We will never know whether Willy Loman's depression fulfilled current operational criteria for a major affective disorder or whether his family had a genetic loading for depression, but his despair and suicide suggest the way in which career, individual, and family variables may be interwoven in the etiology of depression. ' Some spectators may focus on career variables, deploring the cruel way in which a new generation of company leadership discarded him. Others point out the role of Willy's personality liabilities - as Biff, his son, put it - "He never knew who he was."1 Still others recognize how being part of a chronically conflicted family contributed to Willy's depression.
In this way, Vie Death of a Salesman speaks eloquently to the role of multiple variables in the etiology of depressive disorders. A biopsychosocial model is necessary because relying on any single system of etiologic variables invariably oversimplifies and distorts the complex interactions that produce a depressive syndrome. Indeed, when we focus on the role of career and family variables we are selecting only two of the many relevant systems of variables. We run the risk of oversimplification because we are unable to attend the interaction of many important variables. Willy Loman's depression and suicide can be used to illustrate this fact. Let us assume that three of the variables involved in his depression were based in his personality, his career, and his family.
There is much that we need to know about the structure of Willy's personality: Does he carry a heavy genetic loading for affective disorders? Is he chronically burdened with a fragile ego, immature defenses, and a tenuous selfsystem? Is there evidence of early organic change in certain central nervous system functions? Does he have an underlying physical disease? Why does he cling so desperately and so long to a life dream that has no apparent base in reality? These are but a few of the unknowns. If we could really understand Willy Loman, his personality assets as well as liabilities, we would have one piece of the puzzle. We must acknowledge, however, that there are thousands of Willys, yet not all become depressed and commit suicide. Why this Willy?
When we turn to career variables in the search for understanding, our attention is drawn to the fact that Willy was "let go" late in his career. Although we can identify empathically with that stress, we need to know more. For example, what are the chronic stresses of his job? What gratifications? We know that a career offers some an ideal context for personal fulfillment and growth, but many others receive little beyond wages and, too often, severe stress. Can we be specific about certain aspects of his career that are noxious? Moos has provided a framework in the Work Environment Scale which emphasizes three broad dimensions." The relationship dimensions include such factors as the degree to which the job involves friendly and supportive colleagues and a management that encourages and rewards. Personal growth dimensions include the ways in which selfsufficiency is encouraged and work pressures are handled. Systems dimensions focus on the clarity of policies, the emphasis on control of employees, the amount of variety possible, and the degree to which physical comfort is considered. Jobs that are perceived as offering little in the way of support and cohesion are associated with depressive disorders, even in longitudinal studies which control for initial levels of maladjustment. ,_5
Wetzel has published a series of studies that suggest interactions between individual, work, and family variables.6"" She hypothesizes that workers with either strong dependent or independent personality features are more vulnerable to depression if their individual characteristics are unsupported by specific dimensions of their work environment. The distinctions in this demonstration are important steps in clarifying crucial interactions of variables in the etiology of depressive disorders.
To return to Willy Loman, it is clear that more information is needed about the specific career variables that bear on his depression and suicide. What of the failure of his life dream, the loneliness of the traveling salesman, the lack of peer support, and the change to a less supportive and ultimately rejecting management?
When we examine a third group of variables - those involving the family - the complexity of the etiologic field is increased, but we gain a more thorough understanding of the interactions that together may produce a depressive disorder. It is the role of interpersonal variables in general, and family variables in particular, that this article addresses. Although the role of interpersonal and family variables in predisposing, precipitating, and sustaining an individual's depressive syndrome is recognized but incompletely understood at the present time, 1 wish to review representative studies that suggest that the nature of the attachments provided by the family and the role of those attachments in dealing with separation and loss may be critical.
In order to organize diverse data, this article will address: 1) family processes that predispose to depression: 2) family processes involved in recovery from, or relapse to, depressive syndromes; and 3) the impact on a family of a family member's depression.
