Psychoanalysis and Contemporary Thought 7, 1984, 147-194.

Freudian Psychoanalysis and Information Processing:

Notes on a Future Synthesis

ROBERT DE BEAUGRANDE, Ph.D.

THE INFORMATION-PROCESSING APPROACH

The information-processing approach has offered new perspectives on many issues in human psychology, such as language, memory, and comprehension (Lindsay and Norman, 1972; Anderson and Bower, 1973; Loftus and Loftus, 1976; Rumelhart, 1977). The basic analogy between information processing and human functions should be construed on the level of metapsychology, whose domain, according to Pribram and Gill (1976), should now be a "biological cognitive control theory" (p. 14). As noted by Rosenblatt and Thickstun (1977), Freud himself admitted the fundamental possibility of reinterpreting the basic corpus of psychoanalytic observations by means of a new theoretical superstructure. Viewing the human as a complex system traversing a pattern of information states helps us to formulate complex problems and issues in tractable and operational ways, as demonstrated by research on visual perception (Sperling, 1960; Eriksen and Collins, 1967; Horn, 1975; Kosslyn, 1975; Minsky, 1975; Waltz, 1975).

However, the information-processing approach by itself is extremely general and must be specified with careful attention to the concerns of each domain that it is applied to. Some of its fundamental assumptions, which will be elaborated further on, are as follows:

1. Information is a designation for any configuration in a system as compared to the alternative configurations of that system, and processing is a mode of acting upon-building, searching, matching, changing-such configurations.

2. The performance of a processing system can be viewed as problem-solving, defined as the search and selection of pathways from one state to another, e.g., moving from unknown to known' unfamiliar to familiar, plan to goal, and so forth.

3. Psychological entities, such as thoughts, meanings, and emotions, are processing events, where event is defined as a change of state. An object is one outcome of an event that constitutes (identifies, classifies, etc.) an entity. A theory for any given psychological entity should represent the dynamics of the relevant processing events as exactly as possible. Black-box models that discount all operations between input and output are ultimately unproductive because, from a human standpoint, input and output must also be constituted as "events" and hence must reflect their dynamic conditions.

4. Some processes are general; others are specialized. The interaction between general and specialized processes can create and control elaborated complexes of behavior and cognition for a wide variety of occasions.

5. Processes are partly physiological and partly symbolic. On the one hand, a process involves the activation of nonrandom neural patterns with a characteristic result; on the other hand, a process represents something that is related in a symbolic way to those patterns. The physiological and the symbolic' aspects are present in all modes of processing, although their mutual interactions and dominances can fluctuate widely from case to case.

6. A human process can vary; it need not run in precisely the same way for all people, nor even for the same person on different occasions. Variations can easily arise from divergences in motivation, time, automaticity, resource load, -and so on. However, this is not to say that the process is uncontrolled or random. Rather, a process is properly defined as a community of related physiological and symbolic events whose preconditions and results are, for human purposes, classified together. The same results might be attained with different means on another occasion.

7. Human actions should be specified in terms of their intentionality; at some level of processing, the human intends to relate the physiological with the symbolic and to group together associated actions as a single process. An account of what humans do includes their beliefs about what they're doing. However, this intentionality may not be open to introspection or observation. In particular, automatic processes are constructed and refined precisely to operate without attentional supervision, and thus to conserve resources. The mind sets a threshold at which these processes are triggered by appropriate conditions.

8. Reality is a human model that provides orientation and explanation in the face of everyday experience. Social order results from a general (though seldom universal) consensus about the validity of a particular reality-model and of the human actions that enact, maintain, or interpret it.

9. Potentially at least, any process can evolve, given appropriate conditions or intentions. However, not every such evolution is equally successful or persistent. Learning occurs when the human has undergone an evolution that modifies or expands his or her repertory of reliable operations. Ideally, learning could he optimized if we could represent and guide the evolution of the instrumental processes, about which we still know far too little.

10. Errors and dysfunctions occur when there is a serious or persistent problem in processing operations. A certain error margin is a natural side-product of the variety and complexity of human processing. Yet, severe limitations or disabilities can accumulate when operations are non-strategically organized. Learning disabilities may be less a predetermined product of genetic potential than an accidental product of unguided, disadvantageous process evolution.

The value of this approach depends on whether its models of human cognition and behavior support orientation and explanation in the exigencies of research and practice. If they do, they (concurrent with other models) meet the criteria of “reality" as defined above. Some implications and consequences of building such models for psychoanalysis are outlined below.

EMPIRICAL PSYCHOLOGY VERSUS PSYCHOANALYSIS

There have been several disagreements between empirical science and psychoanalysis in the course of their history. One of the major disputes has been that of physiological versus symbolic factors. During his long and busy career, Sigmund Freud combined physiological issues with symbolic ones in a powerful and original way. In his later years, he favored a purely "psychological" approach to psychoanalysis, one that would not rely on biological and neurological concepts. But, as Pribram and Gill (1976) observed, this separation is neither feasible nor desirable from the standpoint of contemporary science; nor was it made in Freud's (1895) original "Project for a Scientific Psychology." To exclude the biological from the psychological is to re-create the old Cartesian mind-body dualism and to dangerously fragment our model of the human being. Moreover, the unconscious is a domain that can have no place in Cartesian dualism (von Bertalanffy, 1964).

Freud certainly countermanded the most basic concepts of scientific positivism when he asserted that the physiological was under the control of the psychological. This new outlook encouraged an undue neglect of biological contexts. The physiological aspect was consequently displaced by metaphorical constructs, such as the "anthropomorphic" and "hydrodynamic" frameworks attacked by Peterfreund and Schwartz (1971). The question is not whether physiology and biology will be included in the foundations of psychoanalysis, but only what implicit or explicit frameworks will he used. If physiology is ignored or discounted, it becomes necessary to invent elaborate substitutes that cloud our view of the issues, such as the problematic and recently criticized construct of "psychic energy" (Rosenblatt and Thickstun, 1970; Peterfreund and Schwartz, 1971).

A negative consequence of one-sided "psychological" viewpoints is the inability to base a consensus on physiological factors. Hence, there has been almost unlimited diversity of terms and models in psychotherapy (Reisman, 1971; Brown, 1976) as well as a corresponding tendency among psychoanalysts to adopt a highly individualistic approach to their work (Ford and Urban, 1963). This has meant that there are no widely accepted standards for constructing theories or for turning them into therapeutic methods.

An equally negative consequence is that Freud's heritage is not integrated with the central concerns of empirical psychology in America. For several decades, behavioral psychology overemphasized the physiological at the expense of the symbolic, at least as radically as the Freudians had done the opposite. The need for a metapsychology of higher-order abstractions for making and interpreting observations was denied for a long time in hopes that simplistic measurements and tight controls would ensure the validity of evidence. To some degree, this stance injected into scientific inquiry itself the mind-body dualism by splitting the scientist into the dichotomy of subject and object. Scientists failed to acknowledge their own symbolic response to the physiological data they were gathering. Only recently has this position been recognized as untenable, especially for psychology (Rosenblatt and Thickstun, 1977).

The dispute over physiological versus symbolic factors is only the first point of contention between empirical psychology and psychoanalysis in America. At least four others should he noted. The second is the dispute over statistical generality versus the individual patient. Whereas experimental findings are customarily stated in terms of what most subjects have in common. Freud and his successors recognized a strong personal component in the subject's tendency to assign symbolic functions, especially in dreams. On the one side, statistical significance is expected to protect empirical science from accidental or idiosyncratic evidence. On the other side, psychoanalysis deals with, and is responsible for, the individual patient, no matter how unique the causalities may be in each case. In fact, many disturbances arise precisely from improbable, idiosyncratic processing that impedes functional interaction with socially accepted reality. The third dispute is closely tied to the second, namely, quantitative versus qualitative discovery. American "science" heavily favors numerical statements, no doubt because numbers seem relatively "objective" across a whole community of sciences. The reliance of empirical psychology on statistical methods conflicted with the need of psychoanalysis to describe syndromes in terms of features and tendencies, often supplied by the patient's own introspection. However, as Rosenblatt and Thickstun (1977) point out, qualitative, introspective data can be quantified with such methods as rating scales; and "premature and misguided attempts at mathematization of behavioral data, in emulation of the physical sciences, have on occasion obscured and stultified the development of original concepts in psychology" (p. 12).

