but it seems that the
-331-
inescapable facts of social living make real happiness dependent on actively friendly, mutually warm relationships with other people.
-331-
inescapable facts of social living make real happiness dependent on actively friendly, mutually warm relationships with other people.
Bowlby - Separation
Many a clinician, unfortunately, imbued with irrelevant theory and untrained in the field of family psychiatry, finds himself sadly ill equipped to see what is happening.
In consequence the family's phobic scapegoat attains the status of psychiatric diagnosis.
Not only are most clinicians untrained in these matters but all too often they show bias. Sometimes the bias is pro-children
____________________
1 Most of the research stemming from this viewpoint deals with interaction in the families of
schizophrenic patients. In this tradition are the works of Batesonet al. ( 1956), Lidzet al. ( 1958), Wynneet al. ( 1958), Laing & Esterson ( 1964), and Scott, Ashworth & Casson (
242
1970). The conclusions to which the findings of these and other studies point are, first, that the pathogenic potential of suppression and falsification as they occur within a family is fully as great as the pathogenic potential of repression and splitting as they occur within an individual, and, second, that processes of the two types interact. This is a field to which we shall return in the third volume. Well-planned research designed to explore this interaction is likely to yield insights of the greatest value to psychopathology.
-319-
and anti-parents. More often it is the other way about. Clinicians are often themselves parents, and so are likely unwittingly to identify over-readily with another parent's viewpoint. Parents may be thought of as experienced and sensible; the patients, by contrast, are young, and seen, perhaps, as inclined to exaggerate or even fabricate. In telling their stories parents may seem more lucid and coherent than their children. Furthermore, parents may be respected citizens, perhaps acquaintances or even friends whose account the clinician is reluctant to question. It may be no coincidence that Little Hans's parents were among Freud's 'closest adherents' ( SE 10: 6). Pervading the scene, moreover, and influencing all parties is the timehonoured commandment 'Honour thy father and thy mother'.
Yet another factor tipping the balance in the same direction is the tendency, as notable in clinicians as it is in laymen, to reify emotions, especially the more uncomfortable ones. Instead of describing the situation in which a person experiences fear, the person is said to 'have' a fear. Instead of describing the situation in which a person becomes angry, he is said to 'have' a bad temper. Similarly, someone 'has' a phobia, or is 'filled with' anxiety or aggression. 1 Once emotions are reified the speaker is spared the task of tracing what is making the person in question afraid or angry, and will hardly notice when family context is omitted or suppressed. Thus any clinician who thinks in these ways is all too apt to fall in with a parent's claim that the behaviour of a child is altogether baffling and unintelligible, and thence to attribute it to some psychological or physiological anomaly inherent in the child. Preoccupation with nosological entities or biochemical anomalies has the same effect. Much present-day theorizing, both psychoanalytic and non-analytic, is of these kinds.
As a result of all these influences, which, as Scott ( 1973a and b) argues, converge to form the cultural image of mental illness as it is today, the dominant bias in psychoanalysis and psychiatry is to give credence to a parent's constructions and to throw doubt on a child's. Discrepancies are attributed with great readiness to the distorting effects of a child's feelings and phantasies, and only reluctantly to the distorting effects of those of a parent.
Nevertheless in certain quarters the boot is put firmly on the other foot. By those espousing anti-psychiatry the patient is
____________________
1 The tendency to reify emotions is discussed further in Appendix III.
-320-
deemed right and well and the parent is wrong or ill. Unfortunately, so strident and condemnatory of parents have some of these claims been that a family perspective becomes discredited and valid points are lost to sight.
243
The position adopted here is that, while parents are held to play a major role in causing a child to develop a heightened susceptibility to fear, their behaviour is seen not in terms of moral condemnation but as having been determined by the experiences they themselves had as children. Once that perspective is attained and rigorously adhered to, parental behaviour that has the gravest consequences for children can be understood and treated without moral censure. That way lies hope of breaking the generational succession.
-321-
Chapter 21
Secure Attachment and the Growth of Self-reliance
People are much greater and much stronger than we imagine, and when unexpected tragedy comes . . . we see them so often grow to a stature that is far beyond anything we imagined. We must remember that people are capable of greatness, of courage, but not in isolation. . . . They need the conditions of a solidly linked human unit in which everyone is prepared to bear the burden of others.
ARCHBISHOP ANTHONY BLOOM 1
Personality development and family experience
Throughout the last half-dozen chapters attention is concentrated on conditions within a family that lead a developing child to grow up more than usually prone to be anxious and fearful. Here, in the penultimate chapter, we examine conditions that lead to an opposite and happier outcome. And just as we found that there is a strong case for believing that gnawing uncertainty about the accessibility and responsiveness of attachment figures is a principal condition for the development of unstable and anxious personality so is there a strong case for believing that an unthinking confidence in the unfailing accessibility and support of attachment figures is the bedrock on which stable and self-reliant personality is built.
Naturally any simple statement of that kind needs elaboration. Thus the family experience of those who grow up anxious and fearful is found to be characterized not only by uncertainty about parental support but often also by covert yet strongly distorting parental pressures: pressure on the child, for example, to act as caretaker for a parent; or to adopt, and thereby to confirm, a parent's false models--of self, of child, and of their relationship. Similarly, the family experience of those who grow up to become relatively stable and self-reliant
____________________
1 The David Kissen Memorial Lecture, 26 March 1969.
-322-
is characterized not only by unfailing parental support when called upon but also by a steady yet timely encouragement towards increasing autonomy, and by the frank communication by parents of working models--of themselves, of child, and of others--that are not only tolerably valid but are open to be questioned and revised.
244
Because in all these respects children tend unwittingly to identify with parents and therefore to adopt, when they become parents, the same patterns of behaviour towards children that they themselves have experienced during their own childhood, patterns of interaction are transmitted, more or less faithfully, from one generation to another. Thus the inheritance of mental health and of mental ill health through the medium of family microculture is certainly no less important, and may well be far more important, than is their inheritance through the medium of genes.
Evidence to support these propositions fully is, inevitably, insufficient. In the studies available the criteria adopted for deciding who are and who are not stable and self-reliant individuals can be challenged; the adequacy of methods used in collecting information about parental behaviour can be faulted; the assumptions made regarding continuity of personality organization over time can be questioned: and the restriction of samples to Western cultures casts doubt on how far findings can be generalized. Even so, the consistency of findings so far reported remains impressive. This means that those who are inclined to challenge either the evidence or the conclusions to which it leads have a case to answer. Only if they present such data as in their judgement point in a different direction can their objections be taken seriously.
In what follows the results of about a dozen studies, all reported since 1960, are drawn upon. It is not an exhaustive list and is, unfortunately, confined to studies undertaken in the United States. So far as is known, however, the findings are not contradicted by the results of any other study. Certainly, such knowledge of how personality development is related to family experience as has been acquired by those who work professionally with families in the United Kingdom does nothing to call in question the American findings.
Sector Studies of the Life-cycle
Since with present facilities it is not possible to study human beings in the course of their development from the cradle to the
-323-
grave, it is necessary to consider sectors of the life-cycle piecemeal. Once a sufficient number of such studies is recorded it is reasonable to hope that, by fitting their findings together as a mosaic, a picture will emerge of a range of personality patterns, each of which will be seen developing along its own typical pathway within that version of family environment that, for good or ill, tends inexorably to promote it. In this chapter a sketch map of such a mosaic is attempted.
The sector of the life-cycle most studied is that lying between ten years of age and the early twenties. Typically the sample chosen is representative either of children in certain specified schools or of students in a specified college. Whereas in most studies, e. g. those of Bronfenbrenner ( 1961), Grinker ( 1962), Rosenberg ( 1965), Coopersmith ( 1967), Megargee, Parker & Levine ( 1971), information regarding personality and family is obtained at a single point in the individual's life-history, in a few the subjects are followed during a number of years. Examples of such studies are one by Peck & Havighurst ( 1960) in which subjects were followed from the age of ten to seventeen, one by Offer ( 1969) in which they were followed from fourteen to eighteen, and one by Murpheyet al. ( 1963) in which they were followed from their last year at high school through their first year at college. Samples range in size from a few dozen to several hundred, with an occasional sample running into thousands. The
245
amount of information available on each individual differs enormously and, as might be expected, varies inversely with size of sample. Whereas most samples include both boys and girls, a few are confined to males.
The findings of the batch of studies that focus on the sector running from pre-adolescence to early adult life provide us with an invaluable vantage-point from which to look both at earlier and at later sectors of the life-cycle. Looking in one direction we can consider the results of three studies of personality development and family experience that cover, respectively, the preschool years ( Baumrind 1967; Heinickeet al. , in press) and the first year of life ( Ainsworth, Bell & Stayton 1971). Looking in the other we can consider the findings of a study of unusually effective and self-reliant men in their early and middle thirties ( Korchin & Ruff 1964). Finally, we can consider the results of a study of almost a hundred adults in their early thirties who, since early childhood, have been the subjects of a longitudinal survey ( Siegelmanet al. 1970).
Among the varied aims of these many studies one that all
-324-
have in common is to relate different degrees and forms of healthy personality organization, and/or of effective performance, to different types of experience within the family. Since in most studies interest focuses mainly on the nature of and conditions for favourable development, many of the samples are deliberately biased so that individuals who are emotionally disturbed or delinquent are under-represented or even excluded. In this way the far more common bias, typical of clinical studies, which leads a sample to be composed mainly or wholly of disturbed or delinquent subjects, is counterbalanced.
Sources of Information
In regard to personality development and its present organization and performance, information can be obtained from at least four main sources:
--from the subject himself, either during interview or in reply to questionnaires and self- rating scales
--from informants who know the subject well, notably parents, teachers, and peers
--from inferences derived from the subject's responses given either during interview or during projective testing
--from first-hand observation of behaviour either in a natural setting, e. g. at home or in
school, or in a laboratory.
Similarly, in regard to family experience, information can be obtained from at least four main sources:
--from the subject's parents, or siblings, either during the course of interview or in reply to questionnaires or self-rating scales
--from the subject himself
--from inferences derived from the parents' responses given either during interview or during projective testing
--from first-hand observation of families in interaction, either at home or in a clinical or laboratory setting.
For their information in regard to either or both of these fields, a few researchers rely on only a single source, and by so doing are able to study a large sample. A majority of researchers, however, draw on information derived from several sources but by so doing are limited to
246
studying only a small sample. That findings from these two very different types of study confirm one another gives added confidence to the findings of each.
-325-
Criteria of Evaluation
A difficulty intrinsic to every study of the kind we are concerned with is that of deciding the criteria to be used in evaluating personality structure. By what criteria, it may be asked, are we categorizing certain persons as well integrated, secure, and mentally healthy, and others as not so? How valid are these criteria? Are we, perhaps, in judging certain personality characteristics favourably doing no more than apply middle-class standards in an area in which they have no relevance? Is there danger therefore that our results are at the best of no more than limited application and at the worst positively misleading? Since criticisms of this type are often voiced (e. g. Spiegel 1958; Miller 1970; Bronfenbrenner 1970), an answer is necessary.
