on the following functions:

1. Value-Category Memory: This special memory area is localized in the frontal, temporal, and parietal brain areas. Its job is to store previous matchings of internal, homeostatic, values to perceptual categorizations of the outside world. These matchings result from linking perceptions experienced by the exteroceptive areas, with self-oriented, species-specific values generated by the interoceptive areas. With experience, this value-category memory becomes filled with previously experienced perceptual categorizations which have been internally evaluated by the interoceptive system. The value-categories change with each new experience, as well.

2. Re-Entrant Communication between current, ongoing perceptions, and the accumulated value-category memory: This ongoing communication creates a composite "scene," which gives importance to particular current perceptions, based on their evaluations which have been remembered from earlier categorizations of similar stimuli.

In this way, primary consciousness "evaluates" ongoing perceptions, before these perceptions are brought to conscious awareness. (15)

Symbolic Representation of Primary Consciousness

Primary consciousness may be represented symbolically as follows:

The letter "C" standing before a parenthesis stands for "Categorization of ..."

"I" stands for interoceptive or "self" input.

"W" stands for exteroceptive or "world" input.

A dot (.) between symbols means linkage through neural connections or synaptic change.

C(W) represents the exteroceptive categorization (perception). C(I) represents the interoceptive categorization (evaluation). C(W).C(I) represents an interoceptively evaluated perception. The value-category memory is created by the accumulation of:

C|C(W).C(I)~

Each of these represents a class of previously evaluated perceptions.

Primary consciousness itself is then achieved by the ongoing interaction of value-free current C(W) with stored C|C(W).C(I)~. These interactions "lead from perceptual categorization to perceptual experience." (16) The ongoing comparison by the conscious mind between the interoceptive and exteroceptive brain areas, constitutes what Edelman terms "the remembered present." Edelman's theory of primary consciousness is illustrated schematically in Figure 1. Consciousness results as "current value-free perceptual categorization interacts with value-dominated memory before further contributing to alteration of that memory." (17)

Concepts and Global Mappings

Recall that a global mapping employs a large part of the entire brain to correlate inputs from multiple senses with outputs for behavioral responses. This creates a categorization of simultaneous input and output signals into a polymorphous group.

Edelman hypothesizes that the brain forms concepts by utilizing special "structures that can categorize, discriminate, and recombine patterns of activity in different kinds of (previous) global mappings." (18) Thus, concepts allow the brain to generalize its own classes of previous internal activity. In this manner, the brain "is able to identify a particular thing or action (as belonging to a general class) and control its future behavior on the basis of that identification in a more or less general way." (19)

In Edelman's hypothesis, concepts are pre-linguistic. Even lower animals which do not possess language may nonetheless be capable of having and using concepts. Edelman sees chimpanzees, and possibly even dogs, as possessing conceptual ability. Edelman states that "in the individual, concepts precede language and meaning. They are driven by the perceptual apparatus." (20) Since concepts are created from the brain's own global mappings, it follows that all concepts should themselves remain tied to internal, homeostatic values. "Concepts serve intentionality." (21)

Rational Emotive Therapy

Rational-emotive therapy belongs to the general movement of cognitive-behavioral psychotherapies (CBT), all of which hypothesize that disturbed emotional states -- especially depression, anxiety, and anger -- result from disturbed cognitions. Ellis (22) has specified RET's particular theory of emotional upset with his "ABCDE" schematic model. (See Table 2.) In this model, "A" stands for the "activating event" in a person's life; "B" stands for his "beliefs," or evaluations about "A"; "C" stands for his disturbed emotional consequence; "D" stands for therapeutic disputations against the "irrational" components of "B"; and "E" stands for the new, "effective" beliefs which result from "D." In Ellis's model, the activating event, "A," does not directly cause "C," the emotional consequence. Rather, the subject's evaluative beliefs about the event ("B") are responsible for his own individual emotional response.

Ellis's scheme is diagrammed in Table 2. Each individual responds to a particular event with his own collection of both "rational beliefs" (rB's), and "irrational beliefs" (iB's) about that event. In general, his rB's express preferences about the event, while his iB's express demands. Ellis maintains that it is only the individual's iB's about this event which cause and sustain his emotional disturbances.

Ellis uses a specialized terminology, and also uses some common terms in a specialized way, so the following glossary may help.

