We see only what we know.
GOETHE
NO UNDERSTANDING OF responses to loss, whether they be healthy or pathological, is possible without constantly invoking concepts of defensive process, defensive belief and defensive activity—the three categories, I argue, into which defences are best grouped. In this chapter a sketch is given of how the phenomena observed and the processes postulated can be understood within the conceptual framework adopted. Although here and there comparisons are made between the present theory and certain of Freud’s concepts of defence and mental structure, for reasons of space no systematic attempt is made to relate the two models.
The conceptual tools on which I draw have been made available by students of human information processing. These tools enable us to examine defensive phenomena from a new point of view, to collect data more systematically and to formulate hypotheses in a language shared by other behavioural scientists. These are great advantages. Nevertheless, there is clearly a long way to go before the theory sketched is within sight of doing justice to the wide range of defensive phenomena met with clinically. Until more work has been done, therefore, it will remain uncertain how successful the new approach is going to be.
In the first volume of this work, at the end of Chapter 6 and throughout Chapter 7, I have drawn attention to current work in neurophysiology and cognitive psychology that points to the central control of sensory inflow. Whether inflow derives from the environment through exteroceptors or from the organism itself through interoceptors, sensory inflow goes through many stages of selection, interpretation and appraisal before it can have any influence on behaviour, either immediately or later. This processing occurs in a succession of stages, all but the most preliminary of which require that the inflow be related to matching information already stored in long-term memory. All such processing is influenced by central control and is done at extraordinary speeds; and all but the most complex is done outside awareness.
For most purposes the inflow of interest to psychologists and the common man alike is that which, having been selected, interpreted and appraised, goes forward to influence mood and behaviour and/or to be stored in long-term memory. The fact that in the course of its being processed the vast proportion of initial inflow is routinely excluded, for one of several reasons, is ignored. For the understanding of pathological conditions, by contrast, the interest lies in the opposite direction, namely in what is being excluded, by what means it is excluded, and perhaps above all why it should be excluded.
In the ordinary course of a person’s life most of the information reaching him is being routinely excluded from further processing in order that his capacities are not overloaded and his attention not constantly distracted. Most selective exclusion, therefore, is both necessary and adaptive. Like other physiological and psychological processes, however, in certain circumstances selective exclusion can have consequences that are of doubtful or varying adaptive value. For example, given certain adverse circumstances during childhood, the selective exclusion of information of certain sorts may be adaptive. Yet, when during adolescence and adult life the situation changes, the persistent exclusion of the same sorts of information may become maladaptive. The defensive processes postulated by psychoanalysts, I believe, belong in this category. To distinguish these unusual instances of selective exclusion, of only temporary adaptive value, from the overwhelming majority of adaptive instances it is convenient to refer to ‘defensive exclusion’.
The basic concept in the theory of defence proposed is that of the exclusion from further processing of information of certain specific types for relatively long periods or even permanently. Some of this information is already stored in long-term memory, in which case defensive exclusion results in some degree of amnesia. Other information is arriving via sense organs, in which case defensive exclusion results in some degree of perceptual blocking. As is made clear later in this volume, the many other phenomena described by clinicians as defensive, notably certain types of belief and certain patterns either of activity or inactivity together with their associated feeling, can be understood within this framework as being the profound consequences of certain significant information having been excluded. Correspondingly, analytic therapies can be understood as procedures aimed at enabling a person to accept for processing information that hitherto he has been excluding, in the hope that the consequences of his doing so will be equally profound.
In presenting the theory, attention is given first to the basic questions of how information, of any sort, can first be selected and then deliberately excluded. Next we consider briefly the nature of the specific information that is liable to be selected for prolonged and defensive exclusion. Only after that do we broach the two further questions: what are the causal conditions that lead certain information to be excluded for long periods of time? and what are the advantages and disadvantages of doing so? In proceeding thus we move from the less controversial questions to the most.
As regards findings from experimental work, it happens that, up to date, more light has been shed on the selective exclusion of information during the processing of sensory inflow than has been shed on the selective exclusion of information already in store. For that reason prior attention is given to studies of subliminal perception and perceptual defence. Since, however, no perception is possible without the interpretation of sensory inflow in terms of matching information already in store, it is plausible to suppose that the mechanisms employed for preventing certain information from being retrieved from store bear some resemblance to the mechanisms employed for excluding from further processing information of similar or related import arriving through the sense organs. Given this, what is known about subliminal perception and perceptual defence can be taken as a paradigm.1
The notion that information of certain meaning could be selectively excluded from perception met with considerable scepticism when first proposed around 1950. How, it was asked, can a person selectively exclude a particular stimulus unless he first perceives the stimulus which he wishes to exclude? At first sight this might seem a conclusive argument, especially if it is assumed that perception is some sort of singular event which either happens or fails to happen. But, as Erdelyi points out, the objection ceases to have any force once perception is conceived as a multi-stage process. For during processing through a sequence of stages it would be at least possible for certain information to be excluded before it reaches some final stage associated with consciousness. There is now abundant evidence that this can happen.
After some decades of controversy and steadily improving experimental techniques a multi-stage theory of perception is now widely accepted. Some features of it relevant to a theory of defence can be summarized.
The recognition of pattern as it occurs during perception proceeds in two directions simultaneously. On the one hand, the arrival of a sensory stimulus triggers an automatic series of analyses that start at the sense organs and continue centrally far up the chain of processing stages. On the other hand and simultaneously, the situation in which the sensory events are occurring triggers expectations based on past experience and general knowledge. These expectations produce conceptually driven processing in which guesses are made about what the input probably means. As the two forms of processing merge the guesses are checked against the data and the task completed.
By proceeding in both directions simultaneously the process of recognition is greatly accelerated. Yet by relying so much on expectations derived from past experience and knowledge the possibility of error is much increased. For example, because it lies outside experience a black three of diamonds when seen briefly is commonly misperceived as a three of spades. Findings of this type cast light on several characteristics common in responses to loss.
A second feature of a modern theory of perception is that sensory inflow can be processed outside a person’s awareness to a stage sufficient for much of its meaning to be determined. Thereafter it can influence his subsequent behaviour, including his verbal responses, without his being aware of it. Experiments using the technique of dichotic listening illustrate these points.