In the first place the criteria used are far from uniform. In some studies the principal criterion is competent performance in the social setting of home, or school, or college, or work. Examples are those by Bronfenbrenner ( 1961) of high-school students rated by their teachers, by Megargeeet al. ( 1971) of college students rated by the researchers on the basis of information given by the students themselves, and by Korchin & Ruff ( 1964) of astronauts in training. In other studies the principal criterion is the subject's self-esteem, measured mainly in terms of how he says he feels about himself in relation to others. Examples are that by Coopersmith ( 1967) of schoolboys aged ten to twelve, and that by Rosenberg ( 1965) of high- school children aged sixteen to eighteen. In other studies, for example of college students by Grinker ( 1962), the criteria applied are complex and derive from psychiatric experience. In several studies, moreover, including those of Grinker ( 1962), Peck & Havighurst ( 1960), and Offer ( 1969), criteria of several sorts are used together. In the multiplicity of criteria used by the different researchers lies some safeguard against unwitting prejudice.
A second reason for having confidence in the criteria is that in several studies evidence is given that the criteria used for healthy development correlate negatively with independent measures of mental ill health. For example, Rosenberg ( 1965) shows that his measure of self- esteem is negatively correlated with a tendency to depression, with a tendency to feel isolated and lonely, and with proneness to psychosomatic symptoms. Similarly, a rather similar measure of self-esteem used by Coopersmith ( 1967) is shown to be negatively correlated with anxiety, as measured by clinical tests, and also with emotional
-326-
problems and destructive behaviour, as reported by the sub. ject's mother.
A third reason for having confidence in the criteria is that, when they are applied to a sample of subjects, the grading of personalities that results is only weakly correlated with the social class from which the subjects come, e. g. Peck & Havighurst ( 1960), Rosenberg ( 1965), Coopersmith ( 1967). This means that, because certain values in respect of personality and family relationships have come to be associated especially with the middle classes, it is mistaken to assume that they are not held also by members of the working classes, though admittedly by a rather smaller proportion of them. Nor, conversely, can it be assumed, as
247
seems often to be done, that these so-called middle-class values are unrelated to mental health. On the contrary, it is very plausible to suppose that certain, though not all, of the psychosocial values and practices of a family that make for a modest degree of educational, social, and economic success in a child are the same as some of those that make for his better than average mental health. The plausibility of this view is much strengthened when it is expressed in its complementary form, namely that certain of the psychosocial values and practices of a family that make for below average mental health in a child are the same as some of those that make for his educational, social, and economic failure. Indeed, those studying the causes of intractable poverty no less than those studying the causes of mental ill health find themselves confronted by certain adverse and self-perpetuating patterns of family microculture that there is reason to believe are causal agents common to both conditions.
These are complex and difficult questions some of which are referred to again later in the chapter. Meanwhile enough has been said to show why in what follows the objection that the findings presented are invalidated because suffused with unwitting middle-class prejudice is not accepted.
All the criteria used in these studies, it is believed, are closely related to each other and all are measures, albeit crude, of a characteristic that might be termed 'adaptability'. By this is meant the capacity to adapt successfully to, and therefore to survive for long periods in, any and all of a wide range of physical and social environments, especially when survival turns on cooperation with others. Although this capacity could in principle be subjected to empirical test, in practice it would be far from easy to do. To illustrate the concept, however, an
-327-
imaginary experiment can be described. In it the experimenter would select several groups of individuals, unfamiliar with each other, and transport each group to a succession of strange and difficult environments--some strange and difficult for reasons of social structure and custom, and others so because of geographical features. The prediction would then be that a group of individuals rated highly on a measure of adaptability would be more likely to succeed and survive over a long period in each one of these environments than would a group of individuals matched in other relevant respects, but rated lower on adaptability.
Thus the criterion of adaptability is distinct from a criterion such as 'adjustment to the status quo', to the use of which in this context there would be strong objection. It is distinct also from the criterion of whether a person tends to accept, to criticize, or to reject the status quo. Indeed, the ways in which personalities rated highly on the criterion of adaptability may contribute, positively or negatively, to the political life of the societies in which they live are little known; and to elucidate them is a task for which psychiatrists are not qualified.
It is thus clearly recognized that the interrelated criteria with which this chapter is concerned are a few only of the many that are applicable to personality. Some of the others, for example, degree of originality, of creative spirit, or of capacity for innovation, are certainly distinct from criteria of mental health and adaptability, and may, perhaps, be correlated with them in only slight degree. It must therefore be strongly emphasized that, in concentrating on the one set of criteria to the exclusion of others, no claim is made that the criteria selected are the only ones of importance. The reason for so concentrating is that in the practice of psychiatry the issues that must be our first concern are those of mental health and ill health. In so far as in
248
our actions we may apply other criteria we are doing so simply as adherents of a professional ethic or as private persons.
The reader interested in considering problems of criteria further is referred to discussions by Grinker ( 1962), Heath ( 1965), and Douvan & Adelson ( 1966), and to a comprehensive review by Offer & Sabshin ( 1966).
Studies of adolescents and young adults The Peck & Havighurst Study
Because clinicians are traditionally sceptical of the results of large samples studied by what they believe to be inadequate
-328-
methods, we start with an extremely detailed and careful study of thirty-four children, seventeen boys and seventeen girls, growing up in a small town of the American mid-west, code-named Prairie City. This study by Peck & Havighurst, published in 1960, is part of a more extensive study begun during the 1940s of social and psychological life in the town. When selected for study the town had a population of about 10,000, 90 per cent of whom were native born and mainly of Norwegian and Polish extraction. The men were engaged either in agriculture or in local industry. Areas of residence were little segregated by social class and there was no socially disorganized area.
The sample of children studied was a sub-sample of all those born in the town during 1933. All children in the cohort, which numbered 120, were first examined in 1943 when they were ten years old. At that time they were given a number of tests of intelligence and personality and were also rated in regard to personality characteristics both by their teachers and by their peers. As a result of this preliminary screening a sub-sample of thirty-four children was chosen as representative (a) of all ranges of moral character and (b) of the social-class structure of the town. Thenceforward the development of these thirtyfour children and of the families in which they lived became the subject of intensive study until 1950 when the children had all turned seventeen.
Since both the criteria used, namely 'moral character' and 'social class', can, as we have seen, give rise to controversy, a word about the place that each holds in this study is necessary.
Although in selecting the sub-sample Peck & Havighurst used a criterion defined in terms of moral character, a reading of their case material makes it plain that there is a high correlation between judgements based on that criterion and judgements based on the degree to which an individual is a wellorganized personality, capable of effective performance in fields both of work and of human relationships, and in good standing with peers. In effect, therefore, the scale used is almost equivalent to scales that might be designed to measure, say, 'integration of personality', or 'ego strength', or 'emotional security', or 'mental health', or adaptability as defined here. 1
____________________
1 Fairly early in their study Peck & Havighurst, in fact, replace the criterion 'moral character'
by that of 'maturity of character'. Reasons why in the present work the latter concept is not employed are already mentioned briefly at the end of Chapter 14, and they are elaborated
249
further in the final one. -329-
As regards the issue of class, it is of advantage that in this study, in contrast to many others, the sample selected is roughly representative of the whole Prairie City population and, as such, comes mainly from the lower half of the socio-economic scale. This is shown in the table below. The criticism that findings are misleading because tangled with middle-class values would, therefore, be of little relevance in this context.
Sample studied
Boys Girls No. No.
All Population
Children Of city % All ages
%
Socio-economic Class
Upper Upper-middle Lower-middle Upper-lower Lower-lower
On each of the thirty-four children in the study a great quantity of data were amassed. Many data came from the child himself, for example, from interviews with him, from standardized tests and questionnaires, and from projection tests. Other data came from sociometric measures given to the whole cohort of 120 children and from teachers' ratings. Data were analysed and evaluated in several steps. First, data from each source were analysed separately. Next, a clinical conference was called in which data from all sources were drawn upon to arrive at a picture of personality structure. A third step was for each research worker in the project to rate each personality on a series of scales aimed at measuring different aspects of character structure; as a result each child became designated by a personality profile. Finally, on the basis of these profiles the children were grouped into what proved empirically to be eight categories arranged according to their degree of maturity, a dimension equivalent to what in this work is termed adaptability. Brief descriptions of these eight character types are quoted below, starting with the 'least mature', and indicating the number of children assigned to each.
-330-
I. The amoral: These five children were characterized by inaccurate perception of social situations, of other people, and of self; poor ability to set clear, realistic, attainable goals of any kind, behavior which is ill-adapted to achieve whatever ends the person does have in
0 0
1 0
4 5
9 10
3 2
1717 100100
0 3 3 11 26 31 56 41 15 14
250
mind; and poor control over impulses which will interfere with successful adaptation to the social world, even in the sense of achieving purely personal, selfish gratification.
They show hostile, immature emotionality. There is, moreover, a pattern of childishly inappropriate emotional lability which mobilizes excessive energy and imposes a severe strain on the individual's already weak self-control. The usual nature of these emotions is that of negativism and hostility. These subjects are unwilling to accept the self-restraints and positive precepts their society suggests. . . .
They suffer from punitive but ineffectual guilt feelings, which are of little use in controlling their behavior. This in itself indicates sharp inner conflict and lack of positive, healthy self- regard or self-respect. . . .
The consequence is that they are no more at peace with themselves than they are with the world, though they might defiantly deny it to any representative of the culture they so strongly reject.
II. A type intermediate between the amoral and the expedient: Three children (not described in detail).
III. The expedient: These four children were characterized by 'taking the easy way out':
Their almost exclusive expediency is . . . not so much an active attempt to manipulate the people and events around them for personal gain as it is an effort to get as much personal gratification as possible by fitting in with their world when they have to, and avoiding as many social demands as possible which would require them to act in a positively socialized way.
Their constriction and relinquishment of direction to the social forces around them leads only to the absence of active immorality . . . . it requires them to suppress enough of their selfish spontaneous impulses to make them tense, restless, and uncomfortable with themselves. . . .
They are would-be hedonists . . .
but it seems that the
-331-
inescapable facts of social living make real happiness dependent on actively friendly, mutually warm relationships with other people. Since they have little conception of such relatedness, their efforts to grasp hedonistic pleasure result in empty satisfactions. They seek but do not find, for they are unable to recognize the human warmth and approval they vaguely but intensely want.
IV. The impulsive yet guilt-ridden: 1 These two boys were characterized by a 'primitive, harsh conscience' which they disowned.
