Hmenican DSournal of Fayohiod es, a BE, Vol IHD CL).
Developmental Psychiatry Comes of Age
John Bowlby
‘Although Meyer and Freud believed that the seeds of
mental health are sown in childhood, systematic re-
earch has proved dificult and disappointing, and
theory has remained intensely controversial. Research
of the past two decades by developmental psycholo-
fists using an ethologically based theory of socioemo-
fional bonds, however, is yielding findings that consis-
tently confirm predictions. The author outlines thi=
research, giving examples of the results, which are
‘shown to be compatible with findings of epidemiolog-
ical studies of depressed adults. He holds that a
person's degree of vulnerability to stressors is strongly
influenced by the development and current state of his
‘or her intimate relationsbips.
(Am J Psychi 1988; 145:1-10)
D uring the first third of this century there were two:
-at proponents of developmental psychiatry—
‘Adolf Meyer and Sigmund Freud. Both believed that
the seeds of mental health and ill health are sown in
childhood and that, if we are to understand the
present-day functioning of a person, itis necessary to
know how he or she has become the man or woman
‘we meet today.
In their approaches to the field, Meyer and Freud
took very diferent routes. Initially, Freed focused on
traumatic family relationships, including incest, but
for reasons that remain obscure, he later claimed that
the real-life events he had originally invoked as patho-
genic had never occurred and that the patient had only
‘imagined them. Thenceforward the emphasis was on
fantasy. Consequently, Freud's interests became fo-
cused on a person's internal world of mental processes,
especially on the powerful influence that unconscious
Pocesses haveon the way a person fel, thinks, and
haves and, above all, on the defensive processes that
actively keep them unconscious.
Meyer, by contrast, continued to emphasize the role
of real-life events in shaping personality, but he was
never very specific about the nature of the ones that
matter, nor did he advance any theory of how being
exposed to some event or situation affects a person's
mental state. Nonetheless, Meyet’s approach played a
‘major part in promoting the mental hygiene movement
fand child psychiatry. In both these revolutionizing
movements the notion that the environment in which a
child grows up plays a critical part in determining his,
‘or her future mental health has always been a stub-
bornly held if ill-defined assumption. Morcover, these
are the fields in which Freud’s original ideas about the
role of childhood trauma have not only persisted but
bore valuable fruits,
"We can see now that the tremendous strides being
made today in developmental pychiaty owe 3
deal to both these pioneers. To Adolf Meyer is due the
credit for having continued to emphasize the influen-
tial role of the events and situations people meet with
during their development. To Sigmund Freud is due
the credit for having emphasized the influence on how
people think, feel, and behave that is exerted by their
Internal world—namely, by the way they perceive,
construe, and structure the events and situations they
encounter. Today we know that the central task of
developmental psychiatry is to study the endless inter-
faction of internal and external and how the one is
constantly influencing the other, not only during child-
hhood but during adolescence and adult life as well.
Evidence that happenings within the family during
childhood and adolescence play a major role in deter-
mining whether a person grows up to be mentally
healthy or not is now formidable, and a review of the
important epidemiological findings appeared in 1985
(1), For that reason, in this paper I will present only a
limited amount of data, some epidemiological and
retrospective and some ethological and prospective,
and will give most attention to the conceptual frame-
work within which I believe the different sets of
Findings can best be comprehended.
‘As an example of fetrospective data I draw on the
findings of the group led by George Brown and Tirril
Harris, which has been undertaking sophisticated ep-
idemiclogical studies to explore the roles of family
experience and other social variables as antecedents of
depressive and anxiety disorders in adult life. As an
‘example of prospective data, I draw on the work of
those developmental psychologists who have been
inspired by and are busy expanding the brilliant pio-54
neering studies of Mary Ainsworth. This research into
the development of a young child’s capacity to make
intimate emotionally mediated relationships with par-
ents is probably little known as yet among general
psychiatrists. Since the findings from these two quite
different approaches ate highly compatible, the result-
8 scientific structure can be likened to’ trilithon
made up of two stout pillars of evidence and a
crosspiece of theory. A major conclusion is. that,
whatever influence variations in genetic endowment
may exert on personality development and psychopa-
thol ony, a immense influence is unquestionably ex-
certed by environmental variables of the kinds now
being systematically explored.
EMOTIONALLY MEDIATED BONDS AND MENTAL
HEALTH
For many years, sensitive clinicians have been aware
that a person’s mental state is deeply influenced by
whether his or her intimate personal relationships are
warm and harmonious or tense, angry, anxious, emo-
tionally remote, or, possibly, nonexistent. Among cli-
ici cael a variry of terms ae in, wie;
te co ceymbiogi”
“object relations.” Nevertheless, although a field of
manifest importance to peychiatry, there has been no
a ‘on how best to conceptualize it. A number
of theorecal systems have been almost all
‘of them derived in some sort of way from psychoanal-
ysis, and a mountain of jargon has accumulated. Yet
none has generated sustained productive research. In
scientific quarters the whole field still struggles for
recognition.
Some years ago, as a family psychiatrist researching
the effects on young children of being removed from
their home to the care of strange people, I was faced
with a problem. Since young children in that situation
are intensely distressed and show every sign that they
desperately wish to recover their mother, the question
arises: How do we understand the origin and nature of
this extraordinarily strong tie between child and
mother? The only theory extant at that time was that
a child becomes emotionally linked to his mother
because she feeds him. Dissatisfied with that expla-
nation, | was lucky enough to be introduced to the
ideas of Konrad Lorenz. Ethologists, I discovered,
t as interested in the strong bonds between
ind parents in a range of animal species as I
cof the human species, Moreover, as pro-
fessional biologists, their approach was radically dif-
reht from anything I had met with among peycholo-
st8 and psychiatrists. Bonds between individuals, 1
eafttd; which occur in a great array of mammalian
anduavian species, could be studied experimentally and
he al function considere light of evo-
it d in the light of evo.