In this type of experiment two different messages are transmitted to a person, one message being received in one ear and the other in the other. The person is then told to attend to one of these messages only, say the one being received by the right ear. To ensure he gives it continuous attention he is required to ‘shadow’ that message by repeating it word for word as he is hearing it. Keeping the two messages distinct is found to be fairly easy, especially when they are spoken by different voices. At the end of the session the subject is usually totally unaware of the content of the unattended message. There are, however, certain exceptions. For example, if his own name or some other personally significant word occurs in the unattended message, he may well notice and remember it. This shows at once that, even though unattended, fairly advanced processing of the unattended message must be taking place.
The results of two experiments that used this technique illustrate how information derived from the unattended message can influence thought and/or autonomic responses even though the message never reaches consciousness.2
In one such experiment subjects were required to attend to and to shadow ambiguous messages of which the following is an example:
they threw stones towards the bank yesterday
Simultaneously with this message either the word ‘river’ was presented in the unattended ear or else the word ‘money’. Later, subjects were presented with a recognition test for the meaning of the sentence in which they were asked to choose between the following:
(a) they threw stones towards the side of the river yesterday;
(b) they threw stones towards the savings and loan association yesterday.
Subjects who had had the word ‘river’ presented to the unattended ear tended to select (a) as the meaning, whereas subjects who had had ‘money’ in the unattended ear tended to select (b). None of the subjects remembered what word had been presented to the unattended ear and were unaware also that their subsequent judgement of meaning had been influenced.
Clearly, in order for the word presented to the unattended ear to have the effect it did in this experiment, it must have undergone sufficient processing for its meaning to have been recognized. A similar conclusion emerges from another experiment that also used the technique of dichotic listening.
Before the experiment proper the subjects went through a few training sessions during which they were exposed to an electric shock when any one of a set of selected words was spoken to them. As a result subjects became conditioned to the word-shock combination so that whenever one of the selected words was heard it was responded to by a change in the GSR (a measure of sweating). In the experiment proper the subjects were required to attend to and shadow a message in one ear while a list of words was presented to the other, unattended, ear. Words in that list were of three kinds: neutral words, some of the words that had been conditioned to shock, and both synonyms and homonyms of those words. Despite the fact that no shocks were given during the experiment itself there was an appreciable rise in the GSR whenever a conditioned word was presented in the unattended ear. Of even greater interest is that there was also a substantial, though lesser, rise when the homonyms and synonyms were presented. Here again the findings indicate that every word presented in the unattended ear must have undergone considerable processing and its meaning established.
From these findings it is but a short step to infer that, just as a person’s judgement and his autonomic responses can be influenced by cognitive processing outside awareness, so also can his mood. Once that is assumed, a mechanism becomes available in terms of which certain changes of mood otherwise inexplicable can be explained.
On the basis of findings such as those described cognitive psychologists propose that an analytical mechanism exists that performs a series of tests outside awareness on all incoming messages. As a result of these tests information can undergo one of several fates amongst which the following are easily specified:
The criteria by which during the series of tests information is judged for allocation are clearly numerous and range from broad and simple to specific and complex. Furthermore, many of these criteria, perhaps all, can be changed by central control. Some such changes, we know well, are a result of conscious and voluntary control as for example when, after receiving a new instruction, attention is shifted from one ear to another or from one voice to another. Other changes, we also know, occur involuntarily and outside awareness as for example when a person’s attention shifts to the other voice when he hears his own name mentioned by it.
Once the possibility of subliminal perception is accepted, theoretical objections to the idea of perceptual defence and its counterpart, perceptual vigilance, drop away. For what the findings from the many hundreds of experiments undertaken in this field show is that, in addition to its being able to influence judgement and autonomic responses, the processing of sensory inflow for meaning outside awareness can influence also the further inflow of that very information itself. Either the inflow may be reduced, as in perceptual defence, or it may be enhanced, as in perceptual vigilance. Examples of these findings are drawn from Dixon’s (1971) detailed examination of the evidence.
Many experiments have been done using a tachistoscope which enables words or pictures to be shown either at different speeds or else at different light levels. Since these speeds and light levels include those that are either too fast or too dim for perception to be possible, a common procedure is to start by showing a word or a picture at a speed or light level known to be impossible, and then gradually to reduce speed or increase light until the subject becomes able to identify the stimulus. A well-attested finding from experiments using this technique is that, when words or pictures known to be emotionally arousing or anxiety provoking are presented, the time taken before they are correctly identified differs significantly from that taken to identify neutral words or pictures. To demonstrate that these results are due to changes in the sensory channels and not in the response channels, other experiments have been done. In some of these a significant change of sensitivity for the sensory inflow being received through one sense modality, say sight, is found to occur when the stimulus being presented through another modality, say hearing, is changed from an emotionally arousing one to a neutral one, or vice versa.
The direction in which a change of this sort occurs differs for different individuals. In some sensitivity to emotionally arousing words is found to be habitually increased, whereas in others it is habitually decreased.
In the experiments so far described the changes being effected in sensory inflow are being effected solely by involuntary means. In certain other experiments, however, it is found that subjects may also be regulating inflow by means of eye movements or eye fixations. In so doing the subjects are employing their voluntary musculature although without being aware they are doing so and, as in all these experiments, without their being aware either of the nature of the stimulus being presented to them. In thus utilizing both involuntary and voluntary effectors, Dixon points out, the systems regulating sensory inflow resemble the systems maintaining body temperature, regulation of which can be achieved either by involuntary means, e.g. reducing peripheral circulation by capillary restriction, or by voluntary means, e.g. putting on extra clothes.
It is not without significance for the scientific status of theories of subliminal perception and perceptual defence advanced by cognitive psychologists that these theories are fully compatible with theories of sensory processing advanced by neurophysiologists. There are in fact many physiological mechanisms that in principle could play the required part.
One possibility, described by Horn (1965, 1976) and for which evidence is accumulating, is that temporary reductions can be effected in the responsiveness of neurones in the sensory pathways. The means to do so are thought to be reductions in the level of a special priming input which these neurones require. Another possibility, different to though fully compatible with the first, is described by Dixon (1971)3 whose review of the neurophysiological literature was guided by Dr R. B. Livingston.
Whether perceptual defence and vigilance are mediated by mechanisms of these sorts or by others is of no great consequence to us. What matters is that we now have good experimental grounds for believing, as every clinician does, that sensory inflow can be processed outside awareness and that, depending on the meaning assigned to it, further inflow can either be enhanced or reduced. That being so, it becomes reasonable to consider whether perhaps there may be other stages of out-of-awareness processing at which it may be possible for analogous processes of defensive exclusion to operate.