A fourth dispute concerns linear versus multiple causality. Again inspired by the physical sciences, empirical psychology has inherited the Newtonian notion of a simple, one-way cause/effect chain; but psychoanalysis has to explore circular and multiple causalities (yon Bertalanffy, 1968). In the classic experimental design, each stimulus event is assumed to elicit a response event in a deterministic, lawlike way. But in psychotherapy, a symptom, such as anxiety, can arise from several causes in interaction and may persist even when such causes are removed, because others have taken over the same function (see Bowlby, 1969). A causality that explains experimentally controlled responses has different concerns from a causality that guides therapy.

A fifth dispute, and one closely related to the issue of causality, is the issue of explicit versus concealed motivation. Behavioral psychology devises special environments in which the subject's actions are controlled by different rewards and punishments. Supplying clear motives is an explicit part of routine experimental design, and much less attention is paid to the subject's symbolic model of everyday reality. By contrast, psychoanalysis assumes that behavioral patterns can arise from elaborately hidden motives which the disturbed subject wants to ignore or repress. What counts as a reward or a punishment for the individual subject might seem fully inconsistent with explicit, general values and goals in society.

Although these disputes are substantive and complex, the isolation and fragmentation of psychoanalysis is regrettable. An important branch of psychological application is cut off from the mainstream of controlled research on perception, comprehension, and memory-the basic processing involved in dealing with human experience and deciding what it means. Future progress crucially depends on overcoming this separation by means of a cumulative general theory (von Bertalanffy, 1964).. In this paper, I shall try to outline one prospective framework. for such a theory and to suggest how some of the Freudian heritage may be restated in terms of the information-processing approach. Obviously, at this stage such an attempt can he no more than a preliminary proposal. But, in view of the pressing need for more concerted interaction between psychoanalysis and empirical psychology, the effort may prove worthwhile.

PHYSIOLOGICAL FOUNDATIONS

It now seems reasonably certain that the physiological basis for thinking, remembering, reasoning, acting, and so on, is electrochemical impulses. In the past it was fashionable to view the mind as a static filing cabinet with documents being entered, removed, and carried about, for example, in the "pigeonhole" view of memory criticized by Kintsch (f977). It now seems more appropriate to regard the mind as a richly structured apparatus whose "contents" are patterns of excitation and transmission. For instance, recalling information in memory requires a change in the activation level of a stored pattern (Morton, 1970), rather than a messenger taking a paper from one pigeonhole and placing it into another. It is fundamental to the nature of electrical impulses that they are ongoing events, not fixed objects. It follows that the human brain and body can be substantially reconfigured by the transmission of neural impulses.

The chemical aspect of human processes is another important dimension. Chemical processes typically change the character of the participating substances. Some of the changes are irreversible; others trigger further changes in a chain reaction. The point is that, potentially at least, chemical impulses are able to fundamentally alter the conditions under which they take place and thus create new conditions for other impulses. To demonstrate that the basic physiological conditions are relevant for understanding the character of human processing, we would have to show that those conditions can account for significant properties of human cognition and behavior. This demonstration is not too difficult. Let us consider how some basic differences between a simple computer and the human brain arise from corresponding differences in their physiological construction.

The simple computer is built of stable conductors and insulators. Operations are mainly executed via the opposition between the open and closed circuits. Each impulse reconfigures a pattern by closing open circuits and opening closed ones. In the human brain, in contrast, an impulse can reconfigure a pattern in a variety of ways that are sensitive to ongoing conditions. One of these conditions is the state of other patterns in the same system and their strength of association with the target pattern. The changes effected by electrochemical impulses during learning, thinking, etc., are vastly complex with respect to the number and nature of states that can be reached. Hence, we should predict that this major physiological difference will be reflected in the operational properties of the simple computer versus those of the brain.

Such is indeed the case in the well-documented trade-off of storage versus search (Loftus and Loftus, 1976). The human mind has vast storage but rather unreliable search capacities, whereas the simple computer has comparatively limited storage but reliable search capacities. Presumably, the results of an impulse being transmitted through the system are relevant to this difference. The computer configures its circuits and leaves them fixed in that pattern and location. (This is as true of adjunct computer storage devices, such as disks and tapes, as it is of on-line memory.) Thus, search can go at once to the right place and get back exactly what was stored. In exchange, though, these fixed and rigidly maintained configurations and access pathways take up a great deal of space in the computer's memory and have to be updated by special commands. The brain, in contrast, configures itself in patterns that may well evolve or update their Vocational properties (loss, reduction, conversion, etc.) as a by-product of subsequent impulses. Hence, search can't count on taking the same pathway again or getting back exactly what was stored before; as a result, search is less reliable. In exchange, -though, these flexible, interactive configurations allow for efficient use of storage space and thus for the staggering memory capacity in humans. Patterns get compacted, simplified, merged, enriched, reused, and redesigned, in such a way that traces can be retained and integrated from enormous quantities of material.

The dynamic construction of the brain was not given due consideration in early theories of learning. In the wake of Darwinism, most such theories incorporated the mechanistic assumption that the brain's capacities were genetically fixed, like those of a hard-wired circuit-board (Galton, 1883; Spearman, 1904). This view provided a handy alibi for the failure of a person-or in Jensen's (1972) "theory," of a whole race-to succeed in school: poor "hardware" in the head. The "fixed mind" became a banner notion for such aberrant projects as eugenics, which proposed to sterilize people with inferior intelligence (Kamin, 1974).

Today, theorists of highly diverse persuasions concur that the potential of the mind for evolution is much greater than is normally attained, and that it is only to a small extent fixed in advance by physiological limits (Combs, 1952; Otto, 1968; Ouspensky, 1968; Maslow, 1971; Marks, 1982). As John McVicker Hunt (1970) concludes from the available evidence, "if the manner in which encounters with the environment foster the development of intellectual interest and capacity were more fully understood, it might be possible to increase the average level of intelligence within the population substantially" (p. 346).

The same lesson can be drawn from mental dysfunctions. Although they can be inherited, they can also result from evolutions during any stage of the person's life. The capacity of electrochemical impulses to alter their own conditions and participating substances could account for the variety and uncertainty of mental evolution. At some point, unfavorable conditions are created for certain processes or classes of processes, and a disability occurs. If not corrected, this effect could be intensified and propagated until gross behavioral disturbances, such as hysteria, are manifested. The fact that such disabilities can be corrected, at least part of the time, is natural if we assume they have originated in a vastly flexible and adaptive system.

It should follow that physiological countermeasures can be devised for psychological disturbances. However, such cures cannot be the kind used by early clinical psychotherapy, whose attitude was like that of a child shaking and banging a damaged toy in hopes that it will fix itself. Electric shocks and other gross impulses were administered, even though clinicians lacked a detailed theory of how or why this might correct a mental disturbance. Clearly, the mind got thoroughly shaken up; but there was no plausible theory to relate these gross physiological interventions with the complex and subtle symbolic causalities of the disturbance. At most, the patient learned to repress or disguise the symptoms in order to escape further persecution.

The lack of knowledge about the symbolic aspect was also a factor in the heavy-handed tinkering of brain operations. The dysfunctional brain location was cut out of, or disconnected from, the organism because nobody understood how to reconfigure the disturbed processes. The patients were "cured" by trading the disturbance for dismemberment. The symptoms disappeared, often along with many other brain functions. There was no way to make the operation specific and controlled enough to provide a cure with the least possible impact on the whole system.

SYMBOLIC FOUNDATIONS

A symbolic relation exists wherever human processes elect to define or treat one thing in terms of another. As Langer (1951) points out, even the basic "perception of form is abstraction" and therefore results from "the process of symbolization" (p. 59). Some symbolic acts are very simple, as when a physical sensation is interpreted as "heat" or "cold." Others are quite elaborated and complex, as when a configuration of experience is interpreted as "personal failure." Undoubtedly, the strongly developed symbolic function contributes to the impressive ability of human processing to conserve resources and storage by grouping comparable things together and interpreting new experiences according to high-powered representations of types. But in exchange, the rich and variegated possibilities for constructing symbolic relations make processing less reliable. Depending on what representation is needed for a given item, the item may be processed in many ways. The mind has to declare its current specifications and then to sort out a large quantity of items that match those specifications in diverse symbolic modes. Understandably enough, this diversity has made experimental psychologists reluctant to study the more complex symbolic action domains.