Here then was a possible new approach to our
problem and one that seemed to ft our data, Not only
that, but the ethologists lad developed sensitive means
for observing and recording and were sophisticated
fefentats with thet feet firmly planted in che biological
sciences. The upshot was that I fell for this approach
and decided to explore its ials. Now, 30 years
later, Lam glad I embarked on the gamble. Thanks to
the work of many gifted colleagues, the bare bones of
theory are becoming clothed with hard-wearing empir-
ical garments, and there are now several productive
research programs studying the different ways in
which intimate personal relationships can develop,
with their causes and consequences. research
programs are as firmly rooted in biology as are those of
the physiological psychiatrists who have improperly
kidnapped the label “biological psychiatry”
‘Not only are ethological concepts proving extremely
fruitful when applied to our field, but present-day
science provides us many other concepts, not available
to Meyer Fred, thar lp us ela our data. For
example, additional ts of high explanatory
value in our field come ‘control theory and
cognitive psychology, including human information
PriAmong the principal propositions of this new ap-
e
proach (2, 3) are the following: 1) Emotionally signif-
icant bonds individuals have basic survival
a ee fabng cporncaeayeene iw
be systems
be goer by pring ene rs
fect of maintaining proximity of ready accessibility
of each partner to the other. 3) In order for the systems
to efficiently, each partner builds in his or her
ind working models of self and of other and of the
patterns of interaction that have developed between
the) Present knowledge requis that theory of
developmental pathways should replace theories that
invoke specific phases of development in which it is
postulated @ person may become fixated and/or to
Gihich he of she may regress. 5
“The bonds with which developmental psychistry is
principally concerned are those of child to parent and
the comy jentary bond of parent to child. The
hypothesis is that variations in the way these bor
areclop and become organized during the infancy and
childhood of different individuals are major determi-
fant of whesher a person grows up to be mentally
healthy. Before discussing variations, however, it is
necessary to consider how the bonds themselves are
best conceived.
FUNCTION AND ORGANIZATION OF EMOTIONAL
BONDS
Ethological theory regards the propensity to make
scfong cteodonal bonds to particular individals 38 a
basic component of human nature, already present in
germinal form in the neonate and continuing through
‘adult life into old age. During infancy and childhood,
bonds are with parents (or parent substitutes), who are55
looked to for protection, comfort, and assistance. situations that terminate care seeking and release the
During healthy adolescence and adult life those bonds child for other activites are comfort and reassurance,
persist but are complemented by new bonds, com- Since control systems are themselves sources of activ-
monly of @ heterosexual nature, Although food and ity, traditional theories of motivation that invoke a
sex sometimes play important roles in such relation- buildup of psychic energy or drive are rendered obso-
ships thetelatonsip cxssininown igh andbasen Tee, _ *
important fanction of its own—namely, Control systems, by nature, can operate effectively
ation, Thin, wishin the atuackment Iumework, only within a epecded exviconment. For exemple the
bonds are acen as neither subordinate to nor derivative system regulating body temperature cannot maintain
from food and sex. Nor is the urgent desire for comfort an appropriate temperature when environmental con-
and support in adversity regarded as childish, as ditions become either too hot or too cold, which means
dependency theory implies: Instead, the cpacty to that whenever conditions go beyond certsn Kimi the
make bonds with other individuals, sometimes in the whole organism becomes stressed or dies. Viewed in an
seeking role and sometimes in the caregiving one, evolutionary ive, i is evident that variation
= ‘asa principal feature of effective personal- and natural have rested in each oganisn's
ity functioning and mental health, physiological sysiems being so constructed that they
‘As a rule, care seeking is shown by a weaker and less operate effectively in the environment to which the
experienced individual toward someone regarded as species has become adapted and that they will become
stronger and/or wiser. A child or older person in the stressed or fail in others.
Gaze geking roe hepa within range ofthe aregivez, An evolutionary perspective k nectar ao if we
the degree of closeness or of ready accessibility are to understand psychological stress and the enviroc-
pending on circumstances; hence, the concept of at- mental conditions that cause it. Like other control
tachment behavior. gystems, the aytem porening harman stachment be.
‘Another basic component of human nature is the havior is 50 a5 to promote survival in the
urge to explore the environment, to play, and to take environment in which man evolved. Since in such an
par in varied aciviies with peers. This behavior is of environment tis essential for survival that an infant ot
course, antithetical to attachment behavior. When older child has an attentive and responsive caregiver to
individuals of any age are feeling secure they arelikely whom to goin an it follows that his or her
to explore away from their attachment figure. When attachment system will be constructed so 2s to operate
alarmed, anxious, tired, or unwell they feel an urge most efficiently in interaction with a person who the
toward proximity. Thus, we se the typical pare of child beleyes wil respond promply and efectvely
interaction between child and parent ‘at explo- when called upon. It is not at all surprising, therefore,
ration from a secure base, Provided the is that a failure in response of his or her familiar
known to be accessible and will be responsive when caregiver, whether due to physical absence or a failure
called upon, a healthy child feels secure enough to to respond appropriately, should always cause stress
explore. At first these explorations are limited both in and sometimes be traumatic.
