From Complexity to Creativity -- Copyright Plenum Press, 1997

Back to " From Complexity to Creativity" Contents

Part IV. The Dynamics of Self and Creativity





    So far, we have discussed psychology a great deal, but mostly in the domains of cognition and perception. We have not had much to say about everyday psychology, the psychology of everyday human behavior. Now it is time to turn to these more nebulous but ultimately more intriguing aspects of the human mind. It is time to ask: Why do we humans act the way we do? What are the dynamics and structures underlying human personality?

    These are complicated questions, and psychology has given us many complicated answers. In this chapter I will provide a new and different answer -- one which, while relying on concepts from complex systems science, is notably simpler than many of the theories of classical clinical psychology. I will use the psynet model to formulate a novel personality theory called "subself dynamics."

    Subself dynamics is a radical and "experimental" theory, which is still in an early stage of development. However, it seems to have a great deal of potential for explaining complex personality phenomena -- phenomena that other approaches to personality dismiss as puzzling or paradoxical. It will be a crucial component of the synthetic theory of creativity to be proposed in the last chapter.

    Psychologists may be alarmed at the degree to which I appear to ignore previous theories of personality. But this is not done out of ignorance or arrogance. I am well aware of the ideas of the "classic" personality theorists -- Freud, Allport, Murray, Rogers, Kelley, Lewin, Bandura, and so forth -- and I have a great deal of respect for these thinkers. In some instances the system-theoretic approach may lead to the rediscovery of classical ideas. Thus, by failing to mention these personality theorists I do not mean to imply that their theories are wrong, but only that, except in a few cases, which will be duly noted, I have not found their work particularly useful in developing a complex systems based approach to personality. This does not mean that relationships do not exist between the present theory and past theories -- only that these relationships are not particularly direct, and thus are not essential to the exposition of the present ideas.

    The idea of the present approach is, instead of formulating theories based on pure intuition as these thinkers have done, to use complex systems science as a guide for the intuitive analysis of personality phenomena. Freudian ideas will pop up here and there; and the final section of this chapter briefly discusses the power of systems theory to unify traditional personalitytheories with cognitive science. But, by and large, I have been content to leave to others the task of elaborating the relationship between the complex systems ideas presented here and the ideas of past personality theorists.

    In the previous chapter, we discussed the notion of self, in the context of artificial intelligence. A self-system, we concluded, is absolutely necessary for adequate knowledge representation and learning. However, no attention was paid there to the actual structure of the self. Here we will turn to this issue in earnest. We will approach human personality and human nature from the perspective of the dissociated self. Every person's self system, we will argue, is naturally divided into a number of situation and state dependent "subselves." The interaction dynamics between these subselves is the crucial factor in a person's complex, internal evolving ecology. Thus, human experience is viewed as subself dynamics.


    Proust, when at age 16 he wrote of "the several gentlemen of whom I consist," was reporting a universal human experience. Each of us acts differently in different situations -- at home, at work, out drinking with friends. Proust's insight was that these different perception and behavior patterns, elicited by different situations, are not merely aspects of the same person, they are fundamentally different people -- different selves inhabiting the same body.

    This may seem a strange idea, but in the end it is almost obvious. For what is a "self," after all? A self is just an autopoietic process system, consisting of a dual network of procedures for mediating between mind and world. There is no reason whatsoever why a mind should not contain many such networks. Different situations require fundamentally different mental processes, perhaps even mutually contradictory mental processes -- it is foolish to expect that this diversity of aims should be accomplished by a unified system.

    Excessive separation of subselves is a problem. One does not want subselves with different names, entirely different identities, mutually exclusive memories. But one does want subselves with slightly different identities, and slightly different memories. One's memory of a situation when at work may be entirely different from one's memory of that situation when at home -- and for good reason, because memory is in large part a constructive process. One builds memories from the raw materials provided by the mind, in order to serve specific purposes. Different subselves will have different purposes and thus different memories.

    The concept of subselves has become popular in psychotherapy circles in recent years (see e.g. Rowan, 1991). There are techniques for "letting one's subpersonalities speak," for coming into contact with one's lover's subpersonalities. A woman might have a "scared little girl" subpersonality, a man might have a "neighborhood bully" subpersonality. A straight-laced societywoman might have a repressed "slut" subpersonality, denied expression since her teenage years. In this type of therapy one deals with subpersonalities on an entirely individual basis: each person must discover, with the therapist's help, what their subpersonalities actually are.