In an attempt to clarify the processes underlying perceptual defence and vigilance Erdelyi (1974) has proposed a flow diagram I find attractive and that is compatible, at least in principle, with ideas advanced by Norman (1976), MacKay (1972), Manner (1975) and Hilgard (1974) on whose work I am also drawing. Amongst their other merits Erdelyi’s proposals suggest a way to understand the role of that small but important part of information processing that occurs within consciousness.
The mental apparatus can be thought of as made up of a very large number of complex control systems, organized in a loosely hierarchical way and with an enormous network of two-way communications between them. At the top of this hierarchy we postulate one or more principal evaluators and controllers, closely linked to long-term memory and comprising a very large number of evaluation (appraisal) scales ranged in some order of precedence. This system, or possibly federation of systems, I shall call the Principal System(s), thus leaving open the question whether it is best regarded as singular or plural.
On the inflow side the task of these Principal System(s) is to scan all raw data as it becomes available (for fractions of a second or at most a second or two in ‘sensory register’4), undertake a preliminary analysis and evaluation of it in terms of stored knowledge and relevant scales, and then send commands to an encoder regarding what should be selected for further processing and what should be discarded. Not only does all this preliminary scanning and sorting take place outside awareness but information rejected at this stage is likely to be permanently lost (although, as the experimental study of hypnosis, discussed later, shows, this may not always be so). This is the stage at which perceptual defence, or vigilance, is postulated to take place.
The reason for having this preliminary scan, Erdelyi suggests, is that the channels engaged in all the more advanced processing are of limited capacity and therefore incapable of dealing with more than a small fraction of inflow. Main bottlenecks appear to be at the stages of encoding, first, for short-term storage and, later, for long-term storage.
Information selected after preliminary scan for further processing, having already been encoded for short-term storage, is then in a form likely to give rise to the conscious perception of objects in a space-time continuum. Thus, in Erdelyi’s words, perception is ‘the conscious terminus of a sequence of non-conscious prior processes’, and occurs probably in the region of short-term storage. ‘While the span of consciousness or conscious perception is small, the span of perceptual processing and analysis is probably vast.’
After the stage of short-term storage and conscious processing, some information is selected for further encoding and eventual storage in long-term memory; other information, having served its purpose, is discarded.
During the past decade experimental psychologists have been giving much thought to the concept of consciousness which is now accepted as being scientifically ‘respectable, useful and necessary’, to quote Mandler (1975) whose ideas I draw upon.5
Consciousness can be regarded as a state of mental structures that greatly facilitates certain distinctive types of processing to occur. Among those are the following:
(a) the ordering, categorizing and encoding of information (which is already in an advanced state of processing) in new and further ways prior to storage;
(b) the retrieving of information from long-term storage by framing simple addresses to extract it from complex memory structures;
(c) the juxtaposition of information of varying kinds, e.g. representational models, plans and sensory inflow, derived from diverse sources; this makes possible reflective thought;
(d) arising from (c) the framing of long-term plans by preparing an array of alternative plans and sub-plans and then evaluating them, thus making possible high level decisions;
(e) the inspection of certain overlearned and automated action systems, together with the representational models linked to them, that may be proving maladapted. As a result of such inspection, systems and models long out of awareness become available for reappraisal in the light of new information, and, if necessary, attempts can then be made to reorganize or, perhaps, to replace them.
The relevance and value of this fifth and last function of conscious processing for the practice of psychotherapy will at once be evident; for it is conceived as enabling certain structures (or programmes) basic to personality to be reappraised and, if necessary, in some degree modified. Let us consider two such basic structures, (a) that which mediates attachment behaviour and (b) that which applies all those rules for appraising action, thought and feeling that together are usually referred to as constituting the super-ego. Both these programmes are conceived as being stored in long-term memory and as being ready to be drawn upon to participate in processing and planning action as inflow from exteroceptors and interoceptors seems to indicate.
Both the nature of the representational models a person builds of his attachment figures and also the form in which his attachment behaviour becomes organized are regarded in this work as being the results of learning experiences that start during the first year of life and are repeated almost daily throughout childhood and adolescence. On the analogy of a physical skill that has been acquired in the same kind of way, both the cognitive and the action components of attachment are thought to become so engrained (in technical terms overlearned) that they come to operate automatically and outside awareness. Similarly, the rules for appraising action, thought and feeling, and the precedence given to each, associated with the concept of super-ego are thought also to become overlearned during the course of childhood and adolescence. As a result they also come to be applied automatically and outside awareness.
Plainly this arrangement has both advantages and disadvantages. On the one hand, it economizes effort and, in particular, makes no demands on the limited capacity channels mediating advanced processing. On the other is the disadvantage that, once cognition and action have been automated, they are not readily accessible to conscious processing and so are difficult to change. The psychological state may then be likened to that of a computer that, once programmed, produces its results automatically whenever activated. Provided the programme is the one required, all is well. Should an error have crept in, however, its correction not only demands skilled attention but may prove troublesome and slow to achieve.
The upshot is that, provided these representational models and programmes are well adapted, the fact that they are drawn on automatically and without awareness is a great advantage. When, however, they are not well adapted, for whatever reason, the disadvantages of the arrangement become serious. As anyone who has developed a bad style in some physical skill knows well, to review the cognitive and action components of a system that has long been automated and to change it is arduous and often frustrating; moreover, it is not always very successful. Hence some of the difficulties encountered during psychotherapy.
This, however, is neither the only problem nor the greatest. For the task of changing an overlearned programme of action and/or of appraisal is enormously exacerbated when rules long implemented by the evaluative system forbid its being reviewed. An example of this, highly germane to what follows, is when a person finds himself unable to review the representational model(s) he has built of his attachment figure(s) because to do so would infringe a long-learned rule that it is against one or both his parents’ wishes that he study them, and their behaviour towards him, objectively. A psychological state of this kind in which a ban on reviewing models and action systems is effected outside awareness is one encountered frequently during psychotherapy. It indicates the existence of another stage of processing at which defensive exclusion can also take place, different to the stage at which perceptual defence occurs.
Yet further evidence of the part played by information processing outside awareness, and of the power of selective exclusion in keeping it so, derives from studies of hypnosis.
As a result of a long experimental programme Hilgard (1973) concludes that during hypnosis, and as the result of the hypnotist’s suggestion, what he calls the Executive Ego, which I shall call Principal System A, assigns control to a Subordinate System, which I shall call Principal System B. Following System A’s assignment of control, the hypnotist’s orders are received, processed and acted upon by System B without System A being in any way aware of what is being processed. Furthermore, these orders are being continuously scanned, still outside System A’s awareness, by an evaluative system. This becomes plain whenever System B receives an order that would be unethical to obey and refuses to comply with it. To criticisms that the Executive Ego is only pretending to be unaware of what orders are being received, Hilgard replies effectively by pointing to the genuine surprise expressed by his subjects when, subsequently, they see or hear taped recordings of their sessions.