Thus, they are not 'masters in their own houses'. They react to impulse or to internalized, irrationally held moral principles in which they do not personally believe . . . They don't care much for other people, and feel themselves to be quite bad. They are not conscious of much of their guilt, for they protect themselves against recognizing their basically low self-regard by consciously picturing themselves in more favourable but not very realistic terms. Even so,
251
their inner conflict is too severe to be successfully ignored. They try hard, but don't find much real pleasure in life.
V. The conforming: These eight children were characterized by a good deal of hostility that was controlled fairly effectively by a punitive conscience.
Two of them, both girls, are described as being unable to express their wishes spontaneously and to derive little satisfaction from life.
They feel strong, chronic guilt about their 'bad impulses', even though they seldom actively express them. Their superegos are almost entirely composed of negative 'Don'ts' which they have incorporated unquestioningly. They feel themselves to be a bad kind of person, just as they see little to like in others. They are unable to check the punitive voices of their conscience (almost a direct echo of harsh parental strictures, it seems) against the reality of daily life, in any rational, self-directing way. They are, in short, depressed, dull, unhappy, and quite unable to stand up to the world even to express their antagonism toward it.
Some of the other children categorized as 'conforming' are described as being 'friendly in their outlook and relatively at
____________________
1 Peck & Havighurst designate this personality type by letters only, and the title given here is
provided by the present author. -332-
ease with themselves'. They were thought, however, to lack inner direction and to conform rather passively to the demands of those around them.
VI. The irrational-conscientious: These three children are described as
walking examples of the 'Puritan conscience'. . . . they have an appreciable degree of generalized hostility. This produces some guilt, but it is not intense since they are so utterly guided by their superego directives. They automatically behave in responsible, 'loyal', honest, 'kind' ways; but it is more by rote than by personal intent. They demand as much of others as of themselves in the way of conventional morality.
Nevertheless, their lack of any strong, positive concern for others as individuals, not to mention their repressed but definite hostility, makes them far too literal minded and rigid in their righteousness to be very easy to live with. . . .
They take some cold satisfaction from rigorously observing the letter of the law. That is about the extent of their joy in living . . . Their peers respect them, but they don't like them.
VII. A type with good integration but less so than in type VIII: These five children show 'a high degree of rationality, friendliness, and altruistic impulse . . . high autonomy and good integration of most major drives'. They are 'thoroughly spontaqeous', like other people, and are liked in return. Although by no means lacking in moral principles, they are judged as rather too apt to put their own enjoyment first to be placed in category VIII. Even so, they are often very considerate of others.
252
VIII. The rational-altruistic: Finally, there were four children who are described as 'well- integrated' and 'emotionally mature' and who possessed 'firm, internalized moral principles' which they applied in an insightful manner.
They enjoy life thoroughly and actively, having as healthy a respect for themselves as they do for other people. There is no false pride in this. They simply are well and accurately aware of their own natures and capacities. Since they are free from serious conflict, free from any irrational need to follow convention blindly for the sake of 'security', they are free to use almost all their emotional energy.
-333-
These judgements of the children made by the researchers in terms of 'moral development', and later equated with level of 'maturity', were closely matched by independent judgements of the children made by their peers in terms of altogether more humdrum criteria. Thus it is found that all nine of the children placed in the highest two categories, by level of maturity are almost uniformly well thought of by their peers in terms of their being friendly and cheerful, good participants in joint enterprises, and capable of self-control and leadership. By contrast, the eight children placed in the lowest two categories by level of maturity are uniformly poorly thought of by their peers. The seventeen children placed in the middle four categories by the researchers occupy intermediate positions also in the eyes of peers. The only disagreements centre on three children, of whom one was categorized as 'expedient' and two were categorized as 'conformers' by the researchers, but who were rated more highly by their peers. It is an open issue which group of judges was the more discerning.
Critical readers will doubtless disagree with the raters in respect of some issues and some children; for example, objection can be brought to rating moral development (or 'maturity' of development) along a single dimension. Nevertheless, a majority of readers will recognize that the overall estimates of the children's 'maturity' conform fairly closely to estimates of their mental health that most clinicians would be likely to reach. Moreover, the substantive findings of the study in which personality development is related to family experience do not turn on details of method.
To turn now to the families from which these children come: by a stroke of good fortune detailed information on the families had been collected when the children were aged between thirteen and fourteen by an independent research group working in the Prairie City study and had been stored away untouched. That made it possible to examine how personality structure is, or is not, related to patterns of family interaction by comparing two sets of data gathered entirely independently of one another. The results show significant correlations in the directions our theory would lead us to predict.
Each of the thirty-four families had been rated by the independent team on a number of scales which, when subjected to factor analysis, yielded four dimensions of family interaction. Two of these dimensions were found to be positively correlated with each other and to be strongly correlated also with the
-334-
253
ratings of the children in terms of their level of maturity, as agreed by the main team. 1 These two family dimensions and the components that contribute to them are as follows:Mutual trust and approval between the child and his parents:
-- parents accept the child as he is and give him much affection and praise
-- parents trust their child's judgement and do not insist on close supervision
-- child feels free to discuss issues with his parents
-- parents encourage the child to make friends and also make his friends welcome -- relations between parents are congenial and compatible.
Consistency of family life:
-- regularity of daily routine
-- predictability, in nature and timing, of parents' methods of control -- frequent participation by family members in shared activities.
When we consider the family experience of the nine children rated most highly in terms of maturity we find that the families of all but one were rated highly on both these dimensions. Conversely, of the eight children rated lowest in terms of maturity the families of all but one were rated low on both dimensions.
The family patterns found to be characteristic of children in each of the five main groups are described below.
Families of the amoral children: 'The most striking feature of these families is that, without exception, they are markedly inconsistent; and [except for one child] they are highly mistrustful and disapproving of their children. These boys and girls have grown up knowing very little love, little emotional security, and little if any consistent discipline. ' It is no surprise, therefore, to learn that a child of this type is found to express 'an active hate for his family and for almost anyone else'.
Families of the expedient children: These children come from 'a laissez-faire home, where the parents give the child indiscriminate freedom to make his own decisions, approve of him,
____________________
1 The other two dimensions of family intervention, namely 'democracy versus autocracy' and
'severity of parental discipline', were not significantly related to the level of maturity of the children's characters.
-335-
and are lenient in their discipline, but also are inconsistent . . . ' Although these children receive 'a good deal of general parental support . . . it is combined with inconsistency, irregularity, and leniency, [and so] does not contain much real recognition or concern for the child as an individual'.
Again, it is no surprise to learn that a child of this type is found to have little feeling for his parents and to be ready to reject them whenever it suits his purpose.
Families of the conforming children: Most of these children come from severely autocratic punitive homes in which there is also much mistrust. When the parents are also inconsistent the resulting character structure of the child comes close to that of the 'amoral' children. When
254
there is higher consistency and less distrust it comes closer to the 'irrational-conscientious' type.
Families of the irrational-conscientious children: These three children, all girls, had parents who were either severe or very severe in their discipline. In no case was mutual trust in the family rated high and in one it was rated very low. Consistency varied from average to high.
Families of the rational-altruistic children: Features that distinguish the parents of children in this most highly rated group are that they are strongly approving of their child and of his activities and friends, that they engage in many activities with him and have a harmonious relationship with each other. Home routine is regular without being rigidly so. Parents trust their child. In matters of discipline they are consistent in what they demand but 'leniency prevails over severity'. The children are found to have strong positive feelings towards both parents, sentiments that they later extend to others. Standards of behaviour, never having been enforced harshly, are open to discussion; they can then be applied later in ways adapted to the special features of a situation.
In the relationship between 'maturity' of character and family experience one case proves a notable exception to what is found otherwise to be an almost perfect correlation. This is the case of a boy rated very highly by the researchers in terms of his maturity, and also highly regarded by his peers, but whose family had been rated very poorly on the various family scales. When the home was first visited it was described as 'a physically unkempt working- class home, in which little regularity or consistency was seen by the interviewer'. Nevertheless, it is of
-336-
interest that a few years later another visitor noted that the boy's 'relationships with his family and his extended kin are mutually acceptant and supportive' although, it was thought, lacking in warm affection. A possible explanation of this apparently anomalous case, an explanation which Peck & Havighurst tend to adopt, is that the original worker had been over-impressed by the obvious material untidiness of the home and too little aware of the less evident but far more relevant strong ties that existed within the family and the mutual support that members gave each other.
Since these thirty-four adolescents and the homes from which they came were studied and observed over a period of seven years, it was possible for the researchers to gauge how much, or little, change occurred during that time in the personalities of the children and in their families. In the event what struck the researchers was the very high degree of consistency that was apparent in the development of both of the parties. Thus '. . . the ratings and the actual case histories both suggest that whatever pattern of moral behaviour and character structure a child shows at ten years of age, he is far more likely than not to display into late adolescence'. Moreover, in so far as data were available also on earlier development, they were found to be of a piece with later development. Similarly, it was found that 'the parents tended to be just as consistently what they were, through the years, as did their children -- particularly in their relationship with a given child'.
This consistency of development over seven years of early and late adolescence is of importance to our thesis for two reasons. First, it lends credence to the research strategy of building up pictures of personality structures as they develop during the whole life-cycle by
255
fitting together, in a mosaic, findings from studies of different sectors of it. Second, it supports the view, discussed further in the final chapter, that different adult personality types are better accounted for in terms of development having taken place along one or other of a number of distinct and divergent developmental pathways than in terms of development having become fixated at one or another of a set of points thought of as occurring at intervals along a single pathway.
Studies of Large Representative Samples
In the much briefer presentation in this section of some findings from the many other studies available on adolescents and young
-337-
adults, emphasis is placed on the regularity with which the findings reported are either similar to or compatible with those of Peck & Havighurst, despite the fact that each of these other researchers studied a differently structured sample, and used a different criterion of character development and also different indices of the pattern of family life.
Because Peck & Havighurst studied so small a sample there is advantage in proceeding next to studies which, because they draw on large representative samples, are able to examine rather different aspects of family life. In considering the findings of these large-sample studies, however, it must be remembered that in most of them information about the families comes entirely from the subjects themselves and must therefore be treated with caution.
In two of the large-sample studies a clear relationship emerges between patterns of personality development and certain basic features of the homes from which the subjects come.
One such study is that by Rosenberg ( 1965) whose sample consisted of no fewer than 5,024 boys and girls; they were aged from sixteen to eighteen years and were attending ten public high schools in New York State, selected to ensure that communities of every sort were represented. The criterion of personality used was a measure of self-esteem, which is best described as a measure of how a person feels towards himself, and especially of how he feels he compares with other people. This Rosenberg measured by means of a checklist of ten questions, each of which was to be answered on a five-point scale, ranging from 'strongly agree' to 'strongly disagree'.