    The present theoretical approach is complementary to this type of psychotherapy. My aim in this and the following chapters will be to seek universal aspects of subself dynamics. These universal aspects are closely tied in with the more situation- specific aspects dealt with by psychotherapists. In order to thoroughly understand any particular individual, both approaches are necessary.

    In Epsteinian terms, it should be understood that a "subself" contains a self system, a world system, and a system of interrelations between them. Insofar as they bring to mind multiple personality disorder, the words "subself" and "subpersonality" may seem too strong. But if this association is put aside, one finds that, if they exaggerate the situation at all, it is only by a very small margin. For in truth, the word "personality" as it is generally understood would seem to be a perversion of the facts. "Personality" is too individualistic; it implies that the qualities of a person are fundamentally a property of that person alone, when in fact these qualities are in every respect social. They are formed through social interaction, and they are also defined by society, in the sense that, to a person from a substantially different society, they would be largely incomprehensible. So a given person's "personality" does not really exist except as part of a network of personalities -- just as a self/reality subsystem does not really exist except as part of a network of self/reality subsystems. And in certain cases two subselves residing in different bodies may be more closely related than two subselves residing in the same body; for instance, the roles that two lovers assume when with one another often have little to do with their other subpersonalities. This fact is exploited frequently in fiction.

    A self/reality subsystem is an autopoietic system; this constitutes its fundamental wholeness. And a person's whole self-reality system is also an autopoietic system, whose interproducing component parts are precisely its subsystems. In other words, a personality is an autopoietic system whose component parts are subpersonalities, or subselves.

    Some psychologists might reject system theory as a foundation for personality, arguing that the proper grounding for a psychological theory is behavioral observation, or neurophysiological data. But while it is certainly important for any theory to agree with the available empirical data, the biological and behavioral data regarding personality are clearly too scanty to allow for inductive theory construction. All serious theorists of the individual personality (and I do not include trait theorists in this category) have formed their theories by making generalizations from their personal experience. But such theories are always sketchy and full ofgaps, because they lack the rigorous logical structure provided by system theory.

    A priori, it would not be unreasonable to extend the subself idea even further, to obtain subsubpersonalities, and so forth. I propose, however, that in the case of human beings this is not necessary. In other words, I suggest that, just as there is a magic number 7+/-2 for human short term memory capacity, and a magic number 3+/-1 for levels of learning (see Bateson, 1980, or EM), there is a magic number 2+/-1 for personality system depth. The "-1" allows for the possibility that some severely retarded people may only be able to develop a single self/reality system without significant subdivisions; and the "+1" allows for the possiblity that in some exceptional cases, subpersonalities might develop subpersonalities. An example of the latter might be certain MPD patients, in which the strongest subpersonalities might eventually develop the flexibility and state-dependence of ordinary "top-level" self/reality systems. The relation of this magic number to the others is a subject for speculation.

Emergent Attractors

    Finally, having discussed subselves at length, I must now introduce a related idea that is very useful for the task of applying the psynet model to the study of personality. This is the concept of "emergent attractors" spanning different subselves and different human beings. Emergent attractors are one of the main reasons why the analysis of traits will never lead to an understanding of personality. The essence of personality lies, not in the particular traits of particular human beings or subselves, but in the behavioral routines that emerge when two or more subselves, or two or more people, interact.

    This fact will become abundantly clear below, when we discuss romantic love. Jack does not love Jill for her "qualities," at least not for any "qualities" that could be listed in words in a brief, comprehensible way. He loves her because of the way she makes him feel, or in other words, because of the pleasurable behavioral routines which emerge between the two of them when they interact. To determine why Jack and Jill lock into mutually pleasurable routines when Jack and Jane do not, is a very difficult problem. Jill and Jane may differ very little by outward "qualities," yet their behavioral routines on contact with Jack may differ greatly. Predicting the behavioral routines emerging between different people is no easier than predicting the behavior of a complex mathematical dynamical system -- something which, as chaos theory has taught us, is a very difficult task.