Many of Hilgard’s experiments were concerned with hypnotic analgesia. Pain was produced by placing the subject’s hand and forearm in circulating ice water for 45 seconds. Ordinarily this causes him to show many signs of discomfort, e.g. grimacing and restlessness, and to report that he is experiencing great pain and distress. In addition, changes are found to occur in his heart rate, blood pressure and other physiological measures, of most of which he is also well aware. When the suggestion is made under hypnosis that his hand will be analgesic to the pain of the circulating ice water, by contrast, the same situation produces no visible signs of discomfort, whilst System B reports neither feeling pain nor being aware of autonomic changes.6 Yet the changes are occurring just as in the unhypnotized subject.
From these findings Hilgard concludes that in the hypnotic condition System B is able to exclude, selectively, sensory inflow from two types of interoceptor, namely the pain-endings and those reporting autonomic activity. The mechanism that he postulates thus appears to be a counterpart to the mechanism responsible for perceptual defence, which excludes inflow from the exteroceptors.
It is interesting to learn how the subjects themselves experience the session, which can be done if, before hypnosis is ended, they are instructed to recall what their experiences have been. Some report concentrating on imagining that their arm is numb. Others use what Hilgard refers to as a ‘dissociative technique’ in which, for example, the subject might concentrate on the separation between his arm and his head, or on imagining himself going away to the country where all is quiet.
In discussing his findings Hilgard (1974) notes that, while some are compatible with Freud’s theory of repression but not with Janet’s theory of dissociation, others are more compatible with dissociation theory than with psychoanalytic. Thus Hilgard’s findings show that the exclusion of information that would normally be accepted is an active process requiring effort, which is a point integral to Freud’s theory but missing from dissociation theory. On the other hand, the findings show that the dissociative process segregates organized systems from one another, as Janet and other advocates of dissociation theories emphasize and in contrast to Freud’s notion of an unorganized, chaotic id. Because the position Hilgard adopts resembles dissociation theory but yet differs from it in certain critical respects, he describes his position as neo-dissociative.
Further information about what is going on in hypnotic states is available from those few subjects who are capable whilst under hypnosis of enabling yet a further system to communicate by means of automatic writing and automatic talking.7 Hilgard refers to this third system as the ‘hidden observer’; I shall call it System C.
When these subjects took part in an experiment in which hypnotic analgesia is induced so that System B is unaware either of pain or of autonomic changes, System C reports, by contrast, being aware of both (though the pain may be at an intensity a little less than in the normal non-hypnotic state) (Knox et al 1974). Here again it is of great interest to learn how the subject himself experiences the session.
The account following was given by one who had taken part in a session during which System C had reported, by means of automatic talking, whenever the subject had felt the experimenter’s hand on his shoulder. (In accordance with instructions the subject refers to System C as the ‘hidden observer’.)
In hypnosis I kept my mind and body separate, and my mind was wandering to other places—not aware of the pain in my arm. When the hidden observer was called up, the hypnotized part had to step back for a minute and let the hidden part tell the truth. The hidden observer is concerned primarily with how my body feels. It doesn’t have a mind to wander and so it hurt quite a bit. When you took your hand off my shoulder, I went back to the separation, and it didn’t hurt any more, but this separation became more and more difficult to achieve.
The experimenter comments that it was very apparent from the subject’s grimaces and movements that he was feeling intense pain whenever her hand was on his shoulder. When, however, her hand was removed his face gradually relaxed and he appeared comfortable again. Yet after the session was over the subject made plain that effecting the dissociation required constant effort and that he found it difficult to maintain.
A further point emphasized by Hilgard is that the interruption to communication between systems that occurs in the hypnotic state is rarely complete. Often one system has some knowledge of what is going on in the other, even if the second has no knowledge of what is going on in the first. The existence of partially permeable barriers may provide a lead for understanding the phenomena that clinicians refer to, paradoxically, as ‘unconscious feelings’.
In recounting these experiments I recognize that only a small minority of individuals are susceptible to being hypnotized and that an even smaller proportion are capable also of automatic writing. (Of the student population tested by Hilgard, no more than one or two per cent proved suitable.) Yet the findings made it clear that, at least in some persons, the mental apparatus is such that not only is a dominant system capable of excluding selectively much sensory inflow that would normally reach consciousness but also that the processing of this excluded inflow may reach a state of consciousness within another system parallel to but segregated from the first.
Experimental findings of this sort, together with comparable findings by clinicians, raise difficult questions of how best to conceive of the self. In the case of Hilgard’s experiments, it may be asked, is Principal System A to be regarded as the self and, if so, what do we make of Principal Systems B and C? Or should all three be regarded as selves? And, in the clinical field, how can we most usefully conceptualize what Winnicott calls a false self and how contrast it with what he calls a true self?
In approaching these problems it is useful to start with Hilgard’s proposal that what he terms the Executive Ego is the system that, being capable of self-perception, becomes capable also of conceiving of the self as an agent; and, further, that the integrity of that system is provided through its constant access to a more or less continuous store of personal memories. Questions that then arise are, first, whether we can conceive of more than one system becoming capable of self-perception and, second, whether there is evidence that the memory store may be sectionalized; and, if so whether it is plausible to postulate that in some individuals barriers to communication are set up between two or more major sections of it.
Within the conceptual framework advanced neither of these proposals raise problems of principle.
As regards the first, MacKay (1972) in his discussion of how we may suppose conflict to be regulated postulates a hierarchy of evaluating and organizing systems, in which the higher systems can be described as meta-systems, meta-meta-systems, and so on with indefinite extension. Whereas in a hierarchy of this sort it is customary to think of the arrangement as moving steadily upward from a large array of lower systems to a single system at the top, other configurations are possible. For example, it is possible to consider two or more systems at the top working in greater or less collaboration with each other. Whereas such an arrangement might be less efficient than one in which the chain of command is unified, it might nevertheless be more flexible. The point I wish to make is simply that a plural arrangement of this kind is well within the bounds of possibility and cannot be ruled out on a priori grounds. It is for this reason that earlier in this chapter I speak of Principal System(s), thereby leaving open the issue singular or plural.
The question whether it is reasonable to suppose that the processing of information can reach the phase of consciousness within more than one Principal System is also not to be ruled out on a priori grounds, especially as we still remain totally ignorant of what the special conditions are that determine whether or not processing ever reaches this phase.