FAMILY PROCESSES AND PREDISPOSITION TO DEPRESSION
It is instructive to take a broad view of the ways in which interpersonal relationships in general may be involved in depression. The idea that the quality of an individual's relationships with others and, in particular, the presence of a "special" or intimate relationship may be a crucial factor finds support in a variety of studies. Those who live alone, for example, have a higher prevalence of depression than do those who are married, and those who lose a loved one are more susceptible than are lifelong isolates.'''10 For those living with another person, the psychological quality of the relationship may influence susceptibility to depressive syndromes. The decisive element appears to be the capacity to share feelings and thoughts in an intimate fashion and, as a consequence, to feel emotionally supported. The presence of such a relationship may diminish vulnerability to depressive disorders.
Data to support this proposition are derived from studies of full-time mothers and housewives. Brown and Harris describe the protective impact of a psychologically intimate relationship for such women." A working-class woman without an intimate relationship with a man is much more likely to become depressed following severe stress than is a woman with such a relationship. Belle and colleagues also studied low-income mothers and their children and report that living with a man who provides emotional support reduces the likelihood of depression. I: Lo went hai and Haven studied intimacy as a critical variable in buffering against social losses in persons aged 60 and older, and their findings offer support for this proposition." The majority of subjects who lost a confidant were depressed, and the majority of those who maintained one were not. Furthermore, having a confidant protected one against depression despite decreased overall social interaction. Increased social interaction, however, in the absence of a confidant did not provide such protection. The odds for depression were overwhelming in subjects who have both low social interaction and the absence of a confidant. The authors also note that a confidant protects against the occurrence of depression following retirement.
These studies, while not necessarily focusing on the family, suggest that whether or not one's environment provides the possibility of intimacy may be critical in the predisposition to depression. The Timberlawn studies of both clinical and nonclinical families support this PrOpOSiIiOn.14"'" Our data support the concept that there is a continuum of family competence (defined as providing support for the parents and producing autonomous children). Highly competent families share a number of structural characteristics. The parents share power and have the capacity for unusual openness with each other. Family members' ego boundaries are clear, they encourage open expression of feelings, are sensitive to each others' messages, and use negotiation in problem solving. A wide range of feelings are expressed, and empathie responsiveness is the rule. These families encourage the development of high levels of individuality and at the same time facilitate closeness between family members. Families rated as highly competent and demonstrating those characteristics were found not to contain individuals with psychiatric syndromes including depression.*
Families functioning at a level between the highly competent and the clearly dysfunctional families share some characteristics of each group. These "competent but pained" families have as a central defect a lack of parental intimacy. Husbands are remote and task-oriented. The wives are lonely and angry, blaming their husbands for their emotional isolation. The husbands, however, view their wives as nagging, complaining, and difficult to love. Each spouse blames the other for the marital failure. These couplés have failed to achieve a mutually satisfactory level of intimacy, and the resulting pain casts a long shadow on the family. There is a propensity for many of these wives to become depressed. Although their depressive syndromes often do not fulfill diagnostic criteria for major affective syndromes, they are clinically depressed and are often treated by primary care physicians with antidepressive agents.
Two types of clearly dysfunctional families are associated with depressive disorders in family members, and in both types of families intimate relationships are absent. The dominant-submissive family is tightly controlled and dominated by one powerful parent. The submissive parent may accept powerlessness in a trade-off for freedom from responsibility, or may rebel in a variety of passive-aggressive or manipulative ways. In the latter circumstance, the submissive spouse may enter coalitions with a child and covertly encourage the child's rebelliousness. These families have an oppressive tone, and family members are often angry or sad. Expressing feelings or thoughts divergent from those of the dominant parent is frightening, and family members become inhibited and constricted. Intimacy is not possible in such a family because it is simply too dangerous to be open about one's feelings and thoughts in the presence of a dictatorial family member.
The second type of dysfunctional family is chronically conflicted. The parents have not resolved who has the power to decide what, and each strives to dominate, neither one accepting a submissive role. Sharing power does not appear to be possible for such parents and, because of this chronic conflict, each family decision precipitates another round of battling. Inevitably, the children are drawn into the fray and learn that closeness is not safe and that people are to be exploited for one's own purposes.
Although very different in many aspects, these two types of families share a feature: intimacy is dangerous and must be avoided.