    And what holds for Jack and Jill, I will argue, also holds for the different sides of Jack's own personality, for Jack's different subselves. To explain how Jack's different subselves get along, or why they don't get along, it is not sufficient to list their traits, because these traits do not contain enough information to predict the behavior of the dynamical system formed by linking the various subselves together. The emergent attractors formed by two subselves may give rise to complex patterns that, in turn, spawn or modify another subself.


    The philosopher Martin Buber, in his famous book Ich und Du, distinguishes two fundamentally different ways of relating: the "I-It" and the "I-You" (sometimes called "I-Thou"). These concepts have not been picked up in any significant way by the psychology community. However, I believe that this has been a serious oversight. An understanding of these two modes of relationship is absolutely essential for the study of human personality. In this section I will present a computational psychology of human relationship corresponding to Buber's philosophical theory, and will apply this theory on the level of subselves within a single person, as well as on the interpersonal level.

    Then, in the following section, I will argue that a healthy mind, as a rule, consists of a population of subselves carrying out mutual I-You relationships with each other. This kind of subself community leads to robust, adaptive belief systems. On the other hand, a subself community containing a preponderance of I-It relationships will be characterized by self-sustaining belief systems of minimal adaptive value. This conclusion is what I call the Fundamental Principle of Personality Dynamics.     The Fundamental Principle connects subself dynamics with underlying thought dynamics, and thus bridges the gap between social and personality psychology, on the one hand, and cognitive psychology on the other. It could, in principle, have been arrived at without any of the apparatus of computational psychology and complexity science. In fact, however, it was not. These modes of thinking lead up to the Fundamental Principle in a remarkably natural way.

Buber on I-It and I-You

    An I-It relationship is a relationship between a self and a thing or a collection of things. An I-You relationship, on the other hand, is a relationship between a self and another self as a self. Two selves may relate in the I-It mode, or in the I-You mode: it is a question of whether the one self recognizes the other one as a genuine self, equal to itself in reality and integrity; or whether it merely interprets the other one as a stiff, inanimate portion of the external world.

    Buber describes the I-You relationship in rather mystical term:

    -- What, then, does one experience of the You?

    -- Nothing at all. For one does not experience it.

    -- What, then, does one know of the You?

    -- Only everything. For one no longer knows particulars.


        The basic word I-You can be spoken only with one's whole being. The concentration and fusion into a wholebeing can never be accomplished by me, can never be accomplished without me. I require a You to become; becoming I, I say You.


        The relation to the You is unmediated. Nothing conceptual intervenes between I and You, no prior knowledge and no imagination; and memory itself is changed as it plunges from particularity into wholeness.

I-You plunges the I into a timeless world of interdependence. I-It, on the other hand, is firmly grounded in the world of time:

    The I of the basic word I-It, the I that is not bodily confronted by a You but surrounded by a multitude of "contents," has only a past and no present.... He has nothing but objects; but objects consist in having been.

    According to Buber, there is no real possibility of perceiving everyone as a You all the time. Rather, real human relations are a process of oscillation between the two extremes:

    This, however, is the sublime melancholy of our lot that every You must become an It in our world. However exclusively present it may have been in the direct relationship -- as soon as the relationship has run its course or is permeated by means, the You becomes an object among objects.... The human being who but now was unique and devoid of qualities, not at hand but only present, not experiencable, only touchable, has again become a He or a She, an aggregate of qualities, a quantum with a shape. Now I can again abstract from him the color of his hair, of his speech, of his graciousness, but as long as I can do that he is my You no longer and not yet again.

In our present culture, however, we are closer to the I-It extreme than was the case in the past. Indeed, Buber views I-You as being a more primitive notion than I or You itself:

    [T]he primitive man speaks the basic word I-You in a natural, as it were still unformed manner, not yet having recognized himself as an I; but the basic word I-It is made possible only by this recognition, by the detachment of the I.

        The former word splits into I and You, but it did not originate as their aggregate; it antedates any I. The latter originated as an aggregate of I and it, it postdates the I.