GLOSSARY OF RET TERMS

Beliefs: Ellis's "beliefs" should be contrasted with similar constructs from cognitive-behavioral therapy, such as Beck's "automatic thoughts" (23) and Meichenbaum's "self-statements". (24)

Beck's automatic thoughts (AT's) consist of inferences and predictions about events, and Beck has concentrated on identifying and correcting reasoning errors common to "dysfunctional" inferences and predictions. Meichenbaum's "self-statements" consist of covert self-instructions which people use to guide their own behavior.

On the contrary, Ellis's "beliefs" contain specifically personal evaluations about events. (25)

Elegant RET: RET which stresses the achievement of profound and lasting cognitive or philosophic change in people's demands that they themselves, other people, and the world conform to their wishes. Elegant RET seeks to change these demands to preferences.

Irrational beliefs: "(1) They have at their core explicit and/or (usually) implicit rigid, dogmatic powerful demands and commands, usually expressed as musts, shoulds, ought to's, have to's, and got to's....(2) They have, usually as derivatives of these demands, highly unrealistic, overgeneralized inferences and attributions." (26)

Rational: Again, Ellis uses the word in a specially defined way. Ellis takes pains to distinguish his own "rational beliefs" from classical rationalism. (27) RET defines "rational" simply as self-helping, and "irrational" simply as self-defeating. This is distinguished from the classical philosophy of "rationalism," which derives from Descartes, and views the mind as predominant in organizing experience on the basis of preexisting schemes.

"Elegant" RET and the "Neglected Cognitive Element"

In 1979, Ellis distinguished between the "elegant" and "inelegant" forms of RET practice. In that paper, he clearly delineated exactly what makes RET different from other forms of cognitive-behavior therapy (CBT). (28) This difference appears mainly in RET's "elegant" or "preferential" version. Elegant, preferential RET, attempts to change a person's fundamental, disturbance-creating philosophies. These philosophies include variations of the following three demanding beliefs:

"I must do well." "You must treat me beautifully" "The conditions I want must be easily available."

In a brilliant 1987 article entitled, "A Sadly Neglected Cognitive Element in Depression," (29) Ellis clarified this argument still further. He analyzed four previous cognitive-behavioral models of depression:

1) "People become depressed when (a) they view themselves negatively, (b) they view their environment negatively, and (c) they view the future negatively." (30)

2) "People make themselves depressed when they give themselves low rates of self-reinforcement and high rates of self-punishment and/or see themselves as receiving less reinforcement than they actually receive." (31)

3) "People become depressed when there is a lack of pleasant events in their lives or a continuing low rate of response-contingent reinforcement." (32)

4) "People who expect that highly aversive things will happen, who also expect that they will be able to do little to improve them, and who tend to attribute internal, stable, and global causes to negative events and to attribute external, unstable, and specific causes to positive events will tend to depress themselves." (33)

Ellis hypothesized that the conditions outlined in these models should not necessarily be sufficient to lead to depression. Many people might experience all four of these conditions, and yet not become depressed. He hypothesized that in addition to any of the above cognitive elements, depressed people possess an additional philosophy that these things must not, should not happen, and that it is awful or terrible that they do happen. However, this still remains an hypothesis, and Ellis has not yet been able to prove that this philosophic cognitive element constitutes a necessary condition for human depression.

Neural Darwinism and the Elegant Disputation in RET

What possible implications may TNGS have for RET?

TNGS implies that "identical" events should trigger different evaluative beliefs in every individual. This is because every individual must rely on his own "primary repertoire" of neuronal groups, which is uniquely created during development prior to birth, and on his "secondary repertoire," which is uniquely created by his experience after birth. As a result, every individual will "create" his own perceptions by correllating his own unique "polymorphous sets." Further, every individual's "value-category memory" will uniquely evaluate these perceptions before they rise to awareness.

Next, the thing that distinguishes RET from other forms of cognitive-behavioral psychotherapy is its "elegant" disputation. In practice, this involves assuming the very worst outcome about a situation and then rationally arguing or imagining or proving that even the very worst is not intolerable, and you can choose to try to change it or else even adjust to it, without hurting yourself with emotional disturbance.