These diverse studies suggest that both the nature of an individual's attachment to others and his or her manner of dealing with the loss of others play crucial roles in the etiology of depressive syndromes. In particular, the studies suggest that the presence of a confidant, an individual with whom one can share deeper feelings and thoughts, or being part of a family in which high levels of expressiveness are facilitated, may diminish vulnerability to depressive disturbances. Families not providing such an interpersonal context may be seen as increasing the vulnerability of individual family members to depression.
FAMILY FACTORS INVOLVED IN RECOVERY OR RELAPSE
In a biopsychosocial approach to depressive disorders emphasizing the role of multiple, diverse, and interacting variables, patients may develop depressive syndromes as the result of the interaction of different sets of variables. Regardless, however, of why a particular individual comes to be depressed, family factors may facilitate recovery, act as an influence that sustains the depression, or increase the likelihood of relapse for the recovered person.
Two types of studies illustrate the possible role of the family in influencing the course of a depressive disorder. One type can be considered macroscopic and is represented by the work of Vaughn and Leff."1 They found that dividing the families of hospitalized patients with depressive neurosis into those high or low on critical comments about the depressed member revealed a significant difference in relapse rate. Those patients returning to critical relatives had a relapse rate of 67% over nine months in contrast to 22% in the low-criticism group. In a subsequent report, Leff and Vaughn describe the interaction of a family's expressed emotion and life events in depressed patients.19 Relapse in depressed patients was associated with the presence of both high stress and a critical spouse, and the investigators suggest that it is the interaction of these two variables that increases the probability of relapse. They raise the question of whether their measure of spouses' criticism is but another description of the lack of intimacy within the marriage.
Davenport and colleagues report a more microscopically focused study of the psychodynamic features of a small group of families with multigenerational, bipolar, manic-depressive illness.20 They emphasize the consistent psychological pattern that maintains the family member's syndrome: separation and individuation are perceived as loss, and intimacy carries with it the threat of abandonment. Rigid conformity is expected, and avoidance and denial of affect lead to a shared family climate of emotional distance and deprivation.
Other clinicians have proposed models of depression that focus on interpersonal processes. Although some define depression as an interactional (rather than intrapsychic) phenomenon, each model can be used in conjunction with variables from other enologie systems. Bonime focuses on the interpersonal relationships of the depressed individual, who is viewed as an extremely manipulative person involved unconsciously in attempts to elicit a subjectively gratifying response from others.21 Thus, sadness, helplessness, and other depressive phenomena are understood as forms of manipulation designed not to produce affection or support, but rather to reassure the depressed person that he or she can elicit a response from significant others. Bonime's model emphasizes, therefore, the need to control interpersonal relationships. He suggests that other interpersonal processes supporting this conceptualization are the patient's aversion to being influenced by others, unwillingness to gratify others, and intense, hostile competitiveness. In this model, guilt is understood as a realistic response to the ways in which the depressed person has to some degree spoiled the lives of others. The patient's manipulations (and, therefore, depression) escalate when they fail to elicit a suitable response from others.
Feldman offers a model of depression and marital interaction focusing on intrapsychic cognitive schemata, social stimulation and reinforcement, and feedback loops.22 He emphasizes the ways in which certain patterns of marital interaction may act as precipitating or sustaining factors in depressive syndromes. In this model, both the depressed individual and the nondepressed spouse are seen to have central, cognitive schemata of self-depreçiation. However, the nondepressed spouse defends against this schemata with protective, rescuing behaviors. When his or her own self-image is threatened, behavior unconsciously designed to increase the spouse's depressive symptomatology is precipitated, thereby allowing the nondepressed spouse once again to become the omnipotent rescuer. In this way Feldman's model emphasizes the interpersonal processes of the nondepressed spouse as sustaining the symptomatic spouse's depressive syndrome.