The checklist for the assessment of self-esteem was given as part of a much larger questionnaire. One part inquired about a teenager's family and the other part about his view of himself, his feelings, and any psychosomatic symptoms to which he might be prone. The questionnaire was presented by teachers and completed during school time. From the information available two types of correlation are possible: (a) correlations of a subject's self- esteem with other statements he might make about himself; (b) correlations of a subject's self- esteem with the structure of his family.
As regards correlations of the first type, Rosenberg found that low self-esteem correlates significantly with several measures related to potential psychiatric disability, for example feelings of loneliness, sensitivity to criticism, anxiety, depression, and psychosomatic symptoms. By contrast, high self-esteem is
256
-338-
correlated with trust in other people, active social participation, and a likelihood of being chosen as leader. As regards correlations of the second type, Rosenberg found that, in level of self-esteem, children of divorced parents tend to compare unfavourably with children living in intact families. These lowered levels of self-esteem occur mainly in children of mothers who married young, had children soon after marriage, and were divorced before their twenty- fourth birthday. In a similar way the children of women who married and were widowed young also show a tendency to lowered self-esteem. By contrast, these ill effects are not seen in the children of women who were older when their child was born and when they lost their husband, whether by death or by divorce. Rosenberg explains his findings by postulating, very plausibly, that early divorce or widowhood leaves a mother of young children in a difficult and vulnerable position, which often results in her feeling insecure, anxious, and irritable, which in turn affects the personality development of her child. Another contingency, not mentioned by Rosenberg, is that the young children of young single-handed mothers are very apt to be subjected to periods of unstable substitute care. In another study with a large sample, comprising 488 university students (280 men and 208 women) of a mean age of nineteen years, Megargee, Parker & Levine ( 1971) report that a systematic relationship is found between a measure of socialization and the state of the parents' marriage. The measure of socialization used, the California Personality Inventory Socialization Scale, is claimed to be a well-validated and standardized instrument that permits the selection of male and female groups characterized as either superior or inferior in regard to socialization by reference to national norms. On this scale groups of disturbed and delinquent adolescents yield low scores. When the 488 students are divided into four groups according to their scores, it is found that the gradient of scores correlates positively with the following features of family life:
-- living with both natural parents
-- parents' marriage rated by student as excellent
-- student's childhood rated by him as having been happy.
By contrast, the gradient correlates negatively with parents being divorced.
In the following table, results are given only for students placed in the highest- and the lowest-scoring of the four groups.
-339-
In every case the findings for the two intermediate groups lie on the gradient between the extremes. When findings for each sex are considered separately, no differences of consequence emerge. Results are expressed as percentages of the students in each group who report that they come from families with the characteristics shown.
% high- scoring group N = 51
Living with both natural parents
of% of low-
scoring
group
N = 110 Family experience
95 78
257
% high- scoring group N = 51
Rating parents' marriage as excellent Rating own childhood as happy Parents are divorced
of% of low-
scoring
group
N = 110 Family experience
85 29 85 42 219
In this study no correlation was found between death of a parent and socialization score. Since only about 7 per cent of the whole sample had lost a parent, it is possible that the absence of correlation is in part because a smaller proportion of bereaved adolescents than of non- bereaved ones had reached college.
A third large-sample study is reported by Bronfenbrenner ( 1961). His aim was to investigate the family background of sixteen-year-old boys and girls rated by teachers in regard to each of two criteria: (a) the extent to which they proved to be leaders or followers at school, and (b) the extent to which they could or could not be relied on to fulfil obligations. Information regarding their families came from a questionnaire, designed to measure twenty different aspects of parent-child relations, which was completed by the subjects themselves.
The sample studied numbered 192, made up of equal numbers of boys and girls and also of equal numbers from each of four socio-economic classes, determined in a rough-and-ready way by the amount of education father is reported to have received.
Results are given separately for boys and girls and for each of the two criteria. Boys tend to be rated higher on leadership than do girls and the reverse is true of responsibility. On each criterion children whose fathers are more educated tend to be rated
-340-
higher than children whose fathers are less educated. Other principal findings are that an adolescent who shows leadership is likely to come from a home in which he is given much time, affection, and support from his parents; and that one who shows a sense of responsibility is likely to come from a home in which parents exercise a good deal of authority, usually by means of reason and reward rather than punishment. Leadership and responsibility in children and affection and authority in homes are all positively correlated with one another.
At the higher ends of the rating-scales in respect of both criteria certain differences in family experience were found between boys and girls. Whereas boys seemed to thrive on high levels of parental support and control, there seemed to be some danger of girls receiving an overdose of one or both from their parents.
At the lower ends of the rating-scales, by contrast, no differences of consequence in family background were found between boys and girls. Moreover, whether the adolescent was rated low on leadership or on responsibility the picture of the home that emerged was much the
258
same: in either case parental indifference or rejection was the rule. The boy (or girl) concerned was likely to describe his parents as inclined to complain about him, to ridicule him and compare him unfavourably with other children, and to spend little time with him and perhaps to avoid his company. Discipline was likely either to be lacking or else to be administered by means of arbitrary and excessive punishment. In respect of a few children whose leadership was rated low, however, a very different picture emerged: so far from being neglected, they had parents who were markedly over-protective of them.
In a fourth study, which draws on a fairly large sample and is reported by Coopersmith ( 1967), information about the family was obtained first-hand, although only from mother. The sample was confined to boys from intact white families.
Coopersmith's sample comprised eighty-five boys, aged from ten to twelve years, who were attending schools in two middlesized towns in New England. The socio-economic classes from which most came were neither high nor low. The sample, which was drawn from a much larger number of children initially assessed, was stratified according to two criteria: (a) the boys' self-assessment on a test designed to measure selfesteem, and (b) the teachers' assessments of the boys in terms of their behaviour. As in the Rosenberg study, low self- esteem was
-341-
found to be strongly correlated with anxiety as measured by clinical tests; it was also, though less strongly, correlated with emotional problems as reported by mother.
Information regarding the boys' families came from: (i) a questionnaire completed by mother, (ii) a two-and-a-half-hour interview with mother by an interviewer uninformed regarding the boy's rating on self-esteem, and (iii) the boy's answers to a series of questions on his parents' attitudes and practices. Fathers were not seen.
In reviewing his findings Coopersmith stresses, above all, the high level of maternal acceptance found in the families of boys with high self-esteem: 'The findings are all consistent, regardless of the instrument or source of information. They reveal that the mothers of children with high self-esteem are more loving and have closer relationships with their children than do mothers of children with low self-esteem. ' Furthermore, as regards the strongly contrasting forms of discipline that were used by the parents of boys in the high and low self-esteem groups respectively, Coopersmith's findings are remarkably similar to those of Peck & Havighurst and to those of Bronfenbrenner, although the criteria of favourable personality development were quite different in the three studies. In the Coopersmith study not only were the boys of high self-esteem expected by their parents to meet high standards but parental control was exercised with care, respect, and firmness, and by the use of reward rather than punishment. By contrast, it was found that the boys of low self-esteem were not only given little care or guidance by their parents but often subjected to harsh and disrespectful punishment, which included loss of love.
Personality Development, Modes of Discipline, and Social Class
The consistency with which differences are reported in the modes of discipline and care to which children who show favourable or unfavourable development respectively are subjected is very striking. Equally striking is the consistency with which some of these same differences
259
are reported to be associated with social class. Thus it is found that less-educated and working-class parents are more likely to use severe and arbitrary punishment, and to ignore or reject a child, than are better-educated and middle-class parents; and working-class fathers are less likely to spend time in joint activities with their adolescent children than are middle-class fathers (see review by Bronfenbrenner 1958). Taken together, these findings regarding
-342-
(a) modes of discipline and personality development and (b) modes of discipline and social class support the hypothesis proposed earlier that the positive, if weak, correlation found between healthy personality development and higher social class may be explained, in part, by the differences in the ways in which parents belonging to different social classes tend to treat their children. The findings of Bronfenbrenner empirical study ( 1961) can be taken to illustrate a set of correlations that appear to be typical:
-- low ratings for leadership and responsibility in children are associated with parents who show little interest in their children and who either adopt arbitrary and punitive methods of discipline or else give them little guidance;
-- arbitrary methods of discipline, including physical punishment and ridicule, are more likely to be used by less-educated parents than by better-educated ones;
-- the children of less-educated parents are likely to be rated lower on leadership and responsibility than are those of better-educated parents.
Rosenberg reports a similar set of correlations between level of self-esteem, the amount of attention and concern fathers give their children, and social class. Further evidence compatible with the hypothesis is already given in Chapter 15 in which the relationship between symptoms of anxiety in a child and parental threats to abandon him or to commit suicide is discussed. These threats, it is found, are used by a larger proportion of parents in the working and lower-middle classes than of parents in the higher social classes. It would be inappropriate to pursue these complex and sensitive matters further in this work. Another large and difficult area which again it is not proposed to pursue here is the differential influence of father and mother on the development of their children, with special reference to the influence of each on boys and on girls. Those interested are referred to the study by Bronfenbrenner ( 1961) and to one by Douvan & Adelson ( 1966) who discuss in much detail the difference in developmental patterns shown by boys and girls between the ages of twelve and eighteen.
Further Studies of Small Samples
Next we revert to more intensive studies by considering the findings of three projects in each of which small samples of men
-343-
or youths, selected especially for their apparently healthy and well-integrated development, were subjected to intensive clinical examination and observation over a period of at least a year. Presented in descending order of the subjects' age, the first study is of astronauts in training, the second of youths attending college, and the third of high-school students bound for college.
260
In respect both of the developmental pathways that these personalities are following and of the family life they have experienced or are still experiencing, the findings of these three studies are in agreement; they are in agreement also with those of Peck & Havighurst. First, these well-adapted personalities are found to show a smoothly working balance of, on the one hand, initiative and self-reliance and, on the other, a capacity both to seek help and to make use of help when occasion demands. Second, an examination of their development shows that they have grown up in closely knit families with parents who, it seems, have never failed to provide them with support and encouragement.
So far as it goes, each study gives the same picture, that of a stable family base from which first the child, then the adolescent, and finally the young adult moves out in a series of ever- lengthening excursions. While autonomy is evidently encouraged in such families, it is not forced. Each step follows the previous one in a series of easy stages. Though home ties may attenuate, they are never broken.
Astronauts rank high as self-reliant men capable of living and working effectively in conditions of great potential danger and stress. Their performance, personalities, and histories have been studied by Korchin & Ruff. In two articles ( Korchin & Ruff 1964; Ruff & Korchin 1967) they publish preliminary findings on a small sample of seven men.
Although these men tend to be individualists who show a high degree of self-reliance and a clear preference for independent action, all are reported to be 'comfortable when dependence on others is required' and to have a 'capacity to maintain trust, in what might seem conditions of distrust'. The performance of the crew of Apollo 13, which met with a mishap en route to the moon, is testimony to their capacity to sustain trust. Not only did they maintain their own efficiency in conditions of great danger but they continued to cooperate trustingly and effectively with their companions at the base on earth.