time and space. Around the middle of the 3rd year, This leads to consideration of the third major com-
however, # cure thid begilt to Become onlin’ ponent of human nature relevant to this exposition —
increase time and distance away—firat to namely, caregiving, which is the prime role of parents
half-days and later to whole days, As he or he grows and complementary to attachment behavior. When
into adolescence, the excursions are extended to weeks looked at in terms of evolution theory the occurrence
> 0x months, but a secure home base remains indispens-<& of altustic care of the young is readily understood,
able nonetheless for optimal functioning and mental since it serves to promote the survival of offspring (and
\ealth. No concept pe the attachment framework often of other relatives as well) and thereby the indi-
‘more central to developmental prychiatry than that —_vidual's own genes. Nevertheless, this form of expla-
of the secure base. nation constitutes a radical shift from most psycholog-
‘An obvious way to conceptualize the behavior ob- ical theorizing, including Freud's, which has mistak-
scrved is to postulate the existence of a control system nly assumed that individuals are by nature essentially
Within the CNS analogous to the physiological control selfish and that they consider the interests of others
systems that maintain physiological measures, such as only when constrained to do so by social pressures and
biood pressure and body temperature, within set lim- sanctions. Nothing, 1 believe, that stems from an
its, Thus, the theory proposes that, ina way analogous —_ethological perspective has more far-reaching implica-
‘0 physiological homeostasis, the at mnt control tions for understanding human nature than this reap-
|_O2tm maieins «person's elation Se na. aaa at eee
y/ Seackment figure between certain limits of distance It is evident that, to operate efficiently, the attach-
and accessibility. As such, it can be regarded as an ment control system within the child requires to have
Example of what can usefully be termed “environmen- at its disposal as much information as possible about
tal homeostasis” (3), Among situations that activate the self and the attachment figure, not only in regard to
care seeking are anything that frightens a child or their respective locations and capabilities but in regard
Signals that he or she is tied or unwell. Among also to how each is likely to respond to the other as56
environmental and other conditions change. Observa-
tions lead us to conclude that toward the end of the
first year of life children are acquiring considerable
knowledge of their immediate world and that during
subsequent years this knowledge is best irded as
becoming organized in the form of internal working
models, including models of self and mother. The
function of these models is to simulate happenings in
the real world, thereby enabling individuals to plan
their behavior with all the advantages of insight and
foresight. The more adequate and accurate the simu-
lation, of course, the better adapted is the behavior
based on it likely to be (4). Although our knowledge of
the rate at which these models develop during the
earliest years is still scanty, there is good evidence that
by their Sth birthday most children are using a sophis-
ticated working mode! of mother or mother-substitute
which includes knowledge of her interests, moods, and
intentions, all of which the child can then take into
account G. ‘With a complementary model of self, the
is already engaging in a complex intersubjective
relationship with the mother, who, of course, has her
‘own working models both of her child and of herself,
Because these models are in constant use, day in and
day out, their influence on thought, feeling, and behav-
ior becomes routine and largely outside of awareness,
As clinicians we know that long before children
reach the age of 5 the patterns of interaction between
them and their mother range vastly in diversity, from
‘smooth running and happy to filled with friction and
distress of every kind and degree. We also know that
these patterns are apt to persist; therefore, the more we
i about how: - oe the ie It is here
it recent researc mental sists
has made such ge shes ements PHO
PATTERNS OF ATTACHMENT AND THEIR
DETERMINANTS
‘Three principal pattems of attachment present dut-
ing the early years have now been reliably identified,
together with the family conditions that promote
them. One of these patterns is consistent with the
child's developing healthily, and two are predictive of
disturbed development. Which patra any one indi
vidual develops during these years has been found to
be profoundly influenced by the way his or her parents
(or other parent-type figures) treat him or her. This
conclusion, which is as important as it has been
controversial, is unpopular in some citcles and, in
consequence, is constantly challenged, Yet the evi-
dence for itis now weighty and derives from a number
of prospective research studies of socioemotional de-
velopment during the first 5 years. This research
tradition was frst set by Ainsworth etal, (6-8) and has
tince been exploited and expanded, notably in the
United States by Main et al, (9-11), Sroufe (12,13),
and Waters et al. (14,15) and in Germany by Gross-
mann et al. (16),
The pattern of at sistent with healthy
devopaent i Beat ee ae ae iment, in which
individuals are confident that their parent (or parent
figure) will be available, responsive, and helpful should
they encounter adverse or frightening situations. With
this assurance, they feel bold in their explorations of
the world and also competent in dealing with it. This
pattern is found to be promoted by a parent (in the
tatty year, expecially by the mother) being. readily
avaiable, sensitive to he cis signals, and lovingly
responsive when he or she secks protection andlor
‘comfort andlor assistance.
‘A second pattern is that of anxious resistant attach-
meat in which the individual is uncertain whether Bis
ot her parent will be available or responsive or
when called upon, Because of this uncertainty the
is always prone to separation anxi
clinging, and is anxious about exploring the world,
This pate is promoted by a parent being available
and helpful on some occasions but not on others, by
separations, and, later, especially by threats of aban-
donment used as a means of control.