Beck has put his finger on this unique method of RET. He discusses a depressed young student who erroneously believes that she would not get into college:

He (the therapist) could have used one approach advocated by rational emotive therapists...and challenged her belief that entrance into college was an index of her worth as a person. Using a different rational emotive technique, the therapist could have explored with her what the catastrophe would be if she could not get into college. (34)

This would constitute a different approach than advising her that she most probably would get into college after all.

If someone makes incorrect inferences, predictions, or assumptions about a situation (such as erroneously predicting failure to get into college), why not just prove that these predictions and attributions about the event are false or of low probability? This is one approach of CBT. However, "elegant" RET relies on two different kinds of disputations at the same time:

1. Empirical: This disputes the likelihood of the incorrect inferences, predictions, assumptions, and attributions.

2. Elegant: The "elegant" disputation (also called the "philosophical" dispute) asks the following questions: (35)

a. "What worst things could actually happen to me if I don't get what I think I must (or do get what I mustn't)?"

b. "What good things could I make happen if I don't get what I think I must (or do get what I think I mustn't)?"

Why should this "elegant" disputation be useful with empirical errors? Why can't a disturbed person just "reality test" his inferences against a "true picture" of the situation? The answer may be that the principle of "degeneracy" of TNGS ensures that there is never any precise, proportionally accurate replica of the situation within the person's brain. According to TNGS there are only personal, idiosyncratic, creative categorizations of the situation. Therefore the depressed student described above really possesses no metrically accurate perception to check reality against.

On the contrary, she has practiced categorizing this situation (worrying about getting into college) in catastrophic terms. Thus, the neuronal groups have been enhanced which execute this "catastrophic" evaluation. The reason her categorization is maladaptive is because her fear is out of proportion to the real situation. But, since her brain does not record a precise, proportional image of the situation -- she cannot call up her own mental "picture" to perceive that her interpretation is out of proportion. Other people can only tell her that their own perceptions differ from hers. On the other hand, since it already hypothetically assumes the very worst outcome anyway, the "elegant" disputation requires no metrically precise perception in advance.

Neural Darwinism and RET's "Insight Number 1"

Ellis proposes three types of "insight" in RET. Insight Number One of RET "sees self-defeating or disturbed consequences as stemming mainly from the irrational beliefs that people bring to the activating experiences occurring in their lives." (36) I propose that Neural Darwinism may provide an explanation of how a normal person could learn to hold such "irrational beliefs":

1. Degenerate Sampling: The principle of degeneracy ensures that there must always be some defect in sampling the environment. This approximate nature of categorization is necessary to "reconcile specificity of recognition with range of recognition." (37)

2. Re-entrant Illusions: Illusions in perceptions result from Neural Darwinism's re-entrant communication between feature recognizers. Re-entry correlates disjunctive features which are sampled in parallel, the output of one feature "recognition" becoming the input of another recognizing area. If the first feature "recognition" is mistaken, re-entry can perpetuate or enhance erroneous categorizations.

Edelman says "illusions are incorrigible; with experience, however, we can detect the incongruities in a perceived situation and even attempt to alter our view of that situation...Only with the advent of rich enough conceptual systems can generalizations be built, allowing us to distinguish between appearance and reality." (38)

3. Autonomic Arousal: It is possible that an erroneous categorization as produced above may serve to arouse strong internal homeostatic appetites which may dominate in the control of behavior. Imagine, for example, that the triangle illusion in Figure 2 was not just a harmless triangle, but instead an illusion of a deadly threat, for example an imminent assault on the life of the observer.

4. Attention: If the illusion led to a strong autonomic arousal, attention could then shift away from reexamining the perceptual illusion, toward dealing with the perceived immediate threat in an emergency fashion. This would prevent the deeper conceptual analysis which is necessary to pierce the illusion.

"Cognitive Darwinism" and Insight Number 2 of RET

RET's Insight Number Two states:

"That people remain disturbed by re-indoctrinating themselves in the present with their absolutistic beliefs." (39)

"No matter how we originally become (or make ourselves) disturbed, we feel upset today because we are still re-indoctrinating ourselves with the same kinds of irrational beliefs that we originated in the past. Even if we learned some of these beliefs from our parents and other early socializing agents, we keep repeating and retaining them ourselves. Therefore our self-conditioning is much more important, than our early conditioning on the part of others." (40) Why do both normal and neurotic humans keep repeating this re-indoctrination, if it continues to defeat their own goals? Ellis specifically refutes theories of unconscious control:

"The psychoanalytic and transactional analysis theories that parents impose cognitive tapes on their children, and therefore, especially in their inter-personal relations, maim them forever has considerable contradictory evidence to assail it." (41)

Why then, do people hang on to their irrational beliefs?