Folk-wisdom maintains that even simple computers are far less subject to error than humans. The notion of "error" is a very literal one here: failure to attain a precisely specified result from an equally precise operation. In a more powerful sense, the simple computer is, on the contrary, much more prone to major errors than the skilled human whenever heuristic reasoning is needed to identify, define, and structure a new problem. The computer's physical construction, which enables such reliable search, also severely limits its symbolic potential. Most programs can't recognize analogies, fit old ideas to new cases, apply common sense, jump to conclusions, correct odd deviations from the norm, and so on (Collins and Quillian, 1972; Collins, 1978). For most computers, the symbolic function demands an explicit statement assigning representations to events (or vice-versa). Some new approaches to programming have tried to offset these limitations by stating more abstract patterns that input should match (Pavlidis, 1977; Waterman and Hayes-Roth, 1978) and by taking everyday world-knowledge into account (Winograd, 1972; Minsky, 1975; Bobrow and Winograd, 1977; Fahlman, 1979; Schank and Riesbeck, 1981). However, as we'd expect in view of the search versus storage trade-off, these new programs tend to consume large amounts of memory and to need considerable respecification for each new domain of knowledge. Evidently, the human ability to reason about vaguely specified or unstable domains allows a complexity and subtlety that reduces major errors, even though it increases minor (literal) ones.

Nonetheless, human errors and breakdowns do occur for many reasons. The syndromes studied by Freud and his successors involve physiological disturbances with highly symbolic correlates. The "normal" causality, where an acute symptom arises from an organic lesion, is replaced by a hidden symbolic causality; when the latter is uncovered during therapy in a way that convinces the patient, the symptoms may disappear, although they may also persist or return. Here lies the central problem of psychoanalysis and psychotherapy. Given the vast, possibly infinite human potential to assign symbolic relations and functions, how can we determine and treat the relevant cause of any particular disturbance?

One first step would be to identify classes of events that are likely to function as pathogenic. Freud himself (1923) devised a classification of mental events according to a trilogy of concurrent constructs he called "ego," “superego," and "id." The “superego" was a monitor created by internalizing the directives of parents and society about what was or was not acceptable. The "id" was the sum of basic "drives." The "ego" was the self-concept ("character") being stimulated by the "id" and controlled by the "superego" in everyday reality. In a repressive society, the "ego" is likely to be caught in severe conflicts between these two sources of commands. Any such conflict could become a plausible candidate for triggering a disturbance. The events of the conflict would he dismissed from conscious awareness and given some symbolic representation that in turn waited to reassert itself in an error, or in severe cases, in a syndrome.

Recent critics have noted that this Freudian scheme tends to propagate an anthropomorphic (or at least animistic) view of the mind as a set of agents (Peterfreund and Schwartz, 1971). Freud himself was prone to describe his constructs in terms that would be more fitting for a person. In one source (Freud, 1940), for instance, he says: "the ego strives after pleasure.... the ego gives up its connection with the external world and withdraws into a state of sleep" (p. 146). Nash (1963) comments that "Freud's metaphor of intrapersonal persons" makes processing look like a "dramatic interaction" of the "inner players of a play-within-a-play" (p. 341). The unity of the human processing system is obscured by these agents whose exact relationships to each other in any specific event or disturbance are very hard to clarify. To view behavior as driven by inner conflicts among undifferentiated animate opponents is to grossly oversimplify the case and to obscure the more subtle modes of therapy.

A more abstract account would be that a causal disturbance can be converted into a symbolic symptom if the person is rejecting it and refusing to deal with it for what it is. This rejection could repress the conscious awareness of the disturbance, but not the unconscious need of the disturbance to be resolved. Eventually, this need engenders a different-order symbolic disturbance that appears to demand an organic (usually medicinal) resolution. Yet, since the causality is elaborately symbolic, organic treatment is more likely to reduce the symptoms than to provide a cure. Only if the patient would recognize and deal with the original disturbance could an effective cure be attained.

Freud reasoned that the patient represses the cause if the latter is in some domain that instills strong negative affects, such as fear, anxiety, or repulsion. Freud's preoccupation with sexuality stemmed from the evident fact that this domain, more than any other, aroused those negative affects in the society of his times. For instance, he assumed that an "instinctual drive" that was strongly repressed might convert desire to the opposite feeling of repulsion-a physiologically complex change, but a symbolically simple one. A thing is more closely associated with its opposite (e.g., "night" and "day") than with an unrelated thing (e.g., "night" and "elbow"). The polarity of opposites is recognizable precisely because of their similarity of structure (Freud, 1916-1917).

Another grave problem is how to gather evidence of symbolic constructs. Freud freely took evidence from his own introspection, other people's reports, observed errors and disturbances, and even myths (e.g., Oedipus). Like many psychoanalysts after him, Freud often had to take the patient's word for what the original experience was like and what it meant. Freud's early work used hypnosis to get patients to reveal clues about pathogenic events. Later, he found that therapeutic discourse was often more useful.[1] Patients were encouraged to perform "free association," whose content was a free flow of symbolically related materials. Freud (1916-1917) hypothesized that this activity would reveal personal constructs and representations tangential to the disturbance, e.g., to the pathogenic event for a syndrome. He thought that the usual monitoring of one's utterances would be relaxed, and that each association would be "strictly determined by important inner attitudes of the mind, which are unknown to us at the moment" (p. 96).

Freud soon realized that his success with neurotic patients suggested a very general phenomenon: the ability of mind and body to create and enact symbolic representations of all sorts. He surmised that "those commonplace occurrences which have been cast aside as all too insignificant" can "betray very important things" (p. 26). The errors of everyday life might be signals of the "interference of two different intentions" (p. 40). He saw dreams as a striking illustration that, without conscious control, symbols for one's real or imagined experience are devised and elaborated with amazing complexity and subtlety. Here, also, Freud assumed that the tendency to transfer or convert would be stronger if the person's "conscious" mind had not resolved the material; and again, sexual issues were the most obvious candidate for conscious rejection. However, many other sources were also uncovered by close analysis, such as extravagant wishes, ambitions, guilt feelings, and unpleasant or salient experiences in one's past.

On the other hand, the extreme variability of symbolic constructs is probably being underestimated if we routinely expect to find sexual issues as the cause. We clearly need further modes of evidence besides introspection, verbal reports, and errors. Despite common beliefs to the contrary, there is more experimental evidence bearing on Freud's theories than on most other theories of personality or development (Fisher and Greenberg, 1977). However, there is little consensus about how Freud's theories can he tested or validated, and practical obstacles abound. Subjects might well conceal or disguise the very data that an investigator would want to evaluate; or if they did not, then repression and avoidance couldn't be as strong as Freud claimed. Many developments would require several years (e.g., the impact of early childhood on adult neuroses), during which continual observation would-be necessary. In addition, strict, stable causalities are hard to find; for example, even if one were to reconstruct an experience and then put the patient to sleep, it is unlikely that the same dream would be triggered again.

There are problems even in using the battery of methods in psychiatry and personality psychology, such as interviews, questionnaires, inventories, Rorschach blots, and anxiety measures (survey in Fisher and Greenberg, 1977). Many of these methods entail an elaborately symbolic component-how subjects want to represent themselves, how they interpret the test and the testing situation, and so forth-whose elucidation is critical for a full evaluation of the findings. The symbolic function adds a complex variable to causalities governing much of human activities, including scientific research itself. Until this function can be reliably accounted for, psychoanalysis and empirical psychology will not be firmly reconciled.

BASIC PROCESSING OF INFORMATION

Here is where the information-processing approach offers a major advantage by defining both physiological and symbolic entities in common terms as structured events in human processing. It seeks to model the complex dynamics involved and to find unified modes of representation. It acknowledges that all human actions have a symbolic aspect, leaving no grounds for excluding from study those traditionally eschewed as "subjective" by experimentalists; in fact, there is no way to do so. No mind-versus-body dualism is needed-the physiological and the symbolic are simply two aspects of processing events.

The classical "stimulus-response" mechanism was usually demonstrated with extremely simple cases. The stimulus was sharp and direct (e.g., an electric shock), and the response followed regularly and immediately (e.g., pushing a lever or running through a maze). This simplicity, combined with a preference for low-order animals as subjects, created the illusion that the whole mechanism involved no symbolic aspect. The most exaggerated extension was Skinner's (1957) attempt to explain language itself as direct "conditioning," without regard for the intensely symbolic character of communication. Predictably, all of Skinner's illustrations came from reductive situations in which people's utterances stimulated, or were stimulated by, external events (e.g., "pass the salt").