A third pattern is at of apxious avoidant attach
ment in which individuals have no confidence that
‘when they seek care they will be responded to helpfull
but, on the contrary, expect to be rebuffed. Su
individuals attempt to live ther life without the love
and support of others, This partern isthe result ofthe
individuals’ mother constantly rebuffing them when
they approach her for comfort of protection, The most
‘extreme cases result from ted rejection and ill-
treatment or prolonged institutionalization, Clinical
evidence suggests that, if it persists, this pattern leads
to a variety of personality disorders, from compul-
sively self-sufcient individuals to persistently delin-
quent ones,
‘There is much evidence that, at least in families
where caregiving arrangements continue to be stable,
the pattern of attachment between child and mother,
‘once established, tends to persist. For example, in two
different samples (Californian (11) and German [17])
the patterns of ateachment to mother at 12 months of
age were found, with but few exceptions, still to be
pvent at 6 years, Furthermore, prospective studies in
inneapolis (12) have shown that the pattern of
attachment characteristic of the pair when the child
was aged 12 months was highly predictive of behavior
outside the home in a nursery group 3¥4 years later,
Thus, children who had shown a secure pattern with
their’ mother at 12 months old ‘were likely to be
described at 3¥4 years old by their nursery teachers as
cheerful and cooperative, popular ‘with other children,
resilient, and resourceful. Those who had shown an
anxious avoidant pattem were likely to be described
later as emotionally insulated, hostile, or antisocial
and as unduly seeking of attention. Those who had
shown an anxious resistant pattern were also likely to
be described as unduly seeking of attention and as
either tense, impulsive, and easily frustrated or passive
and helpless.Ample confirmation of the teachers’ descriptions
comes. from i it observers and laboratory
assessments of the same children (12, 18). Similarly, an
experimental study done in Germany (19) showed that
at 3 year of age children earlier assessed as securely
tached responded to potential failure with increased
tffor, whereas the insecurely attached did the oppo-
ended wih conkdeoce snd Kope tha hey could
sponded with confidence and hope that they cou
Succeed, whereas the insecure already showed signs of
helplessness and defeatism.
In a number of these studies, detailed observations
have been made of the inch the children’s mothers
treated them, Great variability was seen, with high
Contlations between a. mothers style of interaction
and the child’s pattern of attachment to her. For
‘example, in one such study (20), done when the
Ghildren were 24 years old) mothers were observed
while their children were attempting a task that they
could not manage without a little help. Mothers of
secure toddlers ee their ieee re ee
task, respected their attempts to comy it on their
and ‘with the required help. when
upon, thers of insecure infants were less
Toloocnr connetat suenleesead snforioee
fiving support and help when appealed to or
intruding when the chil ‘were striving to solve the
problem themselves,
In discussing these and similar findings, Bretherton
(21) emphasized the easy flow of communication be-
tween a mother and her child in secure partnerships
and concluded that this easy communication is possi-
ble only when a mother is intuitively alive to the
crucal part she plays in providing her child with a
secure base, variously encouraging autonomy, provid-
ing necessary help, of giving comfort according to bet
child’s state of mind.
Mothers of insecus fants deviate from this sensi-
tive pattern of mothering in a great variety of ways.
One, common among mothers of avoidant infants, is
to scoff at her child's bids for comfort and support
(11). Another, well-known to clinicians, the effects of
which are now being observed by deve! lists
(22, 23), sa mother who fails to respect her child's
dete ft atonomy and ducourages exploration, This
is usually a mother who, not having had a secure home
base during her own childhood, is consciously or
unconsciously seeking to invert the relationship by
making her child her own attachment figure. In the
Patt this has too often been labeled “overindulgence”
‘or “spoiling,” which has led to appalling confusion
about what is best for a child.
It is not difficult to understand why patterns of
attachment, once developed, tend to persist. One rea-
son is that the way a parent treats a child, whether for
or for worse, tends to continue unchanged;
another is that each pattem tends to be self-perpetu-
ating, Thus, a secure child is a happier and more
rewarding child to care for as well as less demanding
than an anxious one, An anxious ambivalent child is
new relationships
7
apt so be whiny and clinging, and an anxious avoiden
ild keeps his or her distance, is bad- and is
Bone to ball other chlden In cach of thes later
‘cases the child’s behavior is likely to elicit an unfavor-
able response from the parent, 40 that vicious circles
lop.
For the reasons given, patterns, once formed, are
likely to persist, but this is by no means necessarily s0.
Evidence shows that during the first 2 or 3 years the
pattern of attachment is a property of the relationship.
For example, the pattern of child to mother may dif
from the pattern of child to father; also, ifthe parent
treats the child differently the pattern will change
accordingly. These changes are among much evidence
reviewed by Sroufe (13) that stability of pattern, when
it occurs, cannot be attributed to the child’s inborn
temperament, as has often been claimed. On the
contrary, the evidence points unmistakably to the
corclusion that a host of personal characteristics tra-
ditionally described as temperamental and often as-
cribed to heredity are environmentally induced. True,
nconates differ from each other in many ways. Yer the
evidence is crystal clear from repeated studies that
infants described as dffcult during their early dayr are
‘enabled by sensitive mothering to happy, easy
toddlers. Contrariwise, placid newborns can be turned
into anxious, moody, ing, or awkward tod-
dlers by insensitive of rejecting mothering, Not only
did Ainsworth et al, (6) demonstrate this in their
‘original study, but ithas been found again and again in
subsequent ones. Those who attribute so much to
inborn temperament will have to think again.