I propose that an explanation of this phenomenon may come from extending Edelman's Neural Darwinism theory to the cognitive realm. Edelman has proposed a detailed model of how neuronal group selection accounts for complex motor actions. (42) His theory draws on thinking about muscle control which was originated by the Soviet physiologist, N. A. Bernstein. In their review of Bernstein's approach, Kelso and Tuller analyze how the nervous system may be able to control the huge number of "degrees of freedom" involved in the many different muscles which combine to perform complex actions. (43)

Synergies and Gestures in Motor Actions

Kelso and Tuller point out that complicated muscle systems must be "selectively limited in their actions, not uniquely determined....Systems that do not perform functions |effectively~ are either too tightly constrained (e.g., rigid objects) or hardly constrained at all (e.g., an aggregate of grains of sand.)" (44)

In order to achieve this balance between "flexibility or freedom to undergo change, and limitations on its flexibility," muscle groups are combined into units called "functional synergies" (or synonymously, "muscle linkages" or "coordinative structures"). These functional synergies are "deployed as a unit in a motor task." (45)

Edelman proposes that functional synergies of muscles are generalized in "gestures," which consist of classes of coordinated motions that can approximate "a pattern that is adaptive in a phenotype." These gestures also exhibit "degeneracy." They communicate with neuronal groups which categorize "gestures and postures and the transitions between them."

"Selection carves out effective motions from the large set of postures and gestural components. The two sets of these otherwise disparate mechanical, and neural components, mix to create a gesture."(46)

"Cognitive Gestures"

I propose that a similar mechanism may help to account for RET's "insight number 2."

We consider humans to be unique among animals in that they probe their environment to come up with an emotional response, mainly by thinking about it, rather than by actually carrying out motor actions. Thus, they can make evaluations of the environment by manipulating their own concepts without remaining tied to the immediate time-frame. Recall that Edelman's definition of concepts is that they are structures by which the brain models and generalizes its own previous "global mapping" experiences.

Now any human could create a potentially huge number of evaluating concepts about a particular situation. To cope with this profusion, I propose that neuronal groups exist within human brains which can categorize evaluating concepts in the same way that motor gestures are categorized by neuronal groups in Edelman's motor model. These collections of evaluating concepts would then be deployed in unison as a "cognitive gesture" to probe different situations. Neuronal group selection could act on these "cognitive gestures" to choose an idiosyncratic repertoire of cognitive gestural probes for each individual. This process constitutes what my hypothesis terms "Cognitive Darwinism."

Thus, an individual would cognitively probe many different situations but would have to choose from the same repertoire of cognitive gestures which he had previously practiced. Of course, his repertoire of "cognitive gestures" could change throughout life, as does his repertoire of motor gestures. As seen above, if the brain uses certain neuronal groups habitually at the expense of others, the groups it uses will become internally strengthened. Therefore, an individual should be more likely to use previously his practiced "cognitive gestures," simply because their associated neuronal groups have already been strengthened by habit.

Thus, "Cognitive Darwinism" may help explain RET's somewhat paradoxical "Insight Number 2": that people re-indoctrinate themselves in the present with the same irrational beliefs which have defeated them in the past. The paradox disappears if an individual is forced to employ the same repertoire of "cognitive gestures" to probe the value of entirely different situations. I believe that this psychological model demonstrates a possible "mechanistic explanation" for RET's "insight #2," in a theoretically parsimonious fashion, without requiring any other hypothetical constructs -- such as hidden behavioristic rewards, unconscious directions, original flaws, etc.

In this way, human beings may reinforce their own habits of emotional evaluation, without necessarily experiencing any external reward (or punishment) for them. Like motor gestures, cognitive gestures would become strengthened simply by repetetive use, not only be hedonic reward. In Darwinian biology, survival itself is reward enough to encourage a particular selection.

Pharmacology and Thoughts: Beck's "Cognitive Continuum"

In 1983, Aaron Beck hypothesized a "cognitive continuum" covering such apparently different treatments of depression as cognitive therapy, behavior therapy, and psychoanalysis. (47) Beck suggested the likelihood that psychopharmacology might also eventually become incorporated into such a general cognitive model of depression, if biochemical findings could be integrated into a general psychological scheme. Beck pointed out that previous biochemical models had been too reductionist to include in an overall cognitive model.