From the standpoint of information processing, it is clear that in most human situations the relationship between a stimulus and a response is partly symbolic, obeying the cognition that the stimulus and the response represent things that make them relevant to each other. In the causality of a neurosis, this assignment of representations is atypically personal and mediated; symptoms are triggered by events that, in terms of social agreement and physiological necessity, need not cause such effects. A substantial portion of pathogenic "conditioning" (configuring of processes) is done in a mental domain to which both the patient and society are denied immediate access. Impulse configurations are reconstructed until an idiosyncratic stimulus (converted experience) elicits a manifest response (external symptom). This exceptional causality does not make the symptoms any less real or distressing for the patient. The more the symptoms intervene in one's actions, the more they tend to acquire "reality" in the sense defined at the start of this paper. From an information-processing standpoint, the "conscious mind" is the sum of attentional events, whereas the "unconscious mind" is the sum of nonattentional events. The "preconscious," which Freud (1900) seemed to view as the buffer for materials passing from the unconscious into the conscious, would he the sum of events that are close to the threshold of activation but have not actually arrived (cf. Morton, 1970). If we assume that the electrochemical processes of thought and memory are constantly acting upon their own conditions and substances, then there should be some unconscious activity going on at all times, no matter what the person may be attending to. Of course, attention can direct a large portion of cognitive resources to an intended cognition or operation. But, the prior structure of one's knowledge can "preconfigure" incidental processing whose relatedness to current attentional focus is accidental or remote.

Freud tended to envision the unconscious as the domain of affect and the conscious as the domain of cognition, although the foregoing scheme suggests that no such division is made-there could easily be conscious affects and unconscious cognition. Weiner (1975) has shown that basic cognitive tasks can bring a conflict-resolving insight into disturbing affects, indicating a closer relationship between the affective and cognitive domains than is assumed in traditional psychoanalysis. However, the fact that affects are evidently powerful and persistent states makes it probable that they (or events including them) often elude immediate resolution and call for additional processing on the unconscious level. This probability would be particularly strong if people try to deny or conceal affects in social settings-a long-standing tendency in some cultures.

In automatic processing, the person selects a preconfigured action pattern and lets it run without conscious supervision (Norman and Shallice, 1980; MacKay, 1982). Automatic process,es are hard to set up and monitor, but once they are working, they are rapid and economical (Shiffrin and Schneider, 1977). Freud (1935 himself noted that slips and errors may he absent during automatic processing, yet may occur when attention is in focus:

A great number of actions may be carried out in a purely automatic way with very little attention and yet quite successfully many acts are most successfully carried out when they are not the objects of particularly concentrated attention, and mistakes way occur just on occasions when one is most eager to he accurate, that is, when a distraction of the necessary attention is certainly not present [p. 29].

Apparently, ongoing operations cap be influenced or even controverted by some processing events that are not receiving either attention or intention.

Consider what happens during 'everyday processing. When an event (an experience or cognition) is registered, it is rapidly interpreted in working memory and then stored as a representation in longer-term memory. Its nature and significance are constituted by recognizing it as an instance of, or a n analogy to, something that was already known and stored. This whole operation can he defined as problem-solving. finding a resolution for the unknown or unfinished, and carrying out any tasks involved (Newell and Simon, 197@). To the extent that this particular event is novel, unexpected, or complex, it may not be fully accounted for before it gets stored. It way remain fragmented, ambivalent, incomplete, or problematic, and therefore want further processing, even though the person's conscious resources have been directed elsewhere. It follows that this problem-solving sequel can only be done unconsciously until some threshold of satisfactory account is attained; and even then, the dynamic nature of storage might enact further reorganizations in the wake of still newer events. The ability to construct rich analogies and associations allows considerable freedom for further interpretations of any one event. In general, preference will go to problem-solving operations that impose order and explanation on experiences. Yet, Freud's famous slips and errors might he a side-effect of peripheral processing that resolves problems in unexpected or undesirable ways (e.g., by revealing an intention the person wants to conceal).

Suppose that the event is significantly disturbing because it conflicts with major expectations or motives. The person has to make a choice in a hurry: to retain the disturbing event in working memory where it is most influential and conspicuous, and try to resolve it somehow; or to push it out of working memory into storage where, though still unresolved, it is at least not in the focus of attention. Obviously, the second solution-immediate release from intense disturbance--looks attractive, especially when a satisfactory resolution might demand large and arduous alterations in one's structure of expectations and motives. However, the event thus "repressed" (dismissed from working memory) goes on demanding an account and presumably attracts some resources from the peripheral processes that are steadily acting upon storage. These processes should tend to convert the event in hopes of reducing or removing its disturbing aspects. The symbolic function is activated to solve the problem of creating a version remote from the proximate but unwelcome one. Conditions should now favor the idiosyncratic and surprising constructs typically uncovered during psychoanalysis. The person would naturally be unaware of these constructs manufactured outside of conscious awareness. Later events may trigger associations with these constructs, though the individual has no way of recognizing what is going on. Instead, negative associations are elicited by events whose relation to the original negative sensation is not consciously known, since the progress of symbolic conversion has not been monitored.

A person's set of socially conditioned directives on action constitutes what Freud called the "superego." One's set of biologically conditioned directives on action constitutes what he called the "id." However, these sets are probably neither independent, nor in steady conflict, unless there is some additional division between the two domains, such as all functions of the "superego" being conscious or cognitive (rational) and all functions of the "id" being unconscious or affective (emotional). However, during everyday processing, the thresholds of conscious versus unconscious or cognitive versus affective are flexible in regard to both social and biological motivations. In between the "superego" and the "id," Freud situated the ,&ego" interacting with everyday reality and correlating its specifications and requirements with these internalized sets of directives. The ego should have considerable input from both the cognitive and the affective domains, but so far, it is not clear why the ego must be trapped in conflicts between these two domains of a different order and with different consequences. Peterfreund and Schwartz (1971) suggest that

phylogenetically older systems, clinically those involved in drives and emotions, are able to play especially crucial roles in motivation.... Though they may he relatively less complex information-processing systems in themselves, they may nevertheless enter into the overall information processing of the organism at crucial points; they may enter deeply into the logical organization of programming; they may have high-priority positions. They can therefore exert an extremely powerful influence, appear to "demand attention," and even, at times, make the organism appear to he "out of control" in the clinical sense [p. 181]

Although the biological and affective domains may be older, a phylogenetic account still seems inadequate by itself (as if the "mind versus body" dualism could be exchanged for an "older versus newer" dualism). The priority of these domains in current processing might be accounted for if: (a) a simpler organization tends to be activated faster and more easily than a complex one; (b) predominantly affective states persist in memory without rehearsal better than predominantly cognitive ones; (c) the denial of affect-oriented goals is more unpleasant than the denial of cognition-oriented ones; and so on.

Hence, more research is needed to determine why and how acute or pathogenic conflicts occur between affective and cognitive aspects. Freud (as well as Nietzsche) assumed that society had adopted an unrealistically hostile outlook toward biological needs, such that the conscious actions of the "ego" become diffuse and distorted. In terms of processing, routine events could receive divisive interpretations and elude definitive resolution until an elaborated construct of poorly integrated ideas and experiences engenders manifest disturbances of thought and action, i.e., errors and syndromes.

Freudian psychoanalysis operates on the principle of bringing such idiosyncratic constructs and associations to the patient's attention, where they can finally be@ reviewed and modified. At least, the possibilities for a resolution are greatly enhanced if this can be achieved. The analyst typically encourages free association in hopes of tracing out unconscious relations that a more precisely defined and controlled memory search would miss or avoid. Of course, the patient may "resist," i.e., persist in rejecting or concealing the underlying causalities for the same motives that led to the original act of repression (Fine, 1973). Weiner (1975) advocates having the patient do conflict-free cognitive tasks as a means of circumventing entrenched defenses surrounding a pathogenic conflict. This approach might simplify psychotherapy, which has traditionally been complicated, circuitous, and time-consuming (Hartley and Strupp, 1980).

This sketch indicates, however, that as long as social directives conflict with biological ones, a tendency to interpret events pathogenically will persist throughout society. Psychotherapy often brings about improvements that are later reversed when the patient reenacts the same or similar processes as those that engendered the original syndrome. Such cases may he instances of "regression" back to an earlier and/or more primitive conflict-stage (Freud, 1900). However, they could also result merely because the person's total environment of action and interaction is still laden with conflict. The system unconsciously repeats symbolic conversions, even if the person consciously knows from experience that such a conversion doesn't lead to a resolution.