    I-You relationships are not necessarily happier than I-It relationships. However, they are deeper. It is better,according to Buber, to genuinely hate another person, than to feel toward him only shallow surface sentiments. The former is a path leading to God; the latter leads only to nothingness. For, in perhaps his most radical contention, Buber equates the experience of an I-You relationship with the experience of God.

Modelling Other Selves

    One may ask what sorts of differences in cognitive processing underly the differences between these two ways of relating. What are the mechanisms underlying these two kinds of experiences? This is a rather prosaic question, one which perhaps jars oddly with Buber's flights of mystical enthusiasm, but it is an interesting one nonetheless.

    I claim that the answer to this question is quite simple. The mechanism underlying an I-It relationship is a recognition of the particular patterns displayed by another individual, another self. On the other hand, the mechanism underlying an I- You relationship is an implicit recognition of the overall emergent structure of the other. This overall emergent structure will, in general, be too complex and subtle to enter into consciousness as a whole. Thus it will be experienced as broad, featureless, abstract. But the mind can nevertheless experience it.

    The experience of the You does not depend on the details of the individual being experienced. But that does not contradict the equivalence posed here, between the You and the overall emergent structure of another self. For it is the key insight of complexity science that the high-level emergent structures observed in a system need not depend on the lower-level details of the system.

    It follows from this that perception as an It is a prerequisite for perception as a You. One must first recognize the lower-level patterns out of which the higher-level patterns emerge. And, furthermore, one obtains the concrete prediction that some individuals will be more easily recognized as You than others. This is indisputably the case. Although the experience of the You is independent of the "It"-qualities of the individual being experienced, the probability of entering into an I-You relationship with a given individual is sharply dependent on these qualities. It is necessary that one recognize a sufficiently wide and deep assortment of the individual's qualities, in order to be able to make the leap to the overall ordering structure.

    What, though, is the overall ordering pattern of a self? It is, I have said, a recursive dual network. A dual network which contains an image of itself. The patterns making up the dual network are the things that are recognized in an I-It encounter with someone's mind. An I-You encounter, on the other hand, is marked by the rhythm of the network as a whole passing through itself again and again. It passes through itself -- mirrors itself -- and thus becomes itself, getting back to whereit started (perhaps with minor modifications; but this is not something that can be detected on a high level of perceptual abstraction). To enter into an I-You relationship with someone else is to enter into their time-stream; their inner process of unfolding and self-creation. For it is this inner process of unfolding that makes them a You, a self, a whole person, rather than just a collection of mental and physical habits.

"You" Need Not Be Human

    From this analysis one might conclude that it is only possible to enter into an I-You relationship with another intelligence -- and not with an inanimate object, say, a flower, or a personal computer. But this would be mistaken. The external world also possesses an emergent, self-unfolding structure that can be grasped as a whole. Buber recognizes this in his passage on Goethe:

        How beautiful and legitimate the full I of Goethe sounds! It is the I of pure intercourse with nature. Nature yields to it and speaks ceaselessly with it; she reveals her mysteries to it and yet does not betray her mystery. It believes in her and says to the role: "So it is You" -- and at once shares the same actuality with the rose.

Nature reveals its mysteries to Goethe -- i.e., he paid it careful enough attention to get to know it as an It; as a dynamic, structured, changing system full of intricate details. And this effort was repaid by a glimpse at the whole, by a view of nature as You.

    This sort of experience, reported by Goethe, was commonplace in many past cultures. Communion with nature as a You was, in "primitive" cultures, understood as communion with God. This experience came almost automatically to individuals who lived in nature every day, and who were raised to relate to nature as a thinking being rather than as an inert physical "environment." This is the sad aspect of current political disputes about environmental issues. Businessmen are asked to respect nature on abstract moral grounds. But such respect can never be genuine, because it comes out of an I-It relationship with nature, rather than an I-You relationship. Having lost the I-You relationship with nature, we have, as a society, lost our fundamental motivation for preserving and nurturing the natural world.