The second question posed earlier concerns the possibility of the personal memory store being sectionalized and communication between sections being impeded or blocked. Here again there are no a priori difficulties, whilst such evidence as we have is entirely compatible with such notions.
In his discussion of long-term memory, Norman (1976) emphasizes that there is more than one way in which information can be encoded for storage and that the same information can be encoded, and can coexist in storage, in several different forms, and can also be accessible by any of several different routes. For example, a mental representation can encode information about the world in an analogue form, which mirrors certain selected properties of the world as in a map or a mechanical model, or it can encode information in a propositional form, which comprises a set of interpreted abstract statements about perceptual events as in a prose description. The design of the human cognitive system, Norman believes, allows flexibility in the way it represents information. Not only can it employ whichever system of encoding best suits its purposes but it appears able also to transform one form of representation into another. For example, analogical representation appears to be well suited for the storage of operations and action programmes, whereas propositional representation appears well suited for the storage of the meaning and interpretation of events. A compromise in which different forms of encoding are used in combination to represent different aspects of the world is probably available also. The information conveyed in diagrams, Norman points out, is often partly in analogue form and partly in propositional.
Norman also draws attention to the distinction, introduced by Tulving (1972),8 between storing information according to personal experiences, autobiographically, and storing it according to its meaning, its contribution to personal knowledge. Since I suspect this distinction may have very significant implications for psychopathology, it is worth examining further.
In the episodic type of storage, information is stored sequentially in terms of temporally dated episodes or events and of temporospatial relations between events. It commonly retains its perceptual properties and each item has its own distinctive place in a person’s life history. ‘Thus, an integral part of the representation of a remembered experience in episodic memory is its reference to the rememberer’s knowledge of his personal identity’ (Tulving, p. 389). An example would be a person’s lively recollections of the events that occurred during a particular holiday. In the semantic type of storage, by contrast, information exists as generalized propositions about the world, derived either from a person’s own experience or from what he has learned from others, or from some combination of the two. Inflows into the semantic memory system, therefore, are always referred to an existing cognitive structure. Examples would be any views the person might form about holidays in general and how any particular holiday might compare with others.
A corollary of the distinction between episodic and semantic storage, and one likely to be of much clinical relevance, is that the storage of images of parents and of self is almost certain to be of at least two distinct types. Whereas memories of behaviour engaged in and of words spoken on each particular occasion will be stored episodically, the generalizations about mother, father and self enshrined in what I am terming working models or representational models will be stored semantically (in either analogical, propositional or some combined format). Given these distinct types of storage a fertile ground exists for the genesis of conflict. For information stored semantically need not always be consistent with what is stored episodically; and it might be that in some individuals information in one store is greatly at variance with that in the other.
My reason for calling attention to the different types of storage and the consequent opportunities for cognitive and emotional conflict is that during therapeutic work it is not uncommon to uncover gross inconsistencies between the generalizations a patient makes about his parents and what is implied by some of the episodes he recalls of how they actually behaved and what they said on particular occasions. Sometimes a generalization refers in broad and glowing terms to a parent’s admirable qualities, some or all of which are called sharply in question when episodes of how he or she had actually behaved and/or spoken are recalled and appraised. At other times the position is reversed, with the generalization being uniformly adverse and what is recalled from episodes being appraised more favourably. Similarly, it is not unusual to uncover gross inconsistencies between the generalized judgements a patient makes about himself and the picture we build up of how he commonly thinks, feels and behaves on particular occasions. For these reasons it is often very helpful for a patient to be encouraged to recall actual events in as much detail as he can, so that he can then appraise afresh, with all the appropriate feeling, both what his own desires, feeling and behaviour may have been on each particular occasion and also what his parents’ behaviour may have been. In so doing he has an opportunity to correct or modify images in semantic store that are found to be out of keeping with the evidence, historical and current.
One reason for discrepancies arising between the information in one type of storage and that in another lies in all likelihood in there being a difference in the source from which each derives the dominant portion of its information. Whereas for information going into episodic storage the dominant part seems likely to derive from what the person himself perceives and a subordinate part only from what he may be told about the episode, for what goes into semantic storage the emphasis may well be reversed, with what he is told being dominant over what he himself might think. An everyday example of a large discrepancy between information in episodic storage and what is in semantic storage is found in the images we have of the earth we live on. In our daily round we experience the earth as flat and for most purposes we treat it as though it were so. Yet most educated Westerners, having learned that it is spherical, would claim that their model of it is indeed spherical. In this case, of course, although the discrepancy exists, no conflict of emotional consequence is nowadays experienced. In the case of information about parents and self, however, on which so much of emotional consequence does turn, major discrepancies are likely to produce a disturbing sense of unease.
Let us return now to the questions posed at the beginning of this section, namely how within the framework proposed we can best conceive the self and how it may be possible also to conceive of a person having more than one self. In most individuals, we may suppose, there is a unified Principal System that is not only capable of self-reflection but has more or less ready access to all information in long-term store, irrespective of its source, of how it is encoded and in which type of storage it may be held. We may also suppose that there are other individuals in whom Principal Systems are not unified so that, whilst one such System might have ready access to information held in one type of storage but little or no access to information held in another, the information to which another Principal System has, or has not got, access might be in many respects complementary. The two systems would then differ in regard to what each perceived and how each interpreted and appraised events, which is exactly what we seem sometimes to meet with clinically. In so far as communication between systems is restricted, they can be described as segregated.
When during therapy a patient compares the discrepant images he has both of his parents and of himself that derive from stores of different types, the images from episodic storage are, I suspect, those he most often judges to have the greater validity and to be the ones with which he most closely identifies. If that is so it would be the self that has the readier access to those images that he would experience as his real self.
This is as far as it is useful to take these rather speculative ideas at this stage. In later chapters, e.g. 12, 13, 20 and 21, they are drawn on to provide possible ways to understand certain not uncommon responses to loss.
Whenever information that would normally be accepted for further processing because of its significance to the individual is subjected to defensive exclusion for prolonged periods the consequences are far-reaching. Among them, I believe, are most, perhaps all, of the very diverse array of phenomena that at one time or another have been described in the psychoanalytic literature as being defences.9
Of the many possible consequences there are two major ones, each with certain contingent consequences, to which at this point I wish to draw attention:
(a) One or more behavioural systems within a person may be deactivated, partially or completely. When that occurs one or more other activities may come to monopolize the person’s time and attention, acting apparently as diversions.