How can we envision the details of a nonattentional symbolic conversion? Probably, the physiological (neurological) basis is highly influential. Two events could be contaminated and lead to disturbances or errors if ' in mental storage, they are contiguous (i.e., because they were originally associated in time or causality), or analogous (i.e., because their structure or features are similar in some way). These configurations could "print through" onto each other and result in a seemingly unaccountable action (Norman, 198 1). The contiguity or analogy could be quite abstract if it occurred at the very deep level where events are represented as configurations of electrochemical impulses. Two events might he stored in contiguity merely because they happened to the person at roughly the same time or place, or under comparable circumstances. Structural analogies on this level could obtain between experiences or cognitions that look quite incongruous to common sense observation or introspection. The closeness of opposites, cited above and very prominent in Freud's theories, is a case in point.

A more complex possibility would be that memory processing decomposes an event and works on its components through a variety of associations in such a way that the end product appears incoherent to anyone who doesn't understand its origin and evolution. This additional step should confuse the person or the observer because the complexity of the conversion has now increased by several orders of magnitude. Such seems to be the case with many dreams where elaborate interpretations and revisions of memory items take place.

No doubt, the complexity and subtlety of such processing, and the absence of attention, would make it hard to anticipate, recognize, or control the results. Still, we could try to specify which conditions would encourage conversion:

1. if any event is too novel or complex to be handled by habitual means;

2. if processing is overloaded by the current demands imposed on it;

3. if fatigue or distraction interfere with normal, processing;

4. if the solutions attained by initial problem-solving are unsatisfactory;

5. if an unpleasant event is pushed out of working memory before it can he satisfactorily interpreted or resolved;

6. if motives or ideas strongly conflict with each other; or

7. if an internal motivation conflicts with an external compulsion.

In all of these cases, processing should be prone to conversions which, under certain conditions, could become pathogenic. The more pronounced the problem, the more probable that the converted configuration will appear deviant compared to the usual modes of sense-making and behaving. Complex physiological and symbolic conversions could elicit thoughts and actions that countermand what the person consciously believes or intends. The processing known as spreading activation occurs when an activated item "spreads" its active status to items associated with it in the mind (Collins and Loftus, 1975). The spreading is automatic and helps create contexts that support the recognition and interpretation of ongoing events and the execution of complex actions (MacKay, 1982). Morton (1970) has proposed that activation depends on accruing levels of neural ex- citation, and thus is capable of subtle nuances. Freud's "preconscious" would contain items fairly close to the threshold, while the "unconscious" would contain ones that are relatively remote; in between would he a gradation rather than a clear or fixed border. The outcome of spreading activation is pre- configured in fine detail by the structure of one's current knowledge. Freud (1935) envisioned comparable prior controls on "free association":

An association.... will be strictly determined by important inner attitudes of mind, which are unknown to us at the moment when they operate.... in spite of our apparent freedom of choice in such matters we cannot, in point of fact, think of any name which cannot be shown to be narrowly determined by the immediate circumstances, the idiosyncrasies, of the person [pp. 96-97]

This conjecture is one of the factors that make Freud's theory seem deterministic. Still, the diversity of available associations in mental storage allows considerable "freedom" among possible outcomes (compare Knight, 1946). Freud is merely saying that associations must have some prior structure in order to become operative; but the person might be wholly unaware of what that structure is. Freud's position is that all mental events are motivated, not random. What occurs during processing has to respect both the structure of one's prior knowledge and the organization of one's neurological processes. This claim fits the contention that no symbolic function can be entirely separate from its physiological aspect.

The role of creativity can be defined within this framework. Although a person's knowledge necessarily has some structure by virtue of the contiguities and analogies already constructed, the range of possible new constructs is undoubtedly vast. The difficult question is how such new constructs are interpreted and motivated once they appear (Beaugrande, 1979). Creative associations and representations are typically novel, but not random; they do not deny reality, but renew our perception of reality from an altered perspective.

If unconscious processing is also in part a vehicle for problem-solving, the concept of "unconscious creativity" (a domain of the so-called "primary process") seems quite natural (Neisser, 1967) alongside the more familiar concept of conscious creativity. The argument has been advanced that the creativity of dreams is in part a vehicle for problem-solving (French and Fromm, 1964; Lipton, 1967; Greenberg and Pearlman, 1978). Crick and Mitchison (1983) even contend that dreaming has the task of discharging false or nonsensical memories generated when the non-specific signals from the brain stem periodically excite the neocortex during REM sleep (Melnechuk, 1983); since the goal is to unload the neocortex of worthless signals, the conscious attempt to remember one's dreams might be counter-productive. However, this conclusion need not follow. The discharged ideas and images might afford important clues to the states and preoccupations of the brain and thus shed some light on the nature and origin of pathogenic signalling.

In summary, the information-processing approach would account for the mind's organization and operation in more abstract and general terms than those envisioned by the main Freudian tradition. What metapsychology has viewed as unitary phenomena--ego, drive, regression, depression, and so forth-are explored as complex aspects of processing in a diversified system (Rosenblatt and Thickstun, 1977). For example, the domains of conscious, preconscious, and unconscious are probably constituted not as fixed units, but as adaptable aspects within the functions of perception, comprehension, memory storage, planning, and performance. Whatever dominances are found, such as affect versus cognition, or routine versus creativity, are distributed across this general system, but not in fixed proportions or sequences. Such fixity would indeed imply that major acts of learning or therapy, such as those we can observe in actual practice, should be scarcely attainable. In reality, the most impressive property of the mind is its potential and inclination for evolution. On the negative end, pathogenic events cannot be ruled out; on the positive end, there is always room for processing to be reconfigured in more strategic modes.

Similarly, mental events are not likely to have a unitary causality (an issue I noted as a dispute between empirical psychology and psychoanalysis). In emulation of physics, some psychoanalysts assumed what Peterfreund and Schwartz (1971) call "simple, cause-and-effect, linear, one-to-one, sum-of-parts explanations" (p. 81), or what Rosenblatt and Thickstun (1977) refer to as the "billiard ball conception" (p. 16). The latter authors point out that multiple causalities are more common in a complex interactive system (compare Waelder, 1936; Moore and Fine, 1968); and that human behavior is determined by the person's motives and reasons as well as by physical or external causes (compare Sherwood, 1969). These factors greatly increase the complexity of explanation, therapy, and control. Even the more subtle and ingenious reconstructions of the psychoanalyst are probably gross simplifications of the actual pathogenesis. In the future, researchers might profit by pursuing the precepts outlined so far, such as:

1. Physiological and symbolic entities are defined in common terms as structured events in human processing.

2. The conscious is the sum of attentional events, whereas the unconscious is the sum of nonattentional events; the border between the two is nuanced and flexible.

3. Because electrochemical processes of thought and memory are constantly acting upon their own conditions and substances, unconscious activity is always in progress when the mind is at work.

4. Associations can he formed through any sort of contiguity (time or causality) or analogy (similarity of structure or features).

5. Associations can enable print-through between configurations that, in everyday observation of behavior, might seem disparate or unrelated.

The generality of these precepts reveals the need for extensive specification in psychoanalytic practice. In the next section, I shall try to outline some implications for the origin and resolution of mental disturbances.

PATHOGENESIS AND SYNDROME

Psychological disturbances, such as neuroses, evidently arise from protracted electrochemical conversions. Here, the print-through that conflates contiguous or analogous configurations (or the components of these) must he abnormally extensive and intricate. A chain reaction or a lengthy reiteration (like a looping circuit) in nonattentional processing might add up to a massive conversion of otherwise innocuous or routine configurations.

Who actually contracts a neurosis should depend on several factors. First, some proclivities appear to be genetically transferred from parent to child. These cases should be particularly hard to treat, if, as seems plausible, genetically transferred tendencies are extremely persistent. However, to explain paranoia via the 44 automatic interference" of an "archaic theme," as Peterfreund and Schwartz (1971, p. 261) suggest, seems hardly less vague than Jung's appeal to "archetypes." After all, the system is in its current state at any one moment, and its history and residue must be accounted for in those terms. I suggested already that case of activation or persistence in memory might be involved.

Second, the events of early childhood seem to he strong determinants. Freud himself attributed this factor to the attempts of society to repress child sexuality. However, some negative effects might arise simply because the child's processing system is prone to be overloaded or confused by the demands of less familiar experiences. Fewer skilled or automatic routines are developed to conserve resources. Social directives are not as firmly internalized as in adults. Thresholds for anxiety and unpleasantness are harder to control. External compulsions seem unduly strong and restrictive. Thus, the conditions favoring symbolic conversions, as stated above, would naturally be acute for children, and patterns could arise that persist into adult life without resolution, until a neurosis appears.