Turning to their life-histories we find that they:
-344-
grew up in relatively small well-organized communities, with considerable family solidarity and strong identification with the father.
Not only are most clinicians untrained in these matters but all too often they show bias. Sometimes the bias is pro-children
____________________
1 Most of the research stemming from this viewpoint deals with interaction in the families of
schizophrenic patients. In this tradition are the works of Batesonet al. ( 1956), Lidzet al. ( 1958), Wynneet al. ( 1958), Laing & Esterson ( 1964), and Scott, Ashworth & Casson (
242
1970). The conclusions to which the findings of these and other studies point are, first, that the pathogenic potential of suppression and falsification as they occur within a family is fully as great as the pathogenic potential of repression and splitting as they occur within an individual, and, second, that processes of the two types interact. This is a field to which we shall return in the third volume. Well-planned research designed to explore this interaction is likely to yield insights of the greatest value to psychopathology.
-319-
and anti-parents. More often it is the other way about. Clinicians are often themselves parents, and so are likely unwittingly to identify over-readily with another parent's viewpoint. Parents may be thought of as experienced and sensible; the patients, by contrast, are young, and seen, perhaps, as inclined to exaggerate or even fabricate. In telling their stories parents may seem more lucid and coherent than their children. Furthermore, parents may be respected citizens, perhaps acquaintances or even friends whose account the clinician is reluctant to question. It may be no coincidence that Little Hans's parents were among Freud's 'closest adherents' ( SE 10: 6). Pervading the scene, moreover, and influencing all parties is the timehonoured commandment 'Honour thy father and thy mother'.
Yet another factor tipping the balance in the same direction is the tendency, as notable in clinicians as it is in laymen, to reify emotions, especially the more uncomfortable ones. Instead of describing the situation in which a person experiences fear, the person is said to 'have' a fear. Instead of describing the situation in which a person becomes angry, he is said to 'have' a bad temper. Similarly, someone 'has' a phobia, or is 'filled with' anxiety or aggression. 1 Once emotions are reified the speaker is spared the task of tracing what is making the person in question afraid or angry, and will hardly notice when family context is omitted or suppressed. Thus any clinician who thinks in these ways is all too apt to fall in with a parent's claim that the behaviour of a child is altogether baffling and unintelligible, and thence to attribute it to some psychological or physiological anomaly inherent in the child. Preoccupation with nosological entities or biochemical anomalies has the same effect. Much present-day theorizing, both psychoanalytic and non-analytic, is of these kinds.
As a result of all these influences, which, as Scott ( 1973a and b) argues, converge to form the cultural image of mental illness as it is today, the dominant bias in psychoanalysis and psychiatry is to give credence to a parent's constructions and to throw doubt on a child's. Discrepancies are attributed with great readiness to the distorting effects of a child's feelings and phantasies, and only reluctantly to the distorting effects of those of a parent.
Nevertheless in certain quarters the boot is put firmly on the other foot. By those espousing anti-psychiatry the patient is
____________________
1 The tendency to reify emotions is discussed further in Appendix III.
-320-
deemed right and well and the parent is wrong or ill. Unfortunately, so strident and condemnatory of parents have some of these claims been that a family perspective becomes discredited and valid points are lost to sight.
243
The position adopted here is that, while parents are held to play a major role in causing a child to develop a heightened susceptibility to fear, their behaviour is seen not in terms of moral condemnation but as having been determined by the experiences they themselves had as children. Once that perspective is attained and rigorously adhered to, parental behaviour that has the gravest consequences for children can be understood and treated without moral censure. That way lies hope of breaking the generational succession.
-321-
Chapter 21
Secure Attachment and the Growth of Self-reliance
People are much greater and much stronger than we imagine, and when unexpected tragedy comes . . . we see them so often grow to a stature that is far beyond anything we imagined. We must remember that people are capable of greatness, of courage, but not in isolation. . . . They need the conditions of a solidly linked human unit in which everyone is prepared to bear the burden of others.
ARCHBISHOP ANTHONY BLOOM 1
Personality development and family experience
Throughout the last half-dozen chapters attention is concentrated on conditions within a family that lead a developing child to grow up more than usually prone to be anxious and fearful. Here, in the penultimate chapter, we examine conditions that lead to an opposite and happier outcome. And just as we found that there is a strong case for believing that gnawing uncertainty about the accessibility and responsiveness of attachment figures is a principal condition for the development of unstable and anxious personality so is there a strong case for believing that an unthinking confidence in the unfailing accessibility and support of attachment figures is the bedrock on which stable and self-reliant personality is built.
Naturally any simple statement of that kind needs elaboration. Thus the family experience of those who grow up anxious and fearful is found to be characterized not only by uncertainty about parental support but often also by covert yet strongly distorting parental pressures: pressure on the child, for example, to act as caretaker for a parent; or to adopt, and thereby to confirm, a parent's false models--of self, of child, and of their relationship. Similarly, the family experience of those who grow up to become relatively stable and self-reliant
____________________
1 The David Kissen Memorial Lecture, 26 March 1969.
-322-
is characterized not only by unfailing parental support when called upon but also by a steady yet timely encouragement towards increasing autonomy, and by the frank communication by parents of working models--of themselves, of child, and of others--that are not only tolerably valid but are open to be questioned and revised.
244
Because in all these respects children tend unwittingly to identify with parents and therefore to adopt, when they become parents, the same patterns of behaviour towards children that they themselves have experienced during their own childhood, patterns of interaction are transmitted, more or less faithfully, from one generation to another. Thus the inheritance of mental health and of mental ill health through the medium of family microculture is certainly no less important, and may well be far more important, than is their inheritance through the medium of genes.
Evidence to support these propositions fully is, inevitably, insufficient. In the studies available the criteria adopted for deciding who are and who are not stable and self-reliant individuals can be challenged; the adequacy of methods used in collecting information about parental behaviour can be faulted; the assumptions made regarding continuity of personality organization over time can be questioned: and the restriction of samples to Western cultures casts doubt on how far findings can be generalized. Even so, the consistency of findings so far reported remains impressive. This means that those who are inclined to challenge either the evidence or the conclusions to which it leads have a case to answer. Only if they present such data as in their judgement point in a different direction can their objections be taken seriously.
In what follows the results of about a dozen studies, all reported since 1960, are drawn upon. It is not an exhaustive list and is, unfortunately, confined to studies undertaken in the United States. So far as is known, however, the findings are not contradicted by the results of any other study. Certainly, such knowledge of how personality development is related to family experience as has been acquired by those who work professionally with families in the United Kingdom does nothing to call in question the American findings.
Sector Studies of the Life-cycle
Since with present facilities it is not possible to study human beings in the course of their development from the cradle to the
-323-
grave, it is necessary to consider sectors of the life-cycle piecemeal. Once a sufficient number of such studies is recorded it is reasonable to hope that, by fitting their findings together as a mosaic, a picture will emerge of a range of personality patterns, each of which will be seen developing along its own typical pathway within that version of family environment that, for good or ill, tends inexorably to promote it. In this chapter a sketch map of such a mosaic is attempted.
The sector of the life-cycle most studied is that lying between ten years of age and the early twenties. Typically the sample chosen is representative either of children in certain specified schools or of students in a specified college. Whereas in most studies, e. g. those of Bronfenbrenner ( 1961), Grinker ( 1962), Rosenberg ( 1965), Coopersmith ( 1967), Megargee, Parker & Levine ( 1971), information regarding personality and family is obtained at a single point in the individual's life-history, in a few the subjects are followed during a number of years. Examples of such studies are one by Peck & Havighurst ( 1960) in which subjects were followed from the age of ten to seventeen, one by Offer ( 1969) in which they were followed from fourteen to eighteen, and one by Murpheyet al. ( 1963) in which they were followed from their last year at high school through their first year at college. Samples range in size from a few dozen to several hundred, with an occasional sample running into thousands. The
245
amount of information available on each individual differs enormously and, as might be expected, varies inversely with size of sample. Whereas most samples include both boys and girls, a few are confined to males.
The findings of the batch of studies that focus on the sector running from pre-adolescence to early adult life provide us with an invaluable vantage-point from which to look both at earlier and at later sectors of the life-cycle. Looking in one direction we can consider the results of three studies of personality development and family experience that cover, respectively, the preschool years ( Baumrind 1967; Heinickeet al. , in press) and the first year of life ( Ainsworth, Bell & Stayton 1971). Looking in the other we can consider the findings of a study of unusually effective and self-reliant men in their early and middle thirties ( Korchin & Ruff 1964). Finally, we can consider the results of a study of almost a hundred adults in their early thirties who, since early childhood, have been the subjects of a longitudinal survey ( Siegelmanet al. 1970).
Among the varied aims of these many studies one that all
-324-
have in common is to relate different degrees and forms of healthy personality organization, and/or of effective performance, to different types of experience within the family. Since in most studies interest focuses mainly on the nature of and conditions for favourable development, many of the samples are deliberately biased so that individuals who are emotionally disturbed or delinquent are under-represented or even excluded. In this way the far more common bias, typical of clinical studies, which leads a sample to be composed mainly or wholly of disturbed or delinquent subjects, is counterbalanced.
Sources of Information
In regard to personality development and its present organization and performance, information can be obtained from at least four main sources:
--from the subject himself, either during interview or in reply to questionnaires and self- rating scales
--from informants who know the subject well, notably parents, teachers, and peers
--from inferences derived from the subject's responses given either during interview or during projective testing
--from first-hand observation of behaviour either in a natural setting, e. g. at home or in
school, or in a laboratory.
Similarly, in regard to family experience, information can be obtained from at least four main sources:
--from the subject's parents, or siblings, either during the course of interview or in reply to questionnaires or self-rating scales
--from the subject himself
--from inferences derived from the parents' responses given either during interview or during projective testing
--from first-hand observation of families in interaction, either at home or in a clinical or laboratory setting.
For their information in regard to either or both of these fields, a few researchers rely on only a single source, and by so doing are able to study a large sample. A majority of researchers, however, draw on information derived from several sources but by so doing are limited to
246
studying only a small sample. That findings from these two very different types of study confirm one another gives added confidence to the findings of each.
-325-
Criteria of Evaluation
A difficulty intrinsic to every study of the kind we are concerned with is that of deciding the criteria to be used in evaluating personality structure. By what criteria, it may be asked, are we categorizing certain persons as well integrated, secure, and mentally healthy, and others as not so? How valid are these criteria? Are we, perhaps, in judging certain personality characteristics favourably doing no more than apply middle-class standards in an area in which they have no relevance? Is there danger therefore that our results are at the best of no more than limited application and at the worst positively misleading? Since criticisms of this type are often voiced (e. g. Spiegel 1958; Miller 1970; Bronfenbrenner 1970), an answer is necessary.