(b) One or a set of responses a person is making may become disconnected cognitively from the interpersonal situation that is eliciting it, leaving him unaware of why he is responding as he is. When that occurs the person may do one or more of several things, each of which is likely to divert his attention away from whoever, or whatever, may be responsible for his reactions:
He may mistakenly identify some other person (or situation) as the one who (which) is eliciting his responses.
He may divert his responses away from someone who is in some degree responsible for arousing them and towards some irrelevant figure, including himself.
He may dwell so insistently on the details of his own reactions and sufferings that he has no time to consider what the interpersonal situation responsible for his reactions may really be.
A behavioural system becomes active only when the necessary combination of inflows, from exteroceptors and/or interoceptors and/or memory stores reaches it. Should such inflows be systematically excluded, it follows that the system must be immobilized, together with the thoughts and feelings to which such inflows give rise, and that it must remain so until such time as the necessary inflow is received. In traditional terms the system thus deactivated is said to be repressed. Or, put the other way about, the effects of repression are regarded as being due to certain information of significance to the individual being systematically excluded from further processing. Like repression, defensive exclusion is regarded as being at the heart of psychopathology. Only in their theoretical overtones is it necessary to make any distinction between the two concepts.
The exclusion of significant information, with the resulting deactivation of a behavioural system, may of course be less than complete. When that is so there are times when fragments of the information defensively excluded seep through so that fragments of the behaviour defensively deactivated become visible; or else feeling and other products of processing related to the behaviour reach consciousness, for example in the form of moods, memories, day dreams or night dreams, and can be reported. These psychological phenomena have given rise in traditional psychoanalytic theory to concepts such as the dynamic unconscious and the return of the repressed.
The magnitude of effect on personality functioning of a behavioural system being deactivated will clearly depend on the status of the system within the personality. Should the system be of only marginal importance, the absence of the behaviour from the person’s repertoire may be of no great consequence. Should, however, it be a behavioural system, or set of behavioural systems, as central for personality functioning as, for example, is the set controlling attachment behaviour, the effects are likely to be extensive. For, on the one hand, certain forms of behaviour, thought and feeling, will cease to occur or be experienced and, on the other, forms of behaviour, thought and feeling of some other kind will take their place. For, as Peterfreund (1971) emphasizes, within a network of control systems a major change in one part will have repercussions throughout the whole.
The diversionary role of defensive activity. Many of the patterns of behaviour, thought and feeling judged by clinicians to be defensive can be understood as alternatives to the behaviour, thought and feeling that have disappeared following deactivation. In judging them to be defensive what is usually in mind is that they give the impression, on the one hand, of being carried out under pressure and of absorbing an undue proportion of a person’s attention, time and energy, perhaps in the form of overwork, and, on the other, of being undertaken by him in some way at the expense of his giving his attention, time and energy to something else. They seem thus to be not merely alternatives but also to be playing a diversionary role; and this is probably what they do. For the more completely a person’s attention, time and energy are concentrated on one activity and on the information concerning it the more completely can information concerning another activity be excluded.
Experience suggests that there is no activity, mental or physical, that cannot be undertaken as a diversion. Whether it is work or play, of great social value or none, provided the activity is all-absorbing it meets the psychological requirement. This means that the effects of defensive activity must be judged on a number of distinct scales. For example we can ask:
The answers to these questions may differ greatly.
We are so used to regarding our thoughts, feeling and behaviour as being linked more or less directly to the circumstances in which we find ourselves that it may seem strange that the link may sometimes be missing, or the wrong link be made. Yet at a trivial level this occurs not infrequently. A man comes home from work and finds fault with his son. Subsequently he may, or may not, be aware that his irritation was aroused initially by events at work and that his son’s behaviour was of only marginal relevance. Another man wakes up feeling worried and depressed and may only subsequently identify the situation that is making him feel so. In both examples certain information relevant to his mood and behaviour is being excluded from conscious processing. When exclusion is only partial or temporary, no great harm results. When, however, exclusion is systematic and persistent ill effects may be grave.
This cognitive disconnection of a response from the interpersonal situation that elicited it I believe to play an enormous role in psychopathology. Sometimes the disconnection is complete, in which case the response may appear wholly inexplicable in terms of a psychological reaction and is consequently readily attributed to something quite different, e.g. indigestion or disturbed metabolism. At other times the disconnection is only partial, inasmuch as the person is unaware only of certain aspects of the situation whilst being well aware of other aspects. In such cases it is the intensity and persistence of the response that pose the problems.
Since in the chapters to follow, notably numbers 9 to 13 and 19 to 24, very many examples of pathological responses to loss are attributed to complete or partial disconnection of response from situation, no more need be said about the process here.
Misidentification of the interpersonal situation eliciting a response. Just as defensive activities may serve in part to ensure that attention is not given to inflow that is being defensively excluded, so may the attribution of a response to some insignificant situation serve to direct attention away from the situation truly responsible. Several examples of this are given in Volume II in the discussion of phobias (Chapters 18 and 19). A child afraid to leave home for fear his mother might desert or commit suicide during his absence claims, or is persuaded, that what he is really afraid of is being criticized by the teacher; or an adult, similarly afraid of what might happen at home during his or her absence, claims that what he or she is really afraid of is to go alone into public places.
Redirection of responses away from the person arousing them. To direct anger away from the person who elicited it and towards some more or less irrelevant person is so well known that little need be said about it. In traditional theory it is termed displacement. The term ‘splitting’ is also used in this connection when an ambivalent reaction is aroused, with the loving component being directed towards one person and the angry component redirected towards another.
Not infrequently anger is redirected away from an attachment figure who aroused it and aimed instead at the self. Inappropriate self-criticism results.
Preoccupation with personal reactions and sufferings. It not infrequently happens that when a set of responses has become disconnected from the interpersonal situation that elicited them the person focuses his attention, not on any person or situation relevant or even irrelevant to his state of mind, but solely on himself. In such cases he may dwell at length on the details of his reactions, both psychological and physiological, and especially on the extent of his sufferings. He may then be described as morbidly introspective and/or hypochondriacal. Examples of patients whose introspective preoccupations are effectively diverting their attention away from a difficult and painful situation are described by Wolff and others (1946b) and by Sacher and others (1968) and are referred to further in Chapters 9 and 13.
There are in fact many other consequences of the defensive exclusion of relevant information in addition to those noted above, including the conditions traditionally described as denial or disavowal. Since, however, they occur frequently as responses to loss and since, moreover, their classification as defences needs examination, discussion of them is left to later chapters.