Third, anxiety can encourage a neurotic interpretation of otherwise innocuous entities. In processing terms, anxiety is a fairly common state when a person confronts tasks in which the outcome is uncertain or failure entails unpleasant consequences. The person feels inadequate, disoriented, or threatened. According to Murray (1971), moderate anxiety can stimulate performance, but, beyond a certain threshold, anxiety begins to overload the system and to degrade performance. In the latter case, a further increase in anxiety is likely to occur, followed by further degradation, then more anxiety, and so on, in a self-aggravating loop. If conversion is probable during overload, anxiety would be a good candidate for the extensive and idiosyncratic conversions that produce neuroses.

Fourth, relative probabilities may be a factor (Rosenblatt and Thickstun, 1977). Basically, the whole nervous system could he probabilistic at the level of the firing of individual neurons (Burns, 1968). On higher levels within a large system of possible states and events, there is always some probability of pathogenic interpretations and conversions occurring in nonattentional processing. That probability should increase according to the conditions suggested above-novelty, processing overload, conflicts, and so on. But every system would be subject to at least some. probability for mildly pathogenic occurrences. These might be resolved without the emergence of a neurosis; or the neurosis might take on a harmless nature and eventually subside. However, the chance of more virulent syndromes would he very hard to rule out in any one person.

I shall now look at some specific types of disturbances within the framework applied so far.

ERRORS

An error would be probable whenever the person has a choice among two or more actions whose organization encourages a conflation (Norman and Shallice, 1980). Dyspraxia is the condition where errors occur frequently in persistent patterns. As already noted, structural analogy is one important contributor. Freud (1916-1917) listed such occasions as when a person tries to open an office door with the key to his house door. Contiguity is another important contributor. Reverend Spooner is reported to have sprinkled claret on some salt he'd spilled, whereas it was normal to sprinkle salt on spilled claret to relieve stains (Potter, 1976, p. 3); the actions were interchanged because they were often done contiguously. However, the structural analogies and contiguities that lead to disturbances need not be obvious to introspection or observation. Although neurotics are evidently not conscious of the associations they make, healthy persons might also be unaware of their own mental structures.

Freud suggested that the person unconsciously wishes to commit the error as a form of escape, release, self-punishment, and so on. For example, the man who tried to open his off-ice door with the key to his house door is claimed to be wishing he were back home; the lady who injured her face and the one who broke her leg jumping from a carriage were diagnosed as punishing themselves for self-appraised misconduct (Freud, 1916-1917). Unlike the relation between prior knowledge and processing, this determinism seems unduly strong and implies single rather than multiple causality. The analyst is obliged to reconstruct motives, whether or not the evidence calls for it in a given case. In effect, the probability of accidental, unmotivated error.% is not properly evaluated. Surely, the analogy or contiguity leading to conflation need not involve a hidden wish; at most, a wish could use those contributors to assert itself at a good opportunity.

To see how this might work, let us suppose that a strong wish arose, where "wish" is defined as an "impulse directed toward a representation of a desired future state of the world." Suppose also that the wish is consciously rejected as conflicting with social directives. According to the model I sketched, the impulse would not just go away, but would be dismissed from working memory without resolution, Then, the impulse might wait to assert itself when an analogous or contiguous action that was not rejected came onto the agenda. Or, the impulse might merely search for any weak spot in the system. Then, an error would be made when the person was performing an action normally taken for granted; a mental disturbance could symbolize itself in symptoms of organs that are easily afflicted (headaches, sore throats, upset stomachs, etc.); and so on. Note that the error or disturbance is still a product of problem-solving in that the manifestation implements (albeit in another domain) an impulse waiting for resolution.

However, as Freud noted, errors also intrude precisely when the person is straining to avoid them. Apparently, the impulse to reject the error has been reconfigured in nonattentional processing by converting or filtering out the inhibition component from the total representation. The person, unaware of this unconscious. change, might register and execute the unblocked impulse without recognizing it. The erroneous action then appears to be just as deliberate as a correct one. A rejected wish would naturally try to unblock its execution in this way, but many other actions could occur without a wish simply because pf structural conflation.

Repetition may be another decisive contributor. Each time an idea comes to mind, the corresponding electrochemical configuration is activated. Each repetition could add a few degrees of concentration that eventually add up to a profound substantive change at that location, e.g., a high density of neurotransmitters. The change eventually unblocks an erroneous or rejected action the person had no intention of performing. If so, constantly worrying about an error is likely to bring it about rather than to prevent it. Inattention arising from repetitious actions can lead to costly and dangerous errors, e.g., in the operation of complicated machines (Fitts and Jones, 1961).

ANTISOCIAL ACTS

An antisocial act could be an extreme stage of this conversion process. Either the mental representation of a strongly rejected action was stripped of its inhibition component, or the inhibition was converted to its opposite, the compulsion. As already noted, opposites encourage print-through via their analogous structure. It is striking how often an antisocial act flies directly in the face of basic social directives. Many crimes might have no other motive than the fact that they are, indeed, crimes. Their rejection is contaminated by their (similarly structured) affirmation; if this process happened frequently, then probabilities favor the outcome that, in a certain percentage of cases, antisocial acts are actually performed with the same directedness as their positive counterparts, e.g., murder vs. life-saving, stealing vs. giving, destruction vs. preservation, and so on. The person might never have consciously "wished" to commit the act; the mere knowledge of the act, plus the unconscious conversion that leads to unblocking, could suffice as a trigger.[2]

NEUROSES

Neuroses could result from a variety of conflations and conversions outlined above. In the traumatic syndrome, one very stressful event triggers processing that the person strongly desires to dismiss from attention, but the powerful affects involved preempt considerable resources in unconscious processing and lead to pronounced disturbances. The phobia is a specific reaction to a symbolic object not normally considered to elicit extreme anxiety. The massive strength of the phobia corresponds to its narrow focus, as compared to anxiety in general. A Association interrupts processing, e.g., in loss of memory or orientation, as an escape from possibly undesirable results. A compulsion (or "obsession") causes the constant performance of seemingly meaningless acts that, in extreme forms, entail perversion, aggression, or violence: the unblocking conversion has attained an extreme concentration at the opposite end of the scale. These symptomatic neuroses arise from milder, more sporadic conversions than do the character neuroses. The latter emerge when pathogenic conversions come to dominate a person's routine processing, leading to such syndromes as erratic delusions, jealousy, grandiosity, and feelings of persecution. Apparently, the converted processing establishes itself in a stable and resistant organization and dominates processes that would otherwise he unrelated and unaffected.

FIXED IDEAS

The fixed idea is a striking case of a self-perpetuating repetition, like a looping circuit or an absorption state in systems theory. The pervasive and damaging character of the fixed idea could arise from a drastic chemical imbalance as repetition creates an extreme concentration level, far beyond the normal level of a mental configuration. The idea can then print-through onto an abnormally wide range of otherwise unrelated ideas and comes to dominate cognition and behavior. Naturally, the same imbalance is prone to distort the idea itself as well as the patient's view of the idea.

HYSTERIA

In hysteria, conflation and imbalance are sufficiently intense to cause extreme manifestations, such as acute pain or paralysis. A pathogenic event, or an event subjected to strong pathogenic conversion, continues to build deep-level discomfort until an external release-valve is found. That valve may simply be a weak or susceptible point in the physiological system, as appears to be true of some familiar psychosomatic illnesses. A cure may be brought when the "insight" of psychoanalysis brings the original event and its conversion to conscious attention, where it can also be more readily reconfigured and thereby converted back to a harmless version. The loop is broken and the event loses its power to generate the symptom. Still, this result might not be dependable: the analysis might fail to reconstruct the correct or complete pathogenesis; the patient might use some other event to trigger the same effects; or, in some types of regression, the loop seems to be re-established after the harmless version is displaced from attention and dissipated in the continuing activities of mental storage. Hence, hysteria might resist a lasting cure through analysis-as Freud had to admit.

SCHIZOPHRENIA

Schizophrenia has been defined since Bleuler as a lack of coordination among human systems. Presumably, the reconfiguration of mental storage, while creating new associations via print-through, also produces dissociations. A dominant but abnormal configuration might defend itself against inspection or change by blocking access from other configurations. Soon, this blocking of access disrupts the unity and consistency of cognition and behavior and projects disparate characters for the same person, since "character" is an abstraction across a person's actions and attitudes. Another explanation might be that print-through begins doubling or multiplying alternative configurations for the same set of events, until two or more conflicting systems have been assembled and compete for control of processing. A massive print-through between alternative constructs is followed by a profound confusion of categorical boundaries. To the outsider, the assignment of dominance to one system or another might appear random and erratic; but in the patient's own mind, certain events could, on a deep level, tend to prefer one system over the other and thereby act as triggers for an abrupt character shift.