‘Thus, during the earliest years, features of personal-
ig crucial to payciaey.emsin relatively open to
ange because they are still responsive to the environ-
ment. AS a child
dence shows that ‘pattern of attachment and
the personality features that go with it become increas-
ingly a property ofthe child himself or herself and also
increasingly resistant to change. This means that the
child tends 0 impose it, or some derivative of it, upon
new relationships, as with a teacher in the Minneapolis:
study (12). Similarly, experience shows that the child
tends also to impose it or some derivative of it on a
foster mother or a therapist,
‘These tendencies to impose earlier patterns onto
‘and in some measure to persist in
doing so despite absence of fit are, of course, the
phenomena hat gave bh to psychoanalysis. They ae
also the phenomena that during recent decades have
led an increasing number of analysts to embrace an
interpzrsonal ot object relations version of psychoana-
lytic theory and, in my own case, to advance the
.ttachment version with its postulate of internal work-
ing models of self and attachment figure in interaction.
Thus far, therefore, the picture presented can be
looked upon as a much modified and updated variant
of traditional psychoanalytic thinking in which great
‘emphasis is placed on the particular pattern into which
each personality comes to be organized during the
‘older, however, clinical evi-<—58
FIGURE 1. Developmental Pathway Deviating Toward Antious AI-
tachment and Depressica
early years, with its own distinctive working models
(or, to use the traditional term, “internal world”) and
‘on the strong influence thereafter that each individual's
models have in shaping his or her life.
CONTINUITY AND DISCONTINUITY:
‘VULNERABILITY AND RESILIENCE
One of the long-running debates between analysts
and others in our field turns on the extent to which it
is believed that personality, once developed during the
carly years, is open to change. Analysts have empha-
sized a strong tendency toward stability and continuity
and have explained it as due to the powerful influence
ofthe individual's existing internal world on ow he or
¢ construes and responds to every new situation.
Critics have emphasized the extent t0 which an indi-
‘duals performance can change given changed con.
ditions of life. If we are to do justice to the plethora of
data now available, we must abandon simple dichor-
omies. First, a sharp distinction must be drawn be-
tween current functioning, measured in terms of pres-
‘nee o absence of psychiatric disorder, and personal-
ity structure, measured in terms of greater or less
vulnerability’ to adverse life events and situations.
Linked closely to degrees of vulnerability, moreover,
are the very differenc ways that people feel about their
lives, either as mostly enjoyable and to be lived to the
full or else asa burden to be endured, as an emotion-
ally rich and varied experience or an emotional desert.
Second, we have at all times to think in terms of the
interactions and transactions that are constantly oc-
curring between an ever-developing personality and
FIGURE 2. j Ta it
FIGURE 2. Deeopmental Patheny Deiating Toward Hoey and
the environment, especially the people in it. This
‘means that itis necessary to think of each personality
as moving through life along some developmental
athway, with the particular pathway followed always
Eeing determined By the interaction of the personalty
as it has s0 far developed and the environment in
which it then finds itself. If as developmental psychia-
tists we adopt these ways of thinking, we need to
picture each personality as moving through life along
sown unique pathway. So long as family conditions
are favorable, the pathway will stare and continue
within the bounds of healthy and resilient devel
ment, but should conditions become sufciently unta-
‘orableat any time itmay deviate to a lesser or greater
‘extent toward some form of disturbed and vulnerable
development. Conversely, should an infant be born
into unfavorable conditions, the pathway along which
it develops may become deviant very eatly; yet once
again, should there be change, in this case for the
better, there is a chance of the deviation diminishing to
‘a lesser or a greater extent. Examples of two such
deviant pathways are illustrated in figures 1 and 2.
This leads me back to patterns of attachment, since
my hypothesis is that the pathway followed by each
developing individual and the extent to which he o
she becomes resilient.to stressful life events is deter-
mined to a very significant degree by the pattern of
attachment he or she develops during the early years.
Furthermore, this implies that, in identifying the family
gaperiences that resin diferent children developing
different patterns of attachment, my colleagues are
identifying also some of the major determinants of
‘each person’s future resilience or vulnerability to life's
sihazards as well as the extent to which he or she will be
able to srioy, C2 Jy, the detailed .
To date, unfortunate! 1e rospective
ruhes deeribed earlier of how some of the diferent
ratterns of attachment come into being and of their
Felative persistence have so far not been carried beyond
the 6th year. This means that the hypothesis, however
plausible, will be without rigorous testing for some
Plars yet, Meanviile, however, backers ofthe hypoth
{sis point to much supporting evidence coming from
the ther major schol of rescatch fered neti in
this paper—namely, epidemiological studies of adults
who a suffering one or another of a wide
of psychiatric disabilities, including anxiety and de-
pressive states, suicidal ior, line condi-.
tions, and sociopathic personality.
From among a large array of epidemiological studies
I have chosen the work of George Brown and Tirril
Harris to illustrate my thesis because it not only is of
the highest quality ‘but also pays attention to variables
‘of the kind ‘that the prospective research already
described is showing to be so very influential, In
addition, they are using a sophisticated model of
developmental pathways for interpreting their find-
een sr, splay
large representative samples of the tion,
fwe in'loner Loudon "boroughs, oe in. tn cone
borough, and one in a remote rural area of the Scoctish
Highlands. (In two of their studies they explored the
effects of loss of the father during childhood because
there is often'a higher incidence of such loss am
samples of depressed women than among control
subjects. Since detailed analyses show that loss of the
father has much less influence on vulnerability than
loss of the mother, further reference to itis omitted.)