At the beginning of this chapter it is pointed out that in the ordinary course of a person’s life most of the information reaching him is being routinely excluded from conscious processing in order that his capacities are not overloaded and his attention not constantly distracted; and also that where defensive exclusion differs from the usual forms of exclusion the difference lies not in the mechanisms responsible for it but in the nature of the information that is excluded. In examining the conditions that promote defensive exclusion, therefore, the focus of attention is on the nature of the information being excluded.
The theory that I believe best fits the evidence is one proposed by Peterfreund (1971), namely that the information likely to be defensively excluded is of a kind that, when accepted for processing in the past, has led the person concerned to suffer more or less severely. Whether this formula embraces all cases cannot be known until it has been tried and tested. Meanwhile I adopt it since it appears sufficient for the understanding of the responses with which this volume is concerned.
There are a number of possible reasons why incoming information of certain kinds could, if accepted, lead the person concerned to suffer. One example, long recognized in clinical literature, is when the incoming information might, if accepted, arouse feelings and/or elicit actions that would be evaluated adversely by the person’s own evaluating systems, thereby creating conflict and guilt. Another, closely related to the first, is when the incoming information might, if accepted, result in a serious conflict with parents, with the acute distress that that is likely to bring. There are two situations of that kind that are especially germane to my thesis.
The first is when a child’s attachment behaviour is strongly aroused and when, for any reason, it is not responded to and terminated. In these circumstances the child protests more or less violently and is much distressed. Should the situation recur frequently and for long periods, not only is distress prolonged but it seems that the systems controlling the behaviour ultimately become deactivated. This, the evidence indicates, is more likely to occur should lack of termination be accompanied by active rejection and, perhaps especially, when the child is punished or threatened with punishment for reacting as he is likely to do, for example by crying strongly and persistently, by demanding his mother’s presence or by being generally contrary and difficult.
The deactivation of systems mediating attachment behaviour, thought and feeling, appears to be achieved by the defensive exclusion, more or less complete, of sensory inflow of any and every kind that might activate attachment behaviour and feeling. The resulting state is one of emotional detachment which can be either partial or complete.
Deactivation of attachment behaviour is especially liable to be initiated during the early years, though it can undoubtedly be increased and consolidated during later childhood and adolescence. One reason why a young child is especially prone to react in this way is that it is during the second half of the first year of life and the subsequent two years or so that attachment behaviour is elicited most readily and continues to be so at high intensity and for long periods, leading to great suffering should no one be available to comfort him. As a result it is during these years that he is especially vulnerable to periods of separation, and also to being rejected or threatened with rejection. Another and quite different reason seems likely to be that selective exclusion occurs more readily in children than in adults. An example of this to which Hilgard (1964) draws attention is the ease with which post-hypnotic amnesia is induced in children when compared to adults.
Since there is evidence that the deactivation of attachment behaviour is a key feature of certain common variants of pathological mourning, and also of personalities prone to respond in those ways, the condition is referred to repeatedly in later chapters.
A second class of conflict with parents, and one that I believe accounts for a great many instances of defensive exclusion, arises when a child is in course of observing features of a parent’s behaviour that that parent wishes strongly he should not know about. Most of the data that can be explained by this hypothesis are well known, though the explanations adopted have usually been very different.
In therapeutic work it is not uncommon to find that a person (child, adolescent or adult) maintains, consciously, a wholly favourable image of a parent, but that at a less conscious level he nurses a contrasting image in which his parent is represented as neglectful, or rejecting, or as ill-treating him. In such persons the two images are kept apart, out of communication with each other; and any information that may be at variance with the established image is excluded.
Various views have been advanced to account for this state of affairs. One view, prominent in traditional psychoanalytic theorizing, postulates that a young child is unable to accommodate within a single image the parent’s kindly treatment of him as well as any less favourable treatment he may receive or, much emphasized by some theorists, is disposed to imagine. A second view is that a young child, being totally dependent on his parents’ care, is strongly biased to see them in a favourable light and so to exclude contrary information. A third view, to which clinicians interested in family interaction call attention, and already described, with references, in Volume II (Chapter 20), emphasizes how insistent some parents are that their children regard them in a favourable light and what pressures they put on them to comply. On threat of not being loved or even of being abandoned a child is led to understand that he is not supposed to notice his parents’ adverse treatment of him or, if he does, that he should regard it as being no more than the justifiable reaction of a wronged parent to his (the child’s) bad behaviour.
Since these explanations are not mutually exclusive, it is possible that each of the factors postulated makes some contribution. In evaluating the probable role of each, however, I believe such evidence as there is strongly favours the last, namely the role of parental pressure, and gives least support to the traditional view. Since examples of that evidence are presented in many later chapters (e.g. Chapters 12 and 18 onwards), it is unnecessary to pursue the matter further here.
Finally, let us consider a related matter, namely whether defensive exclusion originates only during earliest childhood, as has been widely assumed by psychoanalysts, or may be initiated also during later childhood and, perhaps, during adolescence and adult life as well. This is an important though difficult question since the evidence is far from clear. A principal problem lies in distinguishing between the conditions that may be necessary for initiating defensive exclusion and those capable of maintaining or increasing it.The following tentative propositions are, I believe, reasonable interpretations of the evidence:
(a) There is reason to suspect that vulnerability to conditions initiating defensive exclusion is at a maximum during the early years of life, perhaps the first three in particular (some reasons why this should be so are already referred to and others are examined in Chapter 24).
(b) Although vulnerability diminishes during later childhood and early adolescence, it probably does so only slowly and remains comparatively high throughout most of these years.
(c) There is probably no age at which human beings cease to be vulnerable to factors that maintain or increase any defensive exclusion already established.
A corollary of this position is that in examining the conditions that initiate defensive exclusion it is as necessary to consider those that may affect older children and young adolescents as it is those to which infants and very young children may be vulnerable.
In considering whether defensive exclusion is biologically adaptive the relevant criterion is whether it contributes in any way to the individual’s surviving and leaving viable offspring.11 Since this is not an easy criterion to apply, it is necessary to weigh arguments.
First, there can be little doubt that those persons in whom defensive exclusion plays a prominent part are handicapped in their dealing with other human beings when compared to those in whom it plays only a minor part. Furthermore, they are more prone to suffer breakdowns in functioning when, for periods lasting weeks, months or years, they may be unable to deal effectively with their environment. Thus, whatever the benefits of defensive exclusion may possibly be, the personality which adopts it pays a penalty, sometimes severe. The question therefore arises whether there are any circumstances in which such benefits as it may confer outweigh these undoubted penalties. This brings us back to the conditions that promote the process.