PSYCHOSIS

In psychosis, the patient's relationship to the whole model of socially established reality undergoes a dissociation, possibly along the lines depicted a moment ago. If we assume that "real" versus "unreal" are components or markers of stored knowledge configurations, then the psychotic patient is one who regularly fails to maintain those components. Despite this change or deterioration, other components of circumstantial detail may remain unaffected. Thus, the psychotic may seem intelligent, alert, and efficient much of the time. Yet when the "real/unreal" parameter becomes acute, behavior can he totally unpredictable to the outsider.

PARANOIA

Paranoia is a collusion of anxieties or phobias that leaves the patient in constant fear of persecution. Presumably, a defense component participates in the routine interpretation of events, leading to such familiar conversions as sublimation, isolation, projection, repression, and so on, which comprise an abstract repertory of displacements in directional, temporal, and numerical contiguity, or in structural analogy (Holland, 1973). These defense conversions might he grossly overactivated by an undue electrochemical concentration. Virtually any event could then he taken as a symbol for danger or disaster, and this pattern would reinforce the concentration each time it was repeated. The intensely symbolic nature of the threat allows the paranoid to persist in unfounded fears despite the lack of genuine occasions for defensive action. When a very advanced stage is reached, the paranoid may be no longer able to initiate any actions (including ones that might alleviate the condition) because all consequences, are feared.

GRANDIOSITY

Grandiosity (or "delusions of grandeur") compensates for feelings of anxiety and inadequacy by converting one's self-concept into an aggrandizing symbolic construct. This conversion might leave intact some domains not closely associated with the explicit role of the self. Thus, the patient appears rational until the parameter of "self versus other" becomes acute. Of course, one's self-concept influences so many actions that grandiosity is likely to be almost as pervasive and debilitating as paranoia. The grandiose patient overperforms as radically as the paranoid patient underperforms. Eventually, repetitive enactment could cause such a high concentration of the imbalance that the customary coherence of thought and action are severely and consistently degraded.

CONTROL AND THERAPY

The sketch of some classical syndromes in the previous section, although obviously tentative, suggests some reasonable conclusions. First, a certain probability of errors and neuroses is built into the system of human processing, with its concurrent physiological and symbolic aspects. Given the complexity and adaptability of the system and the staggering number of possible event sequences, there will always be some probability of pathogenic constructs. Moreover, society will always have motives to reject some of a person's wishes, especially a child's, in the interests of order and unity. Every person will at some time undergo "repression" from without and impose it from within.

Second, it follows that a "healthy" person is not one who has eliminated all chance of pathogenic constructs, but one who can confront and control them or their manifestations before a damaging level of intensity is reached. A certain amount of errors and symptoms in everyday life is presumably a beneficial release-valve for reducing and regulating the load on the system, e.g., rejected wishes waiting for resolution. In a certain percentage of cases, this release becomes imbalanced and detrimental. However, since pathogenic constructs have a physiological basis as well as a symbolic one, and since the nature and control of these constructs are poorly understood, it is senseless to blame the patient for the syndrome. The neurosis is just as motivated and "real" (i.e., participates in one's orientation) as an illness due to an organic lesion; the causalities are just more complex, idiosyncratic, and mediated than in nonneurotic processing.

Third, such oppositions as "healthy" versus "unhealthy" and "organic" versus "neurotic," are not strict demarcations, but subtle shadings sensitive to the Person's context. By overstressing these oppositions, we only exaggerate the difficulty of therapy and cure. Moreover, the traditional labels have the effect of giving the patient a role to act out, thereby reinforcing the imbalance that engendered the syndrome in the first place. In the role of "schizophrenic," "paranoid," etc., the patient finds an elaborated symbolic system for thought and behavior. in incurable or highly resistant patients, this system has evidently enacted a conversion so powerful and extensive that it can't be dislodged, or can be so only temporarily. Its constructs print through onto virtually all events of processing.

Fourth, the classical methods of Freudian and post-Freudian psychoanalysis are inadequate to the extent that they are fixated not merely on the symbolic aspect, but on restrictive assumptions about it. For example, Freud himself found the issue of sexuality implicated in so many patients that he failed to consider the possibility of its being a special case rather than the general case. He tended to view sexual conflict as the cause, rather than the effect, of nearly all syndromes. For instance, he tried to define "personality types" as determined by the stages of childhood sexuality ("oral," "anal," and "genital") (Freud, 1908); however, Freud didn't consider that these types and stages might themselves be only one set of symbols for more abstract symbolic proclivities, e.g., modes of defining self versus nonself or for imposing order on experience. In this view, the “oral" type would be dominated by symbols of ingestion (absorption, accumulation, etc.), while the "anal" type would be dominated by symbols of elimination (purification, externalization, etc.). There is no doubt that childhood sexuality is prone to engender anxieties and conflicts in a society that drastically denies and represses it; but many other events can trigger pathogenic sequels.

Equally disproportionate was Freud's (1915) conviction that all paranoid delusions are defensive attempts to negate and contain homosexual impulses. His grounds were that such impulses appear to be disturbing to many paranoid patients, a suggestion that has been confirmed by some clinical studies (Zamansky, 1958; Sternlof, 1964) but not others (Sheflin, 1969; MeLaughlin, 1972). This divergence of findings is accountable if we assume that paranoids are generally anxious about strong social prescriptions, among which homosexuality may or may not be chosen as a representation. Thus, a more basic and pervasive tendency for exaggerated anxiety could encourage a pathogenic interpretation of specific sexual issues and events-not just the other way around.

These four conclusions indicate that psychoanalysis could become more flexible and general if we were able to redefine its central notions within a more abstract theory of human processing. Above all, we could use a model of how people process their knowledge and experience in order to anticipate pathogenic developments (estimated in terms of probabilities) and to redirect them before a syndrome becomes pronounced or entrenched. This project calls for vastly more differentiated methods of description and diagnosis. Although analysis is obliged to use the patient's introspective reports, we need a more complete theory of how these relate to relevant mental data (Peterfreund, 1975, 1980; Labov and Fanshel 1977). Even under hypnosis, such reports are not always reliable (Loftus, 1980).

Both implicitly and explicitly, psychoanalysis has already evolved from the state Freud left it in toward more abstract and encompassing models. Alfred Adler stressed the balance between inferiority and superiority as motivations. Carl Gustav Jung subsumed in the collective unconscious the individual unconscious that dominated Freud's approach. Harry Stack Sullivan, Erik Erikson, Erich Fromm, and Karen Horney highlighted social and environmental factors Freud had neglected. Carl Rogers and Abraham Maslow defined humans as capable of "self-actualization" in a uniquely experienced world, where strategies must he devised and implemented to fulfill the person's needs. In the wake of such revisions, the influence of social interaction upon the ego has been given as much weight as its internal preoccupations, and the role of sexuality is no longer considered totally pervasive.

All the same, my impression is that these revisions have mainly adopted the Freudian framework as their point of orientation. For instance, the theories are still predominantly conceptualized in anthropomorphic terms with agents: "ego," "id," "superego," "self," "me," "parent," "family," "persecutor", etc., that "defend," "repress," "project," "displace," "threaten," "evaluate," and so on. We should explore the merits of a "Project" in which a fundamentally distinct framework, such as the information-processing approach, is used to model the Freudian framework from the outside. We would improve our understanding and treatment of syndromes if we could build detailed process models of their origin and operation. By recasting the usual agents, such as "ego," in terms of processes, we can restore them to their original dynamic meaning as collective designations for patterns of human actions.

Therapy itself is still sparsely mapped out in terms of processing. Freud and his successors relied mainly on the therapeutic value of uncovering and reenacting past experiences. The analyst is expected to interpret materials, help lower the patient's resistance, and foster insights (Wolberg, 1967; Fine, 1973). However, if analysis itself tends to predict certain findings, it may reduce or bias the range and interpretation of evidence. For example, given Freud's emphasis on childhood events, the analyst is typically cast in the role of a parent (Hartley and Strupp, 1980). Moreover, patients ideally improve when, with the analyst's help, they consciously figure out the origins of their syndrome and restructure the pathogenic conversion of the causal event. The frequent protraction or failure of analysis shows the vulnerability of this ideal and calls for a more process-oriented account of therapeutic evolution.