Beck observed, provocatively, that his own research had found that "the initial change with drug therapy was in cognitions, not the affect of depression." (48) He then raised the question, "What are the cognitions being affected by drugs?...A better model for understanding psychopharmacology may be found." (49)

I believe that applying the TNGS model of consciousness to Ellis's "missing cognitive element" theory of depression may help to shed some light on Beck's important riddle.

The Causal Efficacy of Consciousness

In Edelman's "Extended TNGS," primary (pre-human) consciousness is presumed to give an adaptive advantage, and to have arisen in evolution because it enhanced the survival of "conscious" animals. Instead of having to respond to individual stimuli on an "ad hoc" basis, the animal with primary consciousness has a specific "skill," which enables it to create a meaningful "scene" out of his environment by correlating and evaluating all the more important stimuli which reach it. The animal selects the stimuli for itself to be aware of by continually matching external stimuli to its own internal value-category memory.

Now from RET theory's point of view, one particular advantage of consciousness is that it allows the conscious animal to construct an overall emotional assessment of its situation at any time. That is, an animal with primary consciousness can construct a meaningful mental "scene," by continually evaluating all its perceptions against its own internal value-category memory. Then, the animal can evaluatively assess its own conscious scene in order to select among different behavioral "modes," or anticipatory "tones," in which it can prepare to more effectively play out its various behavioral strategies. Let us call this assessment of the proper behavioral "mode" the animal's "emotional response." Thus, if the animal with primary consciousness assesses its current mental scene as "dangerous," it may take on an alert emotional "mode" in which it anticipates dangers, even though no specific danger may yet be apparent.

Now for the human with what Edelman calls "higher-order," or linguistic consciousness, I propose that the construction of an emotional response may proceed along these same lines, except that the human is also able to conceive of itself in the past and in the future, and thus can make a more sophisticated assessment of its predicament than can the animal with "primary consciousness" only. (50)

"Value-category Memory" in Depression

Based on the above assumptions, we can hypothesize that the "locus of control" by which psychopharmacologic agents affect Beck's "cognitive continuum" may reside in Edelman's value-category memory. Remember that the value-category memory contains perceptual categories which have already been evaluated according to the individual's own internal, homeostatic standards. If enough relevant categories in the value-category memory become evaluated as "depressogenic," and if the individual also perceives these categories as present in his environment, we hypothesize that he will assess his conscious "scene" as of a "depressing" nature, and will then act and feel "depressed."

Referring to Figure 1, we see that at any instant in time, the categories within the value-category memory consist of stored matchings between perceptual (or "world") categories, C(W), and self (or "internal") values, C(I). The foregoing argument implies that remedies which would serve to improve a particular "depressogenic" category within the value-category memory might act in either one of two ways -- they could change C(I), the internal self-evaluation; or they could change C(W), the external world-classification. It follows that persuasion, leading to what Beck calls changes in cognitive organization, might succeed by changing the C(W) (i.e., the "exteroceptive") component of the "value-category." Pharmacologic agents might act on the same value-category by changing its C(I) (or "interoceptive") component.

The above model may help explain one of the more mysterious aspects of RET theory: How can the evaluative, demanding cores of irrational beliefs generate what Ellis calls their derivatives -- "unrealistic, overgeneralized inferences and attributions"? (51) Ellis himself raised this question in his 1987 article on depression's "neglected cognitive element":

"Are the common modes of perceptual and conceptual distortions, such as personalizing and overgeneralizing, significantly increased when depressed individuals subscribe to absolutistic musts?" (52)

The above model implies that the evaluative "absolutistic musts" which form the "core" of Ellis's "irrational beliefs" do their damage by changing values in the value-category memory. Subsequently, the value-category memory itself becomes responsible for altering the perceptions and conceptions which rise to conscious awareness.(53) According to Edelman's model, those perceptions and conceptions then go on to affect the value-category memory in turn.

In Beck's example of the depressed student, her "catastrophic" evaluation of the possibility of not getting into college would affect her value-category memory. This errant value-category would then give unusual conscious salience to specious perceptions of actual failure which otherwise would never rise to conscious awareness.