‘The aim of these researchers has been to identi
women suffering from depression of a clinically signi
icant degree and to compare them with other women
inthe same community who are not depressed in order
toddacover whether there are any fal or other socal
variables that distinguish the two groups. The findings
of the first such study, done in the inner London
borough of Camberwell were published 10 years ago
(24) and are becoming well-known. Init they identified
four classes of variables that were found significantly
more often among the depressed members of the
Population. Three of these concemed current events
and conditions: 1) a severe adverse event, usually
involving an important personal loss or disappoint-
‘ment, that had occurred within the year before onset,
2) the absence of a companion in whom to confide,
and 3) chronically difficult living conditions, including.
extremely bad housing and cesponsibility for caring for
a number of children under the age of 14,
The fourth variable associated with current incepac-
itating depression is a historical one—namely, a
Woman’s loss of her mother due to death or prolonged
Separation before her 11th birthday. In subsequent
studies of two other large and mainly working-class
39
‘samples, in Walthamstow and Islington (25-28), find-
ings of similar sort have been the rule. The incidence
‘current affective disorder in women who had lost
their mother before their 11th birchday compared with
those who had not done so were, in the three samples,
43% versus 14%, 25% versus 7%, and 34% versus
17%, respectively. Moreover, in the second and third
studies (26, 27) it was found that women who had lost
their mother during adolescence (between their 11th
and 17th birthdays) were also more prone to develop
depression than control subjects, although less so than
those who had lost their mother when they were
younger. The consistency of these findings, together
with findings from a long-term prospective study re-
ported by Wadsworth (29), gives strong support to the
clinically derived that childhood loss of
one’s mother is likely to lead a person to become
excessively prone to develop psychiatric symptoms and
to do s0 especially when current personal relationships
g0 wrong.
Jn their Walthamstow study (27), members of the
Brown-Harris group undertook a prolonged and ex-
tensive interview with each woman in order to gain as
Imuch information as possible abou the family cium
stances during her chi ‘and adolescence and also
Pare pent ie cou, ining, for exam-
jot boyfriends, any premar-
ital , and her husband and children (if any).
‘With the sample deliberately structured to include a
high proportion of women who had suffered a child-
hood loss or prolonged separation, it became possible
to analyze the extent to which family variables other
than loss may have contributed to a woman's current
vulnerability.
In keeping with expectations it was found chat both
the family circumstances which had led to the child-
hood lost and the adequacy of care the gil had
received afterwards were it con!
tworve the family circumstances before the lose and the
more inadequate the care after it the more vulnerable
to depression the woman had become.
‘The very detailed analyses of these data carried out
by the researchers have led them to explain their
findings in terms of developmental pathways that,
through the continuing interactions of social and per-
sonality factors, result in a working-class woman in
London who loses her mother for any reason to be at
‘an appallingly high risk of becoming depressed. In th=
first place, they found, such a loss carried with it a high
likelihood that she will not receive adequate care; in
the Walthamstow study (27) the chances were found
to be no better than 50-50, Subsequently, her troubles
were likely to snowball. Should the care she received
have been poor, risks were high that she would become
pregnant before marriage and/or that she would ven-
ture on an early and ill-advised marriage. These two
‘occurrences were found to be strongly associated with
later depression, in part because they usually resulted
in her living in very unfavorable conditions at high risk
of suffering a severe adverse event and with no one in60
FIGURE 3. Female Pathways to Depression®
(hams pss and ow set-in
Asus
saci!
Adapted from Harris tal. (27)
whom to confide. For these reasons a working-class
her mother was al too likely to
find herself on a slippery slope.
Figure 3, adapted from one designed by Hari tal
(27), attempts to illustrate these interacting processes.
‘The thick upper line represents the woman's mental
state. The middle line represents what may for conve
niience be described a8 her family environment. The
thin lower line represents her socioeconomic environ-
ment, in this case urban working class. The broken
lines indicate the endless ways in which mental state
and environment interact, with environmental happen-
ings affecting mental state and mental state, in turn,
affecting the way the girl or young woman deals with
aspects of her environment.
‘Two further findings, both from the Islington study
(28), illustrate this interaction. A woman who had
received inadequate care, either in or out of her family,
was found to be twice as likely to have developed a
negative self-image as a woman who had received
adequate care. The figures were 54% compared with
27%, Similar differences have been reported from
Edinburgh (30).
Furthermore, the marriage she had entered into was
more likely to have failed than was that of the more
fortunate woman. Here the figures were 36% and
23%. The high risk of this happening was evidently the
recult of a chain of adverse happenings. For example,
when a young woman has no caring home base she
may become desperate to find a boyfriend who will
care for her. That, combined with het f-
make unduly intense demands on her husband and,
should he fail to meet them, to treat him badly. No
‘wonder one in three of these marriages broke up.
‘Gloomy though these conclusions are, we must
remember always that a disastrous outcome is not
inevitable, The more wecure an anachment ¢ woman
experienced during years, we can confi.
dently: predict, the ge Oa Ge her chances of
‘eacaping the slippery slope.
CONCLUSIONS
In conclusion, 1 refer, first, to some of the
scientific problems that clamor for attention
second, to some of the practical implications of our
wali sie tak to explore snd tet the
yypothesis, stated earlier, that ¢ach person’s resilience
of vlnerabilty to sre is event deermioed
a very significant degre by the pattem of artachment
he of she develops during the early years and, more
expecially, to clay to what degres and in what wa
the early-developed patterns influence subsequent
velopment, Since, however, there are many subpat-
tems of each of the three main patterns, to elucidate
the influence of each will be a long and time-
consuming task, An integral part of such research is
the further examination of precisely what environmen-
tal conditions, impinging earlier and/or later, enable a
person to retain or attain a greater degree of resilience
and, conversely, what conditions push him or her
toward greater deprees of valerbilty.