An account we might explore is that an insight helps to substitute a balanced counterpart for a pathogenically imbalanced processing configuration. However, assuming that a record of that configuration persists, the patient cannot actually expel it entirely from memory; at best, skillful psychotherapy can bring about a synthesis between the balanced and the imbalanced configurations so that whatever remains of the latter now has a new and positive function. What conditions ought to be necessary or supportive for such a synthesis?

One possibility is that conscious attention paid to a non-pathogenic event allows it to override the pathogenic event that had exerted itself in nonattentional processing. Probably, however, this attentional effort must be of a certain kind before the positive effect can occur and the syndrome is overcome. Otherwise, attention may only fixate the problem but not resolve it, or may even aggravate it. At the very least, the focus of attention should be accompanied by the patient's strong conviction that an improvement will occur. Another possibility is, that the analysis triggers a general destabilization of processing tendencies so that a major reorganization can begin. "Dynamic" therapies, such as those of the “gestaltists" (e.g., Fritz Perls) and the "phenomenologists” (e.g., Carl Rogers, George Kelley), aim at this fundamental reorganization by redefining and transforming personal relationships, especially between analyst and patient. Since their focus is much broader, the chances are better for influencing a syndrome with abstruse or multiple causes. This second possibility lends credibility to the notion that neurosis is in part controlled by the patient's inclination to act out the role of a neurotic. The old practice of incarcerating patients together in isolation from nonneurotics afforded ample opportunity to understudy a role. Because attention is decisive y in learning, remembering, and behaving, paying attention to neurotic symptoms may reinforce them better than nonattentional processing could. Perhaps a neurosis can become a symbolic representation of itself as well as of a converted event in one's experience, and if so, the chain of causalities would be more complex. Such a factor could explain the unregenerate quality displayed by many patients receiving traditional therapy that presupposes the simpler origin. In contrast, the dynamic therapies seek to dismantle the role of "neurotic" and to reinterpret one's available character components as parts of a healthy personality. The information-processing approach is in turn far more encompassing than the dynamic revision therapies. The drawback is that the abstractness, complexity and diversity of information-processing makes the operationalizing of tests and therapies extremely difficult. However, the state of the art in direct testing is not much more refined within traditional approaches. For example, trying to recognize neurotics on the basis of conventional physiological gauges (e.g., electroencephalographs, brain scans, X-rays, angiograms) has turned out to be less reliable than tests having symbolic (e.g., categorizing, speech perception) as well as physiological (e.g., finger-tapping, rhythm) elements, such as the Wechsler Adult Intelligence Scale and the Halstead-Reitan Battery (Levine and Feirstein, 1972; Filskov and Goldstein, 1974; Goldstein and Halperin, 1977). By and large, tests direct people to think and behave in ways that may not reveal what occurs spontaneously in everyday life. Therefore, we need more abstract models and more sensitive methods to determine the reliability of correlations between detectable electrochemical states and processing states. Refined and nonnoxious organic detectors, among which positron emission tomography seems quite impressive (Ter-Pogassian Raichle, and Sobel, 1980), especially with computer enhancement, could analyze and compare a wide range of electrochemical states in healthy, neurotic, and organic patients. If characteristic configurations are found, these three groups could he identified independently of how they classify themselves or are classified by individual therapists or conventional tests. Reliable correlations between physiological versus symbolic functions might even enable us to differentiate patterns Specifically enough to recognize subtypes of psychopathology, such as hysteria, paranoia, and schizophrenia. The same data might serve to develop much more finely tuned means of intervention. Electric and chemical changes could be precisely designed and controlled in order to help reconfigure nonhealthy electrochemical states as healthy ones. Traditional electroshock, medication, and surgery are too much like salvo-bombing -- massive discharge throughout the system in hopes of hitting a small 'elusive target (with impact on many nontargets). A refined knowledge of the nature and constitution of electrochemical states in human processing would enable specialized, accurate physiological targeting, such as changing the concentrations of neurotransmitters.

Furthermore, the discovery of reliable correlations between detectable electrochemical states and processing states would support a more perspicuous interpretation of the patients' verbal evidence-complaints, reports, interviews, and so forth. Several of my students have analyzed discourse samples from patients diagnosed as having schizophrenia, Alzheimer's disease, and closed-head trauma.[3] Although we observed that certain linguistic features are typical of neurotic discourse, such as word-repetition and rhyming, these features are far from adequate for diagnostic purposes, and are not entirely absent from normal discourse. Moreover, a few patients appeared to be cultivating those features (especially rhymes) in order to play their role as neurotics. We also tested the attitudes of outside observers by comparing their judgments of samples from neurotics versus samples from healthy subjects, and the features identified as neurotic were found in the control samples as well. These common conceptions about the presumed features of neurotic discourse could easily become part of a patient's role-playing.

Such findings indicate that, as a configuration of symbols, the surface text by itself is not a very reliable indicator of the speaker's cognitive states; at the very least, the production of the text is a vastly complex processing event (Beaugrande, 1984). A therapist frequently has to rely on intuition and guesswork when inferring a syndrome from what patients say. However, if we could gauge electrochemical states characterizing the syndrome and correlate them with electrochemical states characterizing the production and reception of discourse, we would finally have independent criteria for assessing the ratio between pathological manifestations and symbolic representations in language. We would also he better able to understand and control syndromes by taking into account the pathogenic and therapeutic potential of discourse itself (Labov and Fanshel, 1977). Moreover, we could provide a better account of the larger sociological contexts of human communication, such as those whose importance has been shown in the "dynamic" therapies.

All this presupposes a method of analysis to capture extremely fine differentiations in both physiological and symbolic substrates. Highly detailed theoretical models of cognition, far beyond those now in circulation, will be required for this enterprise. Otherwise, even the most exact physiological data can elude consistent and meaningful interpretation; the greater the exactness, the more stringent the theoretical demands. Models of human knowledge are required on a large and ambitious scale (see Minsky, 1975; Brachman, 1978; Fahlman, 1979; Findler, 1979). For some time, these models will necessarily remain underdifferentiated in comparison to the human processes themselves, but at least steady progress can be expected. This is not true as long as we tolerate an extensive theoretical gap between the physiological and the symbolic.

The recourse I have tried to outline could bear on many issues other than psychotherapy. The most obvious one is developmental learning. So far, we can gauge only the gross results of learning that can he practically manifested in certain classes of performance. We have little notion of the physiological and symbolic changes that occur in successful learning, as compared to failure or limited success. Though still disputed, the phenomenon of "visceral learning" demonstrated by Neal Miller signals that even fairly gross electrochemical intervention can y bring about complex learning (Jonas, 1973). A vastly more refined mode of intervention might well open an unprecedented technology for guiding and optimizing human learning. Here too, the theoretical basis may prove harder to create than the technological implementation.

Other applications are not hard to envision: the improving of self-concepts and social expectations, the circumvention of defenses that block cognitive development, the reduction of routine errors, the alleviation of aggression and hostility, the correction of criminal and other antisocial tendencies, and many more. Of course, a malignant authority might abuse the same methods to bring about harmful effects. Still, it seems reasonable to hope that the positive applications could at least reduce people's inclinations to carry out the negative ones. As such, this particular advance could mediate against human tendencies to abuse it.

To the practicing psychoanalyst and empirical psychologist, this project may seem utopian-as indeed many notions we now take for granted seemed in times past. However, the project is derivable from recognized fundamental principles in physiological and symbolic processing. Admittedly, the details have yet to be supplied in a great many domains, and there is danger that the correlation between these two basic aspects of processing may prove unreliable or at 'least unsystematic within our normal understanding of the notion of "system." Nonetheless, I hardly think that this correlation can be random, given the extent of uniformity needed to support human cognition of reality and communication of ideas. At the very least, we ought to probe the correlation before we make a judgment.

If the project were to succeed, we would probably want to revise the theory and practice of psychoanalysis and psychotherapy. However, we would also want to retain parts of them with a clearer understanding of why they work and under what conditions. The present lack of interchange between psychoanalysis and process psychology is neither in the interests of patients nor investigators, nor indeed of a society consistently beset by its own failure to control and recast pathogenic constructs.

NOTES

*For helpful comments on a draft of this paper, I am greatly indebted to Robert Boice (Albany), Michel Grimaud (Wellesley), Norman Holland (Gainesville), Elizabeth Loftus (Seattle), Don MacKay (Claremont), Donald Norman (La Jolla), Begonia Roji Menchaca (Salamanca), Bernard Paris (Gainesville), David Rubin (Durham), and Insup Taylor (Toronto).