Coyne offers a model emphasizing both the behavior of the depressed person and a significant other.21 He suggests that the depressed person engages others in his environment in such a way that support is lost and "depressive information" is elicited. This increases the level of depression, strengthens the depressed behavior, and leads to more depression-evoking responses from others. Much of what is customarily viewed as the depressed patient's cognitive distortion is held by Coyne to be an accurate perception of messages from significant others. The depressed person is understood to be experiencing worthlessness and helplessness, combined with oversensitivity to messages that impinge upon these feelings. The communication of these affects is considered, however, to be a set of messages that both demand reassurance of the depressed person's place in the relationship and insist that others act to restore the patient to his or her original state. Most often these messages lead to reassurance and support from significant others, but the depressed person is unable to decide whether these responses reflect a genuine attitude or merely acquiescence to the demands for reassurance. As a consequence, the depressed person increases the intensity of depressive messages, and significant others respond with more reassurance and support. The interaction continues until those in the social context become irritated, at the same time, feeling guilty and inhibited. At some point, however, their underlying feelings will be reflected nonverbally if not verbally. They may withdraw or become openly critical; either response diminishes the support offered to the patient and encourages the development of chronicity.
Whether the critical spouses of the Vaughn and Leff studies (70% of the sample) reached that point by way of the interpersonal dynamics suggested by Bonime, Feldman, and Coyne requires empirical demonstration.
Hinchliffe and her colleagues studied the intimate social context of depressed patients and emphasize the role of marital communication.24-28 Depressed patients and their spouses reveal a variety of communication abnormalities that are not found when the depressed patient is talking with a stranger. These findings can be understood as offering some empirical validation of the type of interpersonal dynamics suggested by Bonime, Feldman, and Coyne. However, the absence of longitudinal studies demonstrating that the communication abnormalities existed prior to the development of the depressive syndrome suggest that such interpersonal dynamics may best be understood as playing a sustaining role.
THE IMPACT OF THE DEPRESSED PATIENT ON THE FAMILY
Many factors affect how a family responds to stress, among them the family's basic competence.29 Whether the stress originates within the family or outside the family, is experienced as discrete or diffuse, or is seen as acute or chronic are other important factors.
Although there are no published reports that describe the impact of a family member's depression on the family, there are a variety of studies detailing the impact of other types of stress on the family. These reports suggest that if a family's coping capacity is exceeded, the structure of the family changes, and the changes occur in an orderly sequence. Highly competent and flexible families become increasingly rigid. One family member gradually takes over and comes to play a dominant role. Other families may develop a rigid, enduring pattern of conflict manifested by arguments that may eventually lead to divorce and the family's dissolution. If the coping mechanisms associated with this rigid level of family functioning do not contain the stress, the family may then literally disintegrate and chaos may ensue.
Clearly the relationship between family variables and a family member's depressive syndrome is circular rather than linear. Families may not only predispose, precipitate, and sustain depressive disorders, but they may also bear the brunt of that disturbance. The stress of having a family member who is depressed may lead to substantial changes in the basic structure of the family. These changes often involve increasing the interpersonal distance or increasing conflict between family members. These changes, in turn, can influence adversely the family member's depressive syndrome. The circular process becomes an escalating spiral that, at some point, may reach an intensity at which suicidal ideation, threats, and behavior become the dominant transactional themes.
To return to the artistic, fictional tragedy of Willy Loman, it seems useful to speculate about what might have occurred in a different family context. If the Lomans had been a family that facilitated both individuation and autonomy, would it have made a difference? If the Lomans had been a family in which high levels of cohesiveness had evolved and the expression of a broad range of feelings was facilitated, would it have made a difference? If the Lomans had the capacity as a family to focus on their present problems rather than on Biffs longpast athletic success, would it have made a difference? If Willy and his wife had developed an intimate relationship in which fears could be shared, would it have made a difference? We cannot answer these questions about this Willy Loman, but for some Willy Lomans family variables do play a role, and their recognition offers the clinician another avenue for effective treatment.
1. Miller A: The Death of a Salesman. New York, Viking Press Inc. 1967.
2. Moos RH: Work Bivironment Manual. Palo Alto. California. Consulting Psychologists Press Inc. 1981.
3. Holahan CJ, Moos R: Development and Validation of Qualitative Indices of Social Support. Palo Alto, California, Social Ecology Laboratory, Stanford University and the Veterans Administration Medicine Cerner, 1981.
4. Holahan CJ. Moos R: Social Support and Adjustment: Predictive Benefits of