In the first place the criteria used are far from uniform. In some studies the principal criterion is competent performance in the social setting of home, or school, or college, or work. Examples are those by Bronfenbrenner ( 1961) of high-school students rated by their teachers, by Megargeeet al. ( 1971) of college students rated by the researchers on the basis of information given by the students themselves, and by Korchin & Ruff ( 1964) of astronauts in training. In other studies the principal criterion is the subject's self-esteem, measured mainly in terms of how he says he feels about himself in relation to others. Examples are that by Coopersmith ( 1967) of schoolboys aged ten to twelve, and that by Rosenberg ( 1965) of high- school children aged sixteen to eighteen. In other studies, for example of college students by Grinker ( 1962), the criteria applied are complex and derive from psychiatric experience. In several studies, moreover, including those of Grinker ( 1962), Peck & Havighurst ( 1960), and Offer ( 1969), criteria of several sorts are used together. In the multiplicity of criteria used by the different researchers lies some safeguard against unwitting prejudice.
A second reason for having confidence in the criteria is that in several studies evidence is given that the criteria used for healthy development correlate negatively with independent measures of mental ill health. For example, Rosenberg ( 1965) shows that his measure of self- esteem is negatively correlated with a tendency to depression, with a tendency to feel isolated and lonely, and with proneness to psychosomatic symptoms. Similarly, a rather similar measure of self-esteem used by Coopersmith ( 1967) is shown to be negatively correlated with anxiety, as measured by clinical tests, and also with emotional
-326-
problems and destructive behaviour, as reported by the sub. ject's mother.
A third reason for having confidence in the criteria is that, when they are applied to a sample of subjects, the grading of personalities that results is only weakly correlated with the social class from which the subjects come, e. g. Peck & Havighurst ( 1960), Rosenberg ( 1965), Coopersmith ( 1967). This means that, because certain values in respect of personality and family relationships have come to be associated especially with the middle classes, it is mistaken to assume that they are not held also by members of the working classes, though admittedly by a rather smaller proportion of them. Nor, conversely, can it be assumed, as
247
seems often to be done, that these so-called middle-class values are unrelated to mental health. On the contrary, it is very plausible to suppose that certain, though not all, of the psychosocial values and practices of a family that make for a modest degree of educational, social, and economic success in a child are the same as some of those that make for his better than average mental health. The plausibility of this view is much strengthened when it is expressed in its complementary form, namely that certain of the psychosocial values and practices of a family that make for below average mental health in a child are the same as some of those that make for his educational, social, and economic failure. Indeed, those studying the causes of intractable poverty no less than those studying the causes of mental ill health find themselves confronted by certain adverse and self-perpetuating patterns of family microculture that there is reason to believe are causal agents common to both conditions.
These are complex and difficult questions some of which are referred to again later in the chapter. Meanwhile enough has been said to show why in what follows the objection that the findings presented are invalidated because suffused with unwitting middle-class prejudice is not accepted.
All the criteria used in these studies, it is believed, are closely related to each other and all are measures, albeit crude, of a characteristic that might be termed 'adaptability'. By this is meant the capacity to adapt successfully to, and therefore to survive for long periods in, any and all of a wide range of physical and social environments, especially when survival turns on cooperation with others. Although this capacity could in principle be subjected to empirical test, in practice it would be far from easy to do. To illustrate the concept, however, an
-327-
imaginary experiment can be described. In it the experimenter would select several groups of individuals, unfamiliar with each other, and transport each group to a succession of strange and difficult environments--some strange and difficult for reasons of social structure and custom, and others so because of geographical features. The prediction would then be that a group of individuals rated highly on a measure of adaptability would be more likely to succeed and survive over a long period in each one of these environments than would a group of individuals matched in other relevant respects, but rated lower on adaptability.
Thus the criterion of adaptability is distinct from a criterion such as 'adjustment to the status quo', to the use of which in this context there would be strong objection. It is distinct also from the criterion of whether a person tends to accept, to criticize, or to reject the status quo. Indeed, the ways in which personalities rated highly on the criterion of adaptability may contribute, positively or negatively, to the political life of the societies in which they live are little known; and to elucidate them is a task for which psychiatrists are not qualified.
It is thus clearly recognized that the interrelated criteria with which this chapter is concerned are a few only of the many that are applicable to personality. Some of the others, for example, degree of originality, of creative spirit, or of capacity for innovation, are certainly distinct from criteria of mental health and adaptability, and may, perhaps, be correlated with them in only slight degree. It must therefore be strongly emphasized that, in concentrating on the one set of criteria to the exclusion of others, no claim is made that the criteria selected are the only ones of importance. The reason for so concentrating is that in the practice of psychiatry the issues that must be our first concern are those of mental health and ill health. In so far as in
248
our actions we may apply other criteria we are doing so simply as adherents of a professional ethic or as private persons.
The reader interested in considering problems of criteria further is referred to discussions by Grinker ( 1962), Heath ( 1965), and Douvan & Adelson ( 1966), and to a comprehensive review by Offer & Sabshin ( 1966).
Studies of adolescents and young adults The Peck & Havighurst Study
Because clinicians are traditionally sceptical of the results of large samples studied by what they believe to be inadequate
-328-
methods, we start with an extremely detailed and careful study of thirty-four children, seventeen boys and seventeen girls, growing up in a small town of the American mid-west, code-named Prairie City. This study by Peck & Havighurst, published in 1960, is part of a more extensive study begun during the 1940s of social and psychological life in the town. When selected for study the town had a population of about 10,000, 90 per cent of whom were native born and mainly of Norwegian and Polish extraction. The men were engaged either in agriculture or in local industry. Areas of residence were little segregated by social class and there was no socially disorganized area.
The sample of children studied was a sub-sample of all those born in the town during 1933. All children in the cohort, which numbered 120, were first examined in 1943 when they were ten years old. At that time they were given a number of tests of intelligence and personality and were also rated in regard to personality characteristics both by their teachers and by their peers. As a result of this preliminary screening a sub-sample of thirty-four children was chosen as representative (a) of all ranges of moral character and (b) of the social-class structure of the town. Thenceforward the development of these thirtyfour children and of the families in which they lived became the subject of intensive study until 1950 when the children had all turned seventeen.
Since both the criteria used, namely 'moral character' and 'social class', can, as we have seen, give rise to controversy, a word about the place that each holds in this study is necessary.
Although in selecting the sub-sample Peck & Havighurst used a criterion defined in terms of moral character, a reading of their case material makes it plain that there is a high correlation between judgements based on that criterion and judgements based on the degree to which an individual is a wellorganized personality, capable of effective performance in fields both of work and of human relationships, and in good standing with peers. In effect, therefore, the scale used is almost equivalent to scales that might be designed to measure, say, 'integration of personality', or 'ego strength', or 'emotional security', or 'mental health', or adaptability as defined here. 1
____________________
1 Fairly early in their study Peck & Havighurst, in fact, replace the criterion 'moral character'
by that of 'maturity of character'. Reasons why in the present work the latter concept is not employed are already mentioned briefly at the end of Chapter 14, and they are elaborated
249
further in the final one. -329-
As regards the issue of class, it is of advantage that in this study, in contrast to many others, the sample selected is roughly representative of the whole Prairie City population and, as such, comes mainly from the lower half of the socio-economic scale. This is shown in the table below. The criticism that findings are misleading because tangled with middle-class values would, therefore, be of little relevance in this context.
Sample studied
Boys Girls No. No.
All Population
Children Of city % All ages
%
Socio-economic Class
Upper Upper-middle Lower-middle Upper-lower Lower-lower
On each of the thirty-four children in the study a great quantity of data were amassed. Many data came from the child himself, for example, from interviews with him, from standardized tests and questionnaires, and from projection tests. Other data came from sociometric measures given to the whole cohort of 120 children and from teachers' ratings. Data were analysed and evaluated in several steps. First, data from each source were analysed separately. Next, a clinical conference was called in which data from all sources were drawn upon to arrive at a picture of personality structure. A third step was for each research worker in the project to rate each personality on a series of scales aimed at measuring different aspects of character structure; as a result each child became designated by a personality profile. Finally, on the basis of these profiles the children were grouped into what proved empirically to be eight categories arranged according to their degree of maturity, a dimension equivalent to what in this work is termed adaptability. Brief descriptions of these eight character types are quoted below, starting with the 'least mature', and indicating the number of children assigned to each.
-330-
I. The amoral: These five children were characterized by inaccurate perception of social situations, of other people, and of self; poor ability to set clear, realistic, attainable goals of any kind, behavior which is ill-adapted to achieve whatever ends the person does have in
0 0
1 0
4 5
9 10
3 2
1717 100100
0 3 3 11 26 31 56 41 15 14
250
mind; and poor control over impulses which will interfere with successful adaptation to the social world, even in the sense of achieving purely personal, selfish gratification.
They show hostile, immature emotionality. There is, moreover, a pattern of childishly inappropriate emotional lability which mobilizes excessive energy and imposes a severe strain on the individual's already weak self-control. The usual nature of these emotions is that of negativism and hostility. These subjects are unwilling to accept the self-restraints and positive precepts their society suggests. . . .
They suffer from punitive but ineffectual guilt feelings, which are of little use in controlling their behavior. This in itself indicates sharp inner conflict and lack of positive, healthy self- regard or self-respect. . . .
The consequence is that they are no more at peace with themselves than they are with the world, though they might defiantly deny it to any representative of the culture they so strongly reject.
II. A type intermediate between the amoral and the expedient: Three children (not described in detail).
III. The expedient: These four children were characterized by 'taking the easy way out':
Their almost exclusive expediency is . . . not so much an active attempt to manipulate the people and events around them for personal gain as it is an effort to get as much personal gratification as possible by fitting in with their world when they have to, and avoiding as many social demands as possible which would require them to act in a positively socialized way.
Their constriction and relinquishment of direction to the social forces around them leads only to the absence of active immorality . . . . it requires them to suppress enough of their selfish spontaneous impulses to make them tense, restless, and uncomfortable with themselves. . . .
They are would-be hedonists . . .
but it seems that the
-331-
inescapable facts of social living make real happiness dependent on actively friendly, mutually warm relationships with other people. Since they have little conception of such relatedness, their efforts to grasp hedonistic pleasure result in empty satisfactions. They seek but do not find, for they are unable to recognize the human warmth and approval they vaguely but intensely want.
IV. The impulsive yet guilt-ridden: 1 These two boys were characterized by a 'primitive, harsh conscience' which they disowned.
Thus, they are not 'masters in their own houses'. They react to impulse or to internalized, irrationally held moral principles in which they do not personally believe . . . They don't care much for other people, and feel themselves to be quite bad. They are not conscious of much of their guilt, for they protect themselves against recognizing their basically low self-regard by consciously picturing themselves in more favourable but not very realistic terms. Even so,
251
their inner conflict is too severe to be successfully ignored. They try hard, but don't find much real pleasure in life.