If she were treated with a psychopharmacologic agent, the drug might change the "catastrophic" nature of her evaluation of not getting into college -- C(I) in the v-c memory. Likewise, she might become convinced through RET that her original evaluation of this outcome had been erroneous -- changing C(W) in the v-c memory. Either way, the new value-category would change her conscious awareness of perceptual and conceptual inferences and predictions.

Instruction and Selection in Personality Models

This article has focused on building bridges between specific theories in two different fields -- Gerald Edelman's TNGS model of brain function, and Albert Ellis's RET model of personality. However, this exercise in "interfield theory" may also point to some more general lessons about psychological theories.

Neils Jerne, the immunologist, wrote in 1967:

"The replacement of instructive by selective theories appears to be a general trend in the development of biology.... An analysis of this question leads to the conclusion that the terms instruction and selection can apply to descriptions of the same process at different levels. Each system that is capable of receiving a signal is subject to instruction by this signal. Thus at the level of an entire system, all such signals are instructive, whereas all instructive processes at some lower level imply selective mechanisms, through which products that were already present in the system prior to the arrival of the signal are selected and amplified. In learning, as in all processes resembling learning, a discussion of instruction versus selection serves only to determine the organizational level of the elements upon which selective mechanisms operate." (54)

In addition to the role of selection in species evolution and in the formation of antibodies in the immune system, Jerne cites selective mechanisms in the development of antibiotic-resistant strains of bacteria, changes in coloring within a species, and the role of messenger RNA and transfer RNA in synthesizing protein. A decade later, Gerald Edelman extended this same selective paradigm to the level of neurophysiological learning, as we have seen. It seems likely that the trend may eventually continue to a deeper analysis of personality models in psychology as well.

We may predict that any selective theory of personality might include the following elements:

1) A "population" consisting of variant individuals within isofunctional "repertoires." The members of this population could consist of "beliefs," "desires," "thoughts," "cognitions," "understandings," "conceptions," "behavioral dispositions," "drives," "goals," "assumptions," "fixed ideas," "personal constructs," "traits," "needs," or whatever. For the sake of clarity, I will continue to use Ellis's term "beliefs," although of course it is not interchangeable with the above alternative constructs.

2) The individual member "beliefs" within an isofunctional "repertoire" should exhibit "degeneracy" in their "recognition" of lifetime events. This would appear to mean that an individual must hold many different "beliefs" at the same time about a single event, some of which "beliefs" would even contradict each other in certain aspects.

3) There should be some "amplification" mechanism which enhances those "beliefs" which better "fit" experience and which allows the personality system of "beliefs" to "learn" from its life events and thoughts.

4) There should be internal communication which associates repertoires of "beliefs" with biological, homeostatic "values." These evolutionarily adaptive "values" should guide the overall process of selecting "beliefs."

REFERENCES

1. A. Ellis. How to Think Scientifically About Yourself, Other People, and Your Life Conditions," ETC., 47, No. 2: 135-144, 1990.

2. A. Ellis. How to Live With a Neurotic, at Home and at Work, Rev. Ed. New York: Institute for Rational Living, 1975.

A. Ellis and R. Harper. A New Guide to Rational Living. New York: Institute for Rational Living, 1975.

A. Ellis and W. Knaus. Overcoming Procrastination: Or How to Think and Act Rationally in Spite of Life's Inevitable Hassles. New York: Institute for Rational Living, 1977.

A. Ellis. Anger: How to Live With and Without It. Secaucus, NJ: Citadel, 1977.

3. G.M. Edelman. and V.B. Mountcastle. The Mindful Brain: Cortical Organization and the Group-Selective Theory of Higher Brain Function. Cambridge, MA: MIT Press, 1978.

G.M. Edelman. Neural Darwinism: The Theory of Neuronal Group Selection. New York: Basic Books, 1987.

G.M. Edelman. The Remembered Present: A Biological Theory of Consciousness. New York: Basic Books, 1989.

G.M. Edelman. Bright Air, Brilliant Fire: On the Matter of the Mind. New York: Basic Books, 1991.

4. W. Bechtel. Philosophy of Science: An Overview for Cognitive Science. Hillsdale, NJ: Lawrence Erlbaum, 1988.

L. Darden and N. Maull. Interfield Theories, Philosophy of Science, 43: 44-64, 1977.