Yet another substantial group of variables tobe
taken into account are the heritable differences that
must be a sumed to exist berween diferent individualstal hazards, especially inadequate parenting, which
influence vulnerability.
Im undeaking such a sexes of esearch programs,
an early requirement is the development of psycholog-
ical method for assessing partes of attachment and
their derivatives at each phase ofthe life cycle. Main et
al, (11) and other developmental psychologists are
already making some promising moves in this direc-
tion. To cast light on problems of continuity and
discontinuity, both of patterns of attachment and also
of erent depress of resience and vulersbiliy,
rospective studies following iy ment
Ehrough diferent phates of the life cyce and fn if
ferent environments are plainly indispensable, despite
their being very costly.
aus progam of reewch ie deopmeal pepe,
mous program of research in eycho-
pathology (31), one that will clearly require genera
tions of research workers. Both in magnitude and in
the biological principles informing it it can best be
compared to the vast program of research that has
been and is still being undertaken in immunology.
Here, as in our own eld, research workers are con-
cerned with the extraordinarily complex interactions
and transactions that occur between an organism as it
develops over the years and the array of hazards
present and potentially preset init environment
cepts analogous to those in our field include the
degree to which an organism is immune to a wide
variety of hazards and the extent to which an existing
state of immunity will persist or change over time.
‘What then are the practical implications of our
present knowledge? For the clinician concerned with
the assessment and treatment of a wide range of
psychiatric patients of all ages it provides a develop
mental psychopathology that is biologically based,
coherent, and already empirically wel As
such it can provide guidance for under wa
Patient's problems and symptoms have evolved in his
or her transactions with the particular environments
he or she has encountered from infancy to the present
day, thus providing an updated and detailed version of
the life chart, which Adolf Meyer (32) so rightly
advocated. Furthermore, it can provide guidance also
for planning therapeutic intervention. Although much
detailed work is still required in these areas, promising
starts have been made (33-36).
‘What, however, is even more important (and would
be especially welcome to Adolf Meyer) is the firm
guidance it gives for prevention, As I have emphasized
throughout, 1 believe there is already sufficient evi-
dence, coming from diverse and independent sources,
that points to the very substantial influence on person:
ality development and mental health of the way an
individual’s parents (or in some cases parent substi-
tutes) treat him or her. Given affectionate and respon-
sive parents who throughout infancy, childhood, and
adolescence provide a boy or girl with a secure base
from which to explore the world and to which to
return when in difficulty, it is more than likely that a
61
child will grow up to be a cheerful, socially coopera-
tive, and elective cltzen and to be unlikely to Break
down in adversity. Furthermore, such persons are far
more likely than those who come from less stable and
supportive homes to make stable marriages and to
provide their children with the same favorable condi-
tions for healthy development that they enjoyed them-
selves (37, 38). These, of course, are age-old truths, but
they are now underpinned by far more solid evidence
than ever before,
Unfortunately, as we also know, there is another
and lest happy’ sde to the, picture, Childen and
lescents who grow up without their home base
Providing the necessary and encouragement
are likely to be less to ind life—-especily
intimate relationships—dificult, and to be vulnerable
in conditions of adversity. In addition, they are likely
to have difficulties when they come to marry and have
children of their own. It is fortunate, of course, that
despite these handicaps some manage to stru
through, often at a much cost to their
emotional life than me. the undiscernis
. Nor
‘must the fortunate exceptions blind us to rule.
‘Thus, to take an analogy from physiological medicine,
the fact that some heavy smokers survive is no argu:
ment for continuing to use tobacco.
In this paper Ihave described how I see the state of
developmental psychiatry today. On this basis I think
‘we can claim it has now come of age.
REFERENCES
1. Rutter M: Resilience in the face of adversiy: protective factors
and resistance to paychiatric disorder. Br J Paychiatry 1985;
147598611
2. Bowiby J: Attachment and Loss, vol I: Arachment, 2nd ed.
New York, Basic Books, 1982
). Bowlby J: Arachment and Loss, vol It Separation: Anxiety and
Anger. New York, Batic Books, 1973
Johnson-Laird PN: Mental Models. Cambridge, Cambridge
University Press, 1979
‘5. Light P: Development of a Child's Sensitivity to People. Lon-
‘don, Cambridge Univercy Press, 1979
6. Ainsworth MD, Blehar MC, Waters E, et al: Paterns of
‘Atachment: Assessed in the Stange Sitation and st Home.