V. The conforming: These eight children were characterized by a good deal of hostility that was controlled fairly effectively by a punitive conscience.
Two of them, both girls, are described as being unable to express their wishes spontaneously and to derive little satisfaction from life.
They feel strong, chronic guilt about their 'bad impulses', even though they seldom actively express them. Their superegos are almost entirely composed of negative 'Don'ts' which they have incorporated unquestioningly. They feel themselves to be a bad kind of person, just as they see little to like in others. They are unable to check the punitive voices of their conscience (almost a direct echo of harsh parental strictures, it seems) against the reality of daily life, in any rational, self-directing way. They are, in short, depressed, dull, unhappy, and quite unable to stand up to the world even to express their antagonism toward it.
Some of the other children categorized as 'conforming' are described as being 'friendly in their outlook and relatively at
____________________
1 Peck & Havighurst designate this personality type by letters only, and the title given here is
provided by the present author. -332-
ease with themselves'. They were thought, however, to lack inner direction and to conform rather passively to the demands of those around them.
VI. The irrational-conscientious: These three children are described as
walking examples of the 'Puritan conscience'. . . . they have an appreciable degree of generalized hostility. This produces some guilt, but it is not intense since they are so utterly guided by their superego directives. They automatically behave in responsible, 'loyal', honest, 'kind' ways; but it is more by rote than by personal intent. They demand as much of others as of themselves in the way of conventional morality.
Nevertheless, their lack of any strong, positive concern for others as individuals, not to mention their repressed but definite hostility, makes them far too literal minded and rigid in their righteousness to be very easy to live with. . . .
They take some cold satisfaction from rigorously observing the letter of the law. That is about the extent of their joy in living . . . Their peers respect them, but they don't like them.
VII. A type with good integration but less so than in type VIII: These five children show 'a high degree of rationality, friendliness, and altruistic impulse . . . high autonomy and good integration of most major drives'. They are 'thoroughly spontaqeous', like other people, and are liked in return. Although by no means lacking in moral principles, they are judged as rather too apt to put their own enjoyment first to be placed in category VIII. Even so, they are often very considerate of others.
252
VIII. The rational-altruistic: Finally, there were four children who are described as 'well- integrated' and 'emotionally mature' and who possessed 'firm, internalized moral principles' which they applied in an insightful manner.
They enjoy life thoroughly and actively, having as healthy a respect for themselves as they do for other people. There is no false pride in this. They simply are well and accurately aware of their own natures and capacities. Since they are free from serious conflict, free from any irrational need to follow convention blindly for the sake of 'security', they are free to use almost all their emotional energy.
-333-
These judgements of the children made by the researchers in terms of 'moral development', and later equated with level of 'maturity', were closely matched by independent judgements of the children made by their peers in terms of altogether more humdrum criteria. Thus it is found that all nine of the children placed in the highest two categories, by level of maturity are almost uniformly well thought of by their peers in terms of their being friendly and cheerful, good participants in joint enterprises, and capable of self-control and leadership. By contrast, the eight children placed in the lowest two categories by level of maturity are uniformly poorly thought of by their peers. The seventeen children placed in the middle four categories by the researchers occupy intermediate positions also in the eyes of peers. The only disagreements centre on three children, of whom one was categorized as 'expedient' and two were categorized as 'conformers' by the researchers, but who were rated more highly by their peers. It is an open issue which group of judges was the more discerning.
Critical readers will doubtless disagree with the raters in respect of some issues and some children; for example, objection can be brought to rating moral development (or 'maturity' of development) along a single dimension. Nevertheless, a majority of readers will recognize that the overall estimates of the children's 'maturity' conform fairly closely to estimates of their mental health that most clinicians would be likely to reach. Moreover, the substantive findings of the study in which personality development is related to family experience do not turn on details of method.
To turn now to the families from which these children come: by a stroke of good fortune detailed information on the families had been collected when the children were aged between thirteen and fourteen by an independent research group working in the Prairie City study and had been stored away untouched. That made it possible to examine how personality structure is, or is not, related to patterns of family interaction by comparing two sets of data gathered entirely independently of one another. The results show significant correlations in the directions our theory would lead us to predict.
Each of the thirty-four families had been rated by the independent team on a number of scales which, when subjected to factor analysis, yielded four dimensions of family interaction. Two of these dimensions were found to be positively correlated with each other and to be strongly correlated also with the
-334-
253
ratings of the children in terms of their level of maturity, as agreed by the main team. 1 These two family dimensions and the components that contribute to them are as follows:Mutual trust and approval between the child and his parents:
-- parents accept the child as he is and give him much affection and praise
-- parents trust their child's judgement and do not insist on close supervision
-- child feels free to discuss issues with his parents
-- parents encourage the child to make friends and also make his friends welcome -- relations between parents are congenial and compatible.
Consistency of family life:
-- regularity of daily routine
-- predictability, in nature and timing, of parents' methods of control -- frequent participation by family members in shared activities.
When we consider the family experience of the nine children rated most highly in terms of maturity we find that the families of all but one were rated highly on both these dimensions. Conversely, of the eight children rated lowest in terms of maturity the families of all but one were rated low on both dimensions.
The family patterns found to be characteristic of children in each of the five main groups are described below.
Families of the amoral children: 'The most striking feature of these families is that, without exception, they are markedly inconsistent; and [except for one child] they are highly mistrustful and disapproving of their children. These boys and girls have grown up knowing very little love, little emotional security, and little if any consistent discipline. ' It is no surprise, therefore, to learn that a child of this type is found to express 'an active hate for his family and for almost anyone else'.
Families of the expedient children: These children come from 'a laissez-faire home, where the parents give the child indiscriminate freedom to make his own decisions, approve of him,
____________________
1 The other two dimensions of family intervention, namely 'democracy versus autocracy' and
'severity of parental discipline', were not significantly related to the level of maturity of the children's characters.
-335-
and are lenient in their discipline, but also are inconsistent . . . ' Although these children receive 'a good deal of general parental support . . . it is combined with inconsistency, irregularity, and leniency, [and so] does not contain much real recognition or concern for the child as an individual'.
Again, it is no surprise to learn that a child of this type is found to have little feeling for his parents and to be ready to reject them whenever it suits his purpose.
Families of the conforming children: Most of these children come from severely autocratic punitive homes in which there is also much mistrust. When the parents are also inconsistent the resulting character structure of the child comes close to that of the 'amoral' children. When
254
there is higher consistency and less distrust it comes closer to the 'irrational-conscientious' type.
Families of the irrational-conscientious children: These three children, all girls, had parents who were either severe or very severe in their discipline. In no case was mutual trust in the family rated high and in one it was rated very low. Consistency varied from average to high.
Families of the rational-altruistic children: Features that distinguish the parents of children in this most highly rated group are that they are strongly approving of their child and of his activities and friends, that they engage in many activities with him and have a harmonious relationship with each other. Home routine is regular without being rigidly so. Parents trust their child. In matters of discipline they are consistent in what they demand but 'leniency prevails over severity'. The children are found to have strong positive feelings towards both parents, sentiments that they later extend to others. Standards of behaviour, never having been enforced harshly, are open to discussion; they can then be applied later in ways adapted to the special features of a situation.
In the relationship between 'maturity' of character and family experience one case proves a notable exception to what is found otherwise to be an almost perfect correlation. This is the case of a boy rated very highly by the researchers in terms of his maturity, and also highly regarded by his peers, but whose family had been rated very poorly on the various family scales. When the home was first visited it was described as 'a physically unkempt working- class home, in which little regularity or consistency was seen by the interviewer'. Nevertheless, it is of
-336-
interest that a few years later another visitor noted that the boy's 'relationships with his family and his extended kin are mutually acceptant and supportive' although, it was thought, lacking in warm affection. A possible explanation of this apparently anomalous case, an explanation which Peck & Havighurst tend to adopt, is that the original worker had been over-impressed by the obvious material untidiness of the home and too little aware of the less evident but far more relevant strong ties that existed within the family and the mutual support that members gave each other.
Since these thirty-four adolescents and the homes from which they came were studied and observed over a period of seven years, it was possible for the researchers to gauge how much, or little, change occurred during that time in the personalities of the children and in their families. In the event what struck the researchers was the very high degree of consistency that was apparent in the development of both of the parties. Thus '. . . the ratings and the actual case histories both suggest that whatever pattern of moral behaviour and character structure a child shows at ten years of age, he is far more likely than not to display into late adolescence'. Moreover, in so far as data were available also on earlier development, they were found to be of a piece with later development. Similarly, it was found that 'the parents tended to be just as consistently what they were, through the years, as did their children -- particularly in their relationship with a given child'.
This consistency of development over seven years of early and late adolescence is of importance to our thesis for two reasons. First, it lends credence to the research strategy of building up pictures of personality structures as they develop during the whole life-cycle by
255
fitting together, in a mosaic, findings from studies of different sectors of it. Second, it supports the view, discussed further in the final chapter, that different adult personality types are better accounted for in terms of development having taken place along one or other of a number of distinct and divergent developmental pathways than in terms of development having become fixated at one or another of a set of points thought of as occurring at intervals along a single pathway.
Studies of Large Representative Samples
In the much briefer presentation in this section of some findings from the many other studies available on adolescents and young
-337-
adults, emphasis is placed on the regularity with which the findings reported are either similar to or compatible with those of Peck & Havighurst, despite the fact that each of these other researchers studied a differently structured sample, and used a different criterion of character development and also different indices of the pattern of family life.
Because Peck & Havighurst studied so small a sample there is advantage in proceeding next to studies which, because they draw on large representative samples, are able to examine rather different aspects of family life. In considering the findings of these large-sample studies, however, it must be remembered that in most of them information about the families comes entirely from the subjects themselves and must therefore be treated with caution.
In two of the large-sample studies a clear relationship emerges between patterns of personality development and certain basic features of the homes from which the subjects come.
One such study is that by Rosenberg ( 1965) whose sample consisted of no fewer than 5,024 boys and girls; they were aged from sixteen to eighteen years and were attending ten public high schools in New York State, selected to ensure that communities of every sort were represented. The criterion of personality used was a measure of self-esteem, which is best described as a measure of how a person feels towards himself, and especially of how he feels he compares with other people. This Rosenberg measured by means of a checklist of ten questions, each of which was to be answered on a five-point scale, ranging from 'strongly agree' to 'strongly disagree'.
The checklist for the assessment of self-esteem was given as part of a much larger questionnaire. One part inquired about a teenager's family and the other part about his view of himself, his feelings, and any psychosomatic symptoms to which he might be prone. The questionnaire was presented by teachers and completed during school time. From the information available two types of correlation are possible: (a) correlations of a subject's self- esteem with other statements he might make about himself; (b) correlations of a subject's self- esteem with the structure of his family.