    The patterns of nature as a whole are relatively similar to the patterns of the human mind. In principle, however, it is even possible to carry out I-You relationships with a truly inanimate object. Say, with a rock. Putting aside ultimate philosophical questions of the reality of the external world, one may at least say that a rock as perceived by a given I, is a collection of perceptual patterns. This collection of perceptual patterns is, under ordinary circumstances, an It. It is not alive; it does not enter into itself and create its own time-stream. However, there are states of mind -- mystical, "oceanic" states -- in which even such inert collections of patterns are perceived to "come to life." The fundamental self-creating dynamic of self and awareness is extended throughout the more rigid layers of the mind. In a state of mind such as this, inanimate objects may be perceived as You. Everything becomes a You. This state of mind is touched on by Buber in the third part of his book. It is the state of mind of the truly enchanted being: the seer, the saint, the Zen master. Buber rejects mysticism that is based on finding the self within. True insight, he said, is based on deep and lasting recognition of the world as You.


    Suppose one has a collection of selves, all relating to each other in the I-You manner. Each self thus recognizes the others as selves. What benefit is this to the collection as a whole? What good does it do a self to be recognized by the others as a self?

    This question has relevance both for personality and for social psychology. For it applies equally to the case where all the selves in question are part of a single mind, and the case where the selves are located in different physical bodies.

    From a common-sensical, human-relations point of view, the answer to our question is almost obvious. From a mathematical point of view, on the other hand, it is an extremely difficult question, one whose complete resolution lies far beyond the grasp of current analytical techniques. This situation is, unfortunately, typical of all the really interesting psychological questions. Nothing is guaranteed in psychological systems: there are exceptions to every would-be rule. One is always talking about probable outcomes in certain situations, where there is, however, no apparent way to quantify the actual probabilities involved, nor to place strict boundaries on the set of situations involved.

    The first idea needed to resolve the question is that the I-You relationship is beyond judgement. To enter into an I-You relationship with a self is to accept that self as a law unto itself, to understand the autopoietic network of relationships by which the different patterns making up that self relate to one another. If one understands and directly perceives this network, then one is bound to be tolerant of the particular patterns making up this other self, even if, taken individually, they might not seem agreeable. On the other hand, in the I-It relationship, the holistic integrity of the other is not immediately perceived, and thus the intuitive reaction must be to reject those aspects of the other's being that are not agreeable in themselves. Intellectually, the I in an I-It relationship might reason that the other self has some reason for doing what it is doing. But this will not be felt, which is the important thing.

    So, in a network of I-You relationships, there will be atendency of selves not to interfere with each other's actions on a microscopic basis. There will tend to be a certain respect -- based on a mutual understanding of the contextuality of actions, on a direct perception of the roles of individual patterns in autopoietic self-and reality-systems. What effect does this respect, this leeway, have on mental functioning?

    Clearly, an atmosphere of tolerance and respect will lead to a decreased need for defensiveness on the part of individual thought-systems. If each individual thought-system proposed by a certain self is immediately going to be shot down by other selves, then the self in question will have to be very careful to protect its thought-systems, to make them self-sufficient and resistant to attack. On the other hand, if there is an atmosphere of relative leniency and tolerance, then resilience is no longer so important, and other aspects of thought-systems may be emphasized.

    But recall that, according to the psynet model, there are two ways for thought systems to survive in the mental network: by autopoiesis or by adaptation. Each thought system must have a little of both features. As a very general rule, it is usually the most dogmatic, entrenched and unproductive thought systems that survive primarily by autopoiesis. And, on the other hand, it is the shallowest and least profoundly insightful belief systems that survive primarily by adaptation, by constant self- adjustment to the fluctuations of the environment. The best belief systems are the ones which use a bit of autopoiesis, to keep themselves going when the environment fluctuates against them, and a bit of adaptation, to ensure their relevance to external situations, their contextual "fitness" and productivity.

    This line of reasoning leads up to the conclusion that a system of selves characterized by I-It relationships will tend to produce overly dogmatic, entrenched thought systems. A system of selves characterized by I-You relationships will, on the other hand, tend to produce more well-balanced thought systems, which are adaptively effective as well as adequately self-preserving. In scientific terms, a statistical correlation is posited: between the adaptivity and productivity of thought systems, and the I-You nature of inter-subself relationships. This correlation is what I call the "Fundamental Principle of Personality Dynamics"