In the last section it was proposed that much of the information liable to be defensively excluded, because when accepted previously it has led to suffering, falls under two main heads: (a) information that leads a child’s attachment behaviour and feeling to be aroused intensely but to remain unassuaged, and perhaps even to be punished, and (b) information that he knows his parent(s) do not wish him to know about and would punish him for accepting as true. The question arises, therefore, whether, in the conditions that make these types of information unacceptable, the behaviour to which its exclusion leads may, at least in some cases, confer benefits that outweigh the penalties.
Let us consider each of these cases.
Main (1977) describes observations of infants aged twelve to eighteen months with their mothers and reports finding that those who fail to show attachment behaviour in circumstances in which it would be expected, e.g. after a separation lasting a few minutes in a strange setting, are highly likely to be the infants of mothers who habitually reject their advances. In the conditions described an infant of this sort, instead of showing attachment behaviour as infants of responsive mothers do, turns away from his mother and busies himself with a toy. In so doing he is effectively excluding any sensory inflow that would elicit his attachment behaviour and is thus avoiding any risk of being rebuffed and becoming distressed and disorganized; in addition he is avoiding any risk of eliciting hostile behaviour from his mother. Yet he remains in her vicinity. This type of response, Main suggests, may represent a strategy for survival alternative to seeking close proximity to mother. Its advantages are that the child avoids becoming disorganized but yet remains moderately close to and on fair terms with his mother, the chances being that, should risk of danger become high, she would then protect him.
Nevertheless, even should this suggestion prove valid, Main emphasizes, there is much evidence that the strategy is no more than second best and to be adopted only when a mother’s attitude is adverse. This is shown by the readiness with which the response is replaced by attachment behaviour once a child has become confident that his mother will respond to him kindly.
The same argument can be applied in the second case, that in which information that a child knows his parents do not wish him to know about is subjected to defensive exclusion. Here again, it is suggested, the advantages of conforming to the parent’s demands may outweigh the disadvantages. For, as children know in their bones, when mother is prone to be rejecting it may be better to placate her than to risk alienating her altogether.
If this reasoning is correct, there must none the less come a point at which the advantages of conforming to a parent’s requirements may be outweighed by the disadvantages. This is the case in those adolescents and adults who, having for long adopted the strategy of placating a parent, now find themselves unable to do anything else.
In this chapter I have tried to indicate the lines along which it may be possible to develop a theory of defence using concepts derived from recent studies of human information processing. In the chapters to follow I attempt to use these ideas to shed light on responses to loss.
1 In what follows I am indebted to Dixon’s Subliminal Perception (1971), to Norman’s introduction to human information processing (1976) and to a paper by Erdelyi (1974) on perceptual defence and vigilance.
2 The accounts given here are derived from those given in Norman (1976, pp. 31–2).
3 Evidence suggests that conscious perception may require that inflows of two different sorts, each of adequate intensity, should be received at a higher centre. One sort carries specific information and is routed via the classical afferent system. The other sort carries non-specific stimulation and is routed via the reticular activating system. Because the conduction rate through the classical system is faster than it is through the reticular system there would be time enough both for (a) the sensory inflow through the classical system to be processed for meaning outside awareness and also for (b) a message, dependent on that meaning, to be sent to the reticular system before the intensity of the non-specific stimulation to be relayed forward from the reticular system had been determined. In this way it would be possible for the intensity of the non-specific stimulation relayed forward to be so regulated that it was set either above the level required for conscious perception or below it, depending on the meaning that had been assigned to the inflow during its preliminary assessment outside awareness.
4 Information received through sense organs is believed to be held initially in a number of extremely brief stores, each linked to a single sensory mode and capable of handling large amounts of minimally processed information. Those accepting visual and auditory data have been termed by Neisser (1967) ‘iconic’ and ‘echoic’ respectively.
5 All workers recognize the formidable problem of relating the phenomenal world of consciousness to concepts of information processing. Shallice (1972) argues that the problem bears some resemblance to that of relating two neighbouring fields of science.
6 Casey (1973) has discussed the neural mechanisms mediating awareness of pain and the means by which awareness may be suppressed, as it is in conditions of great excitement, e.g. battle, strenuous sport, and under hypnosis. As in the case of visual and auditory perception, there is evidence that conscious awareness of pain requires that inflows mediated by two different systems should be received at a higher centre. One is a rapidly acting system that provides information relating to the location of a disturbance, the other, which acts more slowly, provides for the aversive and emotional components. Evidence suggests that, as in the case of visual and auditory perception, a mechanism may exist whereby neural excitation in the slower acting system can be blocked from reaching the higher centre so that the aversive and emotional components would be excluded and no pain be experienced. Even so, there would often be limited awareness that in some part of the body all is not well.
Another means for the physiological suppression of pain is suggested by the discovery of substances secreted by the pituitary and the brain (endorphins and enkephalins) which have an analgesic action comparable to opiates (Jeffcoate and others 1978).
7 Hilgard (1974) describes the procedure for inducing automatic writing: while the left hand and forearm were kept in the ice water without any discomfort, the right hand was placed in a box arranged for automatic writing or for reporting pain on a numerical scale through key pressing. The hypnotized subject was then told that ‘the hand would tell us what we ought to know, but that the subject would pay no attention to this hand and would not know what it was communicating or even that it was doing anything at all’. Procedure for automatic talking is described in detail in Knox et al. (1974).
8 In a review chapter, Tulving notes that experimental work on memory falls naturally into these two classes and concludes, therefore, that the distinction may prove to have heuristic value. He compares the two memory systems in terms of the nature of the information selected for storage, the networks within which it is stored, and the means whereby it is retrieved.
9 In the comprehensive list of defences compiled by Sperling (1958) all the following appear: disease entities, character, symptom complexes, affects, physiological states and processes, psychological states and processes, art forms, and behaviour both of social and anti-social kinds. From among that array I am confining attention to those phenomena which, together, appear to be central to the concept, namely defensive process, defensive behaviour and defensive belief. The position adopted is in all important respects the same as Peterfreund’s (1971).
10 Whereas the term ‘inactivation’ would be grammatically correct, I follow Peterfreund (1971) in using ‘deactivation’. The advantage of the latter term is that it keeps the condition distinct from that of a behavioural system which merely happens at a given moment to be inactive but which remains accessible, in the usual way, to all potentially activating inflow.
11 This, of course, is a very different criterion to those traditionally adopted by psychoanalysts which are concerned either with the distribution of psychic energy or else with the degree of mental pain experienced.