As regards correlations of the first type, Rosenberg found that low self-esteem correlates significantly with several measures related to potential psychiatric disability, for example feelings of loneliness, sensitivity to criticism, anxiety, depression, and psychosomatic symptoms. By contrast, high self-esteem is
256
-338-
correlated with trust in other people, active social participation, and a likelihood of being chosen as leader. As regards correlations of the second type, Rosenberg found that, in level of self-esteem, children of divorced parents tend to compare unfavourably with children living in intact families. These lowered levels of self-esteem occur mainly in children of mothers who married young, had children soon after marriage, and were divorced before their twenty- fourth birthday. In a similar way the children of women who married and were widowed young also show a tendency to lowered self-esteem. By contrast, these ill effects are not seen in the children of women who were older when their child was born and when they lost their husband, whether by death or by divorce. Rosenberg explains his findings by postulating, very plausibly, that early divorce or widowhood leaves a mother of young children in a difficult and vulnerable position, which often results in her feeling insecure, anxious, and irritable, which in turn affects the personality development of her child. Another contingency, not mentioned by Rosenberg, is that the young children of young single-handed mothers are very apt to be subjected to periods of unstable substitute care. In another study with a large sample, comprising 488 university students (280 men and 208 women) of a mean age of nineteen years, Megargee, Parker & Levine ( 1971) report that a systematic relationship is found between a measure of socialization and the state of the parents' marriage. The measure of socialization used, the California Personality Inventory Socialization Scale, is claimed to be a well-validated and standardized instrument that permits the selection of male and female groups characterized as either superior or inferior in regard to socialization by reference to national norms. On this scale groups of disturbed and delinquent adolescents yield low scores. When the 488 students are divided into four groups according to their scores, it is found that the gradient of scores correlates positively with the following features of family life:
-- living with both natural parents
-- parents' marriage rated by student as excellent
-- student's childhood rated by him as having been happy.
By contrast, the gradient correlates negatively with parents being divorced.
In the following table, results are given only for students placed in the highest- and the lowest-scoring of the four groups.
-339-
In every case the findings for the two intermediate groups lie on the gradient between the extremes. When findings for each sex are considered separately, no differences of consequence emerge. Results are expressed as percentages of the students in each group who report that they come from families with the characteristics shown.
% high- scoring group N = 51
Living with both natural parents
of% of low-
scoring
group
N = 110 Family experience
95 78
257
% high- scoring group N = 51
Rating parents' marriage as excellent Rating own childhood as happy Parents are divorced
of% of low-
scoring
group
N = 110 Family experience
85 29 85 42 219
In this study no correlation was found between death of a parent and socialization score. Since only about 7 per cent of the whole sample had lost a parent, it is possible that the absence of correlation is in part because a smaller proportion of bereaved adolescents than of non- bereaved ones had reached college.
A third large-sample study is reported by Bronfenbrenner ( 1961). His aim was to investigate the family background of sixteen-year-old boys and girls rated by teachers in regard to each of two criteria: (a) the extent to which they proved to be leaders or followers at school, and (b) the extent to which they could or could not be relied on to fulfil obligations. Information regarding their families came from a questionnaire, designed to measure twenty different aspects of parent-child relations, which was completed by the subjects themselves.
The sample studied numbered 192, made up of equal numbers of boys and girls and also of equal numbers from each of four socio-economic classes, determined in a rough-and-ready way by the amount of education father is reported to have received.
Results are given separately for boys and girls and for each of the two criteria. Boys tend to be rated higher on leadership than do girls and the reverse is true of responsibility. On each criterion children whose fathers are more educated tend to be rated
-340-
higher than children whose fathers are less educated. Other principal findings are that an adolescent who shows leadership is likely to come from a home in which he is given much time, affection, and support from his parents; and that one who shows a sense of responsibility is likely to come from a home in which parents exercise a good deal of authority, usually by means of reason and reward rather than punishment. Leadership and responsibility in children and affection and authority in homes are all positively correlated with one another.
At the higher ends of the rating-scales in respect of both criteria certain differences in family experience were found between boys and girls. Whereas boys seemed to thrive on high levels of parental support and control, there seemed to be some danger of girls receiving an overdose of one or both from their parents.
At the lower ends of the rating-scales, by contrast, no differences of consequence in family background were found between boys and girls. Moreover, whether the adolescent was rated low on leadership or on responsibility the picture of the home that emerged was much the
258
same: in either case parental indifference or rejection was the rule. The boy (or girl) concerned was likely to describe his parents as inclined to complain about him, to ridicule him and compare him unfavourably with other children, and to spend little time with him and perhaps to avoid his company. Discipline was likely either to be lacking or else to be administered by means of arbitrary and excessive punishment. In respect of a few children whose leadership was rated low, however, a very different picture emerged: so far from being neglected, they had parents who were markedly over-protective of them.
In a fourth study, which draws on a fairly large sample and is reported by Coopersmith ( 1967), information about the family was obtained first-hand, although only from mother. The sample was confined to boys from intact white families.
Coopersmith's sample comprised eighty-five boys, aged from ten to twelve years, who were attending schools in two middlesized towns in New England. The socio-economic classes from which most came were neither high nor low. The sample, which was drawn from a much larger number of children initially assessed, was stratified according to two criteria: (a) the boys' self-assessment on a test designed to measure selfesteem, and (b) the teachers' assessments of the boys in terms of their behaviour. As in the Rosenberg study, low self- esteem was
-341-
found to be strongly correlated with anxiety as measured by clinical tests; it was also, though less strongly, correlated with emotional problems as reported by mother.
Information regarding the boys' families came from: (i) a questionnaire completed by mother, (ii) a two-and-a-half-hour interview with mother by an interviewer uninformed regarding the boy's rating on self-esteem, and (iii) the boy's answers to a series of questions on his parents' attitudes and practices. Fathers were not seen.
In reviewing his findings Coopersmith stresses, above all, the high level of maternal acceptance found in the families of boys with high self-esteem: 'The findings are all consistent, regardless of the instrument or source of information. They reveal that the mothers of children with high self-esteem are more loving and have closer relationships with their children than do mothers of children with low self-esteem. ' Furthermore, as regards the strongly contrasting forms of discipline that were used by the parents of boys in the high and low self-esteem groups respectively, Coopersmith's findings are remarkably similar to those of Peck & Havighurst and to those of Bronfenbrenner, although the criteria of favourable personality development were quite different in the three studies. In the Coopersmith study not only were the boys of high self-esteem expected by their parents to meet high standards but parental control was exercised with care, respect, and firmness, and by the use of reward rather than punishment. By contrast, it was found that the boys of low self-esteem were not only given little care or guidance by their parents but often subjected to harsh and disrespectful punishment, which included loss of love.
Personality Development, Modes of Discipline, and Social Class
The consistency with which differences are reported in the modes of discipline and care to which children who show favourable or unfavourable development respectively are subjected is very striking. Equally striking is the consistency with which some of these same differences
259
are reported to be associated with social class. Thus it is found that less-educated and working-class parents are more likely to use severe and arbitrary punishment, and to ignore or reject a child, than are better-educated and middle-class parents; and working-class fathers are less likely to spend time in joint activities with their adolescent children than are middle-class fathers (see review by Bronfenbrenner 1958). Taken together, these findings regarding
-342-
(a) modes of discipline and personality development and (b) modes of discipline and social class support the hypothesis proposed earlier that the positive, if weak, correlation found between healthy personality development and higher social class may be explained, in part, by the differences in the ways in which parents belonging to different social classes tend to treat their children. The findings of Bronfenbrenner empirical study ( 1961) can be taken to illustrate a set of correlations that appear to be typical:
-- low ratings for leadership and responsibility in children are associated with parents who show little interest in their children and who either adopt arbitrary and punitive methods of discipline or else give them little guidance;
-- arbitrary methods of discipline, including physical punishment and ridicule, are more likely to be used by less-educated parents than by better-educated ones;
-- the children of less-educated parents are likely to be rated lower on leadership and responsibility than are those of better-educated parents.
Rosenberg reports a similar set of correlations between level of self-esteem, the amount of attention and concern fathers give their children, and social class. Further evidence compatible with the hypothesis is already given in Chapter 15 in which the relationship between symptoms of anxiety in a child and parental threats to abandon him or to commit suicide is discussed. These threats, it is found, are used by a larger proportion of parents in the working and lower-middle classes than of parents in the higher social classes. It would be inappropriate to pursue these complex and sensitive matters further in this work. Another large and difficult area which again it is not proposed to pursue here is the differential influence of father and mother on the development of their children, with special reference to the influence of each on boys and on girls. Those interested are referred to the study by Bronfenbrenner ( 1961) and to one by Douvan & Adelson ( 1966) who discuss in much detail the difference in developmental patterns shown by boys and girls between the ages of twelve and eighteen.
Further Studies of Small Samples
Next we revert to more intensive studies by considering the findings of three projects in each of which small samples of men
-343-
or youths, selected especially for their apparently healthy and well-integrated development, were subjected to intensive clinical examination and observation over a period of at least a year. Presented in descending order of the subjects' age, the first study is of astronauts in training, the second of youths attending college, and the third of high-school students bound for college.
260
In respect both of the developmental pathways that these personalities are following and of the family life they have experienced or are still experiencing, the findings of these three studies are in agreement; they are in agreement also with those of Peck & Havighurst. First, these well-adapted personalities are found to show a smoothly working balance of, on the one hand, initiative and self-reliance and, on the other, a capacity both to seek help and to make use of help when occasion demands. Second, an examination of their development shows that they have grown up in closely knit families with parents who, it seems, have never failed to provide them with support and encouragement.
So far as it goes, each study gives the same picture, that of a stable family base from which first the child, then the adolescent, and finally the young adult moves out in a series of ever- lengthening excursions. While autonomy is evidently encouraged in such families, it is not forced. Each step follows the previous one in a series of easy stages. Though home ties may attenuate, they are never broken.
Astronauts rank high as self-reliant men capable of living and working effectively in conditions of great potential danger and stress. Their performance, personalities, and histories have been studied by Korchin & Ruff. In two articles ( Korchin & Ruff 1964; Ruff & Korchin 1967) they publish preliminary findings on a small sample of seven men.
Although these men tend to be individualists who show a high degree of self-reliance and a clear preference for independent action, all are reported to be 'comfortable when dependence on others is required' and to have a 'capacity to maintain trust, in what might seem conditions of distrust'. The performance of the crew of Apollo 13, which met with a mishap en route to the moon, is testimony to their capacity to sustain trust. Not only did they maintain their own efficiency in conditions of great danger but they continued to cooperate trustingly and effectively with their companions at the base on earth.
Turning to their life-histories we find that they:
-344-
grew up in relatively small well-organized communities, with considerable family solidarity and strong identification with the father.