    Obviously, I-It interself dynamics is not the only explanation for dogmatic, inflexible thought. Inflexible thought systems can arise for any number of reasons. For example, in Chaotic Logic I give a detailed example of an inflexible thought system, the conspiracy theory of a paranoid individual. Everything said to the paranoid person becomes "part of the conspiracy." The basic principles of the system never adapt, and the new explanations which these principles generate bear little relation to the situations they allegedly respond to. The system generates little emergent pattern in conjunction with its environment -- it is "un-fit." However, this system did not arise primarily as a consequence of subself dynamics, but rather as a consequence of unvaryingly I-It interactions with the outside world. The paranoid individual, more than anyone else perhaps, does not accept anyone else as a You. Their suspicionholds them back from this intimacy. Their world is an object to be manipulated and analyzed. They have no respect for the emergent wholeness of the world, or the others in it, and it is precisely for this reason that they react so oddly to the specific patterns that confront them in their lives.

    There is a possible connection with developmental psychology, which should be noted. Apparently, we learn to relate by observing our parents, siblings and others relate. Thus a person who was presented with predominately I-It relationships during their youth, would be expected, on average, to develop primarily I-It relationships internally, among their subselves. This leads to the hypothesis that individuals whose early childhoods are deficient in I-You relating will tend to display inflexible belief systems throughout their lives. Their thought systems will be inflexible in regard to the external world; and particular components of their personality will be inflexible in their relations to each other. This prediction is not quite a necessary consequence of the Fundamental Principle: one can imagine mechanisms by which early I-It relationships might lead to internal I-You relationships. However, it is a "likely" consequence of the Fundamental Principle and as such would seem worthy of empirical exploration.

    Finally, although I have been mainly thinking of communities of selves within a single mind/body, it should be reiterated that the same conclusions apply to groups of human beings. For instance, in a marriage, which is a group of two, the Fundamental Principle would suggest that the degree of inflexibility of the partners' belief systems should be inversely correlated with the frequency of I-You encounters between the partners. One of the remarkable things about the subself perspective is the way it erodes the border between social psychology and personality psychology. Both are seen to follow the same high-level dynamic patterns.


    Before moving on to apply and extend these ideas, it is worth briefly pausing to ask: how does the system-theoretic view of personality outlined in this chapter and the last relate with previous personality theories? This is a complex question, far too large to be substantially explored in this brief treatment; but a few simple observations may be made.

    The key point I would like to make here is that systems theory bridges the gap between modern cognitive science, which provides incisive understanding of particular mental processes, and intuitive personality theorizing as represented e.g. by Freudian psychology, which provides deep but vague understanding of high-level personality phenomena. The Fundamental Principle described in this chapter is an example of this kind of "gap- bridging" at which systems theory is so adept.

Freudian Theory and Systems Theory

    Let us begin from the top, with the "high-level" theories. Many of the great psychologists have had theories of personality: Freud, Jung, Allport, Murray, Kelly, and so forth. These theories are extremely difficult to assess. One may observe that psychiatrists, in controlled tests, have proved no better than anyone else at the task of predicting behavior. But on the other hand, one may quite correctly argue that, unlike, say, trait theories of personality, these non-quantitative theories of personality were never intended for prediction. They were intended to provide understanding of specific cases. And if chaos theory has taught us anything, it is that prediction and understanding are very different things.

    My quarrel with these personality theories is not their poor prediction ability, but something much more meaningful, their lack of subtlety. These theories simply do not do justice to the complexity of personality. The system-theoretic treatment given here does not fully do justice to this complexity either, but it is a step in the right direction. What past personality theories have done is to replace complexity with simple stock ideas -- ideas pulled "out of a hat," with no scientific or mathematical foundation, for the sole purpose of making personality theory simpler. This point is best made clear by a concrete example. Let us briefly discuss Freudian theory -- which, for all its obvious faults, has come closer to an understanding of the complexity of personality than any of its competitors.

    The weakness of Freudian psychology is well symbolized by the concept of psychic energy. This is a metaphorical idea, pulled out of a hat and inserted at precisely the point in Freudian theory where a scientific idea would have been most valuable. And what were the conclusions to which Freud was led by this idea? Chief among them was the idea that it is pointless to treat the symptoms of a psychologically troubled person, because training a person not to express a certain symptom will simply cause the psychic energy underlying that symptom to be redirected elsewhere. However, over the past 40 years many psychoanalysts and other psychotherapists have come to the conclusion that this Freudian hypothesis is generally not true. Very often, symptoms are treated, and no other symptoms arise to take their place. Now, how is this observation to be integrated into Freudian theory? If only some symptoms are worth treating, is there some way to determine which ones? Freudian theory is not equipped to answer these questions, because it is not grounded in any scientific or mathematical discipline; it is a "castle built in the air."

    Freud was trying to formulate laws to explain the behavior of a complex system, the personality, but he had no concept of the brain/mind as a complex system to fall back on, and one can see this shortcoming in the details of his personality theories. The whole idea of a symptom as a consequence of an underlying problem bespeaks a failure to appreciate the circularity of complex system dynamics. The symptoms maintain and produce the underlying problem, and the underlying problem maintains and produces the symptoms: the whole "complex" is an autopoietic,self-organizing system. In some situations, removing one element from an autopoietic system will destroy the whole system; in other situations it will not. This is not a mysterious matter, it is a question of basin size and shape, of stability with respect to perturbations. One can imagine that the diagnostic manuals of the future will contain, along with their lists of different disorders, explicit information about the size and shape of the basin of the autopoietic subsystem characterizing each disorder. In all but a few instances, Freud overlooked the very possibility of this kind of complexity.

    It is important to make a distinction here. Freud, like the other great personality psychologists who followed him, had a unique and profound understanding of human personality. In their analyses of specific cases, in the organization of their descriptions, in their detailed decisions as to what to emphasize and what to omit, they revealed the depth of their personal understanding, a depth comparable to that achieved by Dostoevsky, Nietszshe, Proust, and other great artists and philosophers concerned with human nature. But the great psychologists were much more accomplished at understanding personality than at theorizing about it. Their intuitive power far exceeded their power to codify their intuitions. And this does not necessarily reflect any innate inabilities on their part; they lacked appropriate tools to aid them in the process of codification. Freud, for one, was acutely aware of this: he once formulated a neurological theory of personality, only to set it aside instead of publishing it, grimly but realistically aware that neuroscience was not yet sufficiently advanced to allow him to do what he wanted to do. In place of this neurological theory, he promoted a theory based on "psychic energy," an idea with absolutely no scientific foundation.

    These strengths and weaknesses of psychoanalytic theory will become quite obvious below, as we consider the concrete phenomena of romantic love. On the one hand, confusing, jargony concepts such as the Oedipus complex are invoked excessively often, in contexts where much less esoteric explanations clearly suffice. But on the other hand, one also finds deep and penetrating analyses of the varieties of human experience. Even if one rejects the Oedipal aspects of Freud's analysis of masochism, one still often finds that he has captured the basic structure of the phenomenon.

    So, in conclusion, it is interesting to contrast the system- theoretic approach with the Freudian approach. Freud was, above all, concerned with modeling the mind. Modern science has given us new tools with which to carry out this modeling, but the basic goal is still the same: to understand the hidden structures and processes which cause us to feel and act the ways we do. The biggest difference is that Freud built his models from concepts like id, ego and super-ego, which were fabricated especially for the purpose of modeling mind. Thus one finds that many of Freud's models are clear and sensible in their abstract structure but unclear or unreasonable in their details. In the system- theoretic approach, on the other hand, the details are not left "dangling" into a sea of jargon and ad hoc concepts, but arerather grounded in concrete computational and complex systems models. Even if, as a consequence of lack of data, a system- theoretic model of some personality phenomenon must be constructed on intuitive rather than deductive grounds, at least the concepts used as building blocks for the model will be scientifically meaningful.

Systems Theory and Cognitive Science

    Now let us turn to the "low-level" side of things. One might hope that a new, scientifically-based theory of personality would emerge from cognitive science -- an interdisciplinary endeavor whose influence is pervading more and more aspects of psychological theory, so much that Mandler (1985) was moved to pronounce that "modern psychology is cognitive psychology." But this has not yet occurred, and, for reasons that should be clear from the previous chapters, I suggest that it is unlikely to occur in the near future.

    The basic metaphors of cognitive science, I suggest, are just not up to the task. Their focus is excessively micro-level; very little attention is paid to crucial questions of high-level structure and dynamics. If mainstream cognitive science were to turn to something like the psynet model, or Kampis's component systems or Varela's autopoietic systems, then it might have a reasonable chance of dealing with high-level psychological issues. But as it stands now, the leaders in the field of cognitive science have no such inclinations, and so the prospects are not good.

    Cognitive science is not so much a unified science as an amalgam of three different subdisciplines: artificial intelligence, cognitive neuroscience and cognitive psychology. And a great deal of work in cognitive science is still pervaded by the logicist paradigm of rule-based artificial intelligence. To oversimplify just a little, the metaphor is, "mental processes as simple algorithms, which have short programs in our high-level programming languages, and which are largely understandable in isolation from each other."

    Carried over from artificial intelligence to cognitive psychology, the logicist approach works well with certain issues -- e.g. factual memory, short-term memory, and perceptual processing -- but is very poorly suited for personality. For personality is deeply dependent on self-organization and interdependence, the very things which traditional cognitive science fails to capture. Cognitive psychologists have made positive contributions to personality theory: for instance, they has shown that we tend to use ourselves as a prototype for understanding others, and that memory depends on personality, e.g. in that certain types will tend to remember successful experiences, while others will tend to remember unsuccessful ones. But while this sort of study is enough to help one adjust the "parameters" of a personality theory, it is not, in itself, a strong enough foundation on which to build a theory of human nature.

    The cognitive science community has begun to growdisenchanted with the logicist metaphor, and has embarked upon a fair amount of research embodying a "connectionist," rather than logicist, point of view. Connectionism, unlike logicism, is strong on self-organization and interdependence. The basic connectionist metaphor, the "neural network," is, like the mind, a self-organizing dynamical system. However, the central weakness of the connectionist approach is precisely its inability to address issues of large-scale cognitive structure. In technical language, neural network models do not "scale up," and the reason these models do not scale up is that they do not incorporate any of the intermediate-scale or large-scale structures of the brain/mind. A neural network psychology which focused on these structures would bear little resemblance to the neural network psychology of today. Thus, although system- theoretic psychology is related to connectionist cognitive science, the relationship is not so close as it might seem. The ideas of autopoietic system theory are unlikely to be "happened upon" through experimentation with toy neural networks; they are external to the conceptual repertoire of neural network psychology.

    So we find that mainstream cognitive science has precisely the opposite flaw of Freudian theory. Freud's penetrating and sophisticated explanations of high-level structure and dynamics had no solid grounding in lower-level dynamics. On the other hand, cognitive science, with its broad-based and incisive understanding of details, has no way of reaching up toward higher-level structures. The connection between the two, I propose, is system theory. System theory connects the micro level with the macro level, the behavior of the relatively simple components with the behavior of the structured whole. It is thus ideally suited to serve as a bridge between modern cognitive science and old-style intuitive personality theorizing.

The Fundamental Principle as a Bridge

    System theory, in general, has the ability to bridge intuitive personality theory with modern cognitive psychological theory. The Fundamental Principle of Personality Dynamics proposed here is, I would claim, a manifestation of this ability.     The concept of I-It versus I-You relationship is an high- level philosophical idea which is at home among intuitive theories of personality. Likewise, the idea of subpersonalities comes from clinical rather than experimental psychology. These ideas tie in closely with everyday human experience.

    On the other hand, the study of thought systems and their rationality, adaptiveness and coherence lies squarely within cognitive psychology. Because of the practical difficulties of gathering adequate time series data, questions of the dynamical adaptivity and resilience of thought systems have not been studied extensively to date. However, these questions are natural extensions of the cognitive science research programme. From an abstract mathematical point of view, they have been studied extensively within the artificial intelligence community.

    Thus, the Fundamental Principle connects the personality level and the cognitive level in a way that has never been donein either personality psychology or cognitive psychology alone. By treating the mind as a whole system, with complex dynamical interactions on several different levels, one comes to view the different aspects of mind studied by different subdisciplines of the discipline of psychology as parts of a unified whole. Of course, on a general level, this unified view of mind is absolutely commonsensical -- the separation into different aspects and levels is purely a consequence of the sociology of the discipline of psychology. But the key thing is to show that the unified view can lead to definite insights. This is what system theory accomplishes.