Hillsdale, NJ, Lawrence Enbaum Associates, 1978
7. Ainsworth MD: Parts of infant-mother stachment: ant-
‘cedents and effects on development. Bull NY Acad Med 1985;
61771791
8. Ainsworth MD: Arcachment across the lifespan. Bull NY Acad
Med 1985; 61:792-812
9. Main M, Weston DR: The quality of the toddler's relationship
to mother and to father: related to conflict behaviour and the
teadines to establish new relationships. Child Dey 1981; 52:
932-940,
10, Main M, Stadtman J: Infant response to rejection of physical
‘contact by the mother: aggression, avoidance and confi. J Am
‘cad Child Psychiatry 1961; 20:292-307
11, Main M, i Cassidy J: Security in infancy, childhood
and adulthood: 4 move #0 the level of representation, in
Growing Points in Atachment Theory ard Reseach: Mono-
szaph ofthe Socery for Research in Child Development Serial
4209. Edited by Bretherton I, Waters E. Chicago, University of
(Chicago Pres, 1985,
12, Sroufe LA’ Infantcaregver attachment and parterns of adapta-od
sion in preschool: the root of maladaptation and competence,
in Minnesota Symposium in Chit Prychology, 16. Edited by
Femur M. Mineapats, Unity of Mans Pres
158
13, Stoule LA: Artchment classification from the perspective of
infantcaregver telatonships and infant temperament. Child
Dev 1985; S6:1-14
14, Waters Ey Vaughn BE, Egeland BR: Individual diferences in
infancmatherarachmentflatonships at age one: antecedents
im neonatal behavior in an urban, economically disadvantaged
sample. Child Dev 1980; $1:208-216
15, Waters E, Deane KE: Defining and atsesing individual ditfer-
ences in atachment relationships: Qmethodology and the
‘reprizaion of behaviour im infancy and eatly childhood, in
‘Growing Points in Artachment Theory and Reseach: Mono:
tzaph ofthe Society for Research in Child Development Serial
209. Edited by Brethereon I, Waters E. Chicago, University of
Chicago Press 1985
16. Gronsmann KE, Groumann K, Schwan A: Caputing the wider
tiew of aachient a teanalyis of Ainwors setnge suse
tion, n Measuring Emotions in Infante and Children, vol 2-
Edited by lard CE: Read PB. New York, Cambridge University
Pres, 1986
17, Warner VG: Attachment in Infancy and at Age Six, and
Children's Sci-Concept: A Follow-Up of» German Longitudi-
nal Sra (doctoral ncenation), Charlottesville, University of
Virginia, Department of Psychology, 1986
Lafrenie [ Sroufe LA: BraBles of peet competence in the
pre-school: intr elations between measures, influence of social
Ecology and relation othe artachment history. Developmental
Paychol 1985; 2156-69
19, Litkenhaus P, Grosamann KE, Grosumann XK Infancmother
attachment a twelve months and syle of interaction with
Stronger a the age of three years. Child Dev 1985; 56:1538—
isan
Maas L, Arend RA, Scoufe LA: Continuity of adaptation in the
second yea: te relationship Between qual of attachment and
ister competence. Child Dev 1978; 49:547~556
21, Breeton Rw perspec on ache el
olintant Developmen, 2nd ed. Edited by Orotsy J.
John Wiley 8 Sons, 1987
22, Stoufe LA, Jacobvie D, Mangehdort $, etl: Generational
boundary isolation berween mothers and their preschool
iden relationship systems approach. Chik Dev 1985; $6
517-335
23, Main M, Solomon J: Discovery of an insecure disorganized
soriented arachment pattem, in Aflecive Development in
20,
mM.
26.
2.
28,
2»
a
3.
3
3s.
36,
a7.
Infancy. Edited by Yogman M, Brateron TB. Norwood, NJ,
Able, 1986
Brown GW, Harris T: The Social Origins of Depression: A
Study of Pacha Dioderin Women New York Free Pres
‘Brown GW, Harris TO, Bifulco A: Long-term effects of early
fous of pacen, in Depression in Young People: Developmental
and Clinical Perspecnves. Edited by Rutter M, lzard CE, Read
B, New York, Guilford Press, 1986 °
falco ATM: Death of Mother in Childhood and Clinical
Depression in Adult Life: A Biographical Approach to Aetiol-
287 (doctoral dissertation}. London, University of London,
Department of Philosophy, 1985
Harris T, Brown GW, Bfulco A: Loss of parent in childhood
in prom pepuney, Peel Med 1907 1731S
tole bie cut TO; Glo lu of xe
(4 adequate parental cate and acule depression: 2 Feplic
ton, } Afective Disord 1987, 12:115-128 ae
‘Wadiworth ME}: Early stress and associations with adult
health, behaviour and parensing, in Sires and Disability it
Childhood: The Longrem Problems. Edited by Butler NR,
Comet BD. Bristol, England, Joha Wright, 1984
Ingham JG, Kreitman NB, Miller FM, etal Seltesteem, wul-
erably and psychiatic disorder in the community. Be J
Paychiatry 1986; 148:375-385
Sroufe LA, Rutter M: The domain of developmental psychopa-
thology. Child Dev 1984; $5:17-29
leyet A: The life chart (1919), in The Commonsense Paychi-
47 of Adel Meyer. Eated by Lic A, New York, McGraw.
il, 1
Heard DH: The relevance of attachment theory to child psychi-
atic practice. J Child Paychol Psychiatry 1981; 22:88-96
Heard D: Family astems andthe stachment dynamic. |Family
‘Therapy 1982; 4:99-116
[Belsky J, Nezworski T (eds: Clinical Implications of Arzach-
‘ments. Hilldate, NJ, Lawrence Erlbaum Associates, 1987
Lieberman AF, Paw! JH: Disorders of tachment in the second
year: clinical developmental perspective, in Attachment inthe
Preschool Years, Ete by Grenberg Mt, Cech, Cm
mings M. Chicago, University of Chicago Press (imprest
Guin, Ror mehr fete ein
care, in Longitu ies im ology and Pry
try: Prac Letsons From Research Experience. Edited by
Nicol AR. New York, Joha Wiley & Sons, 1985
Dowdhney 1, Skuse D, Rutter My etal: The nature and
Parenting provided by women raised im institution
PaycholFspchiatry 1985; 26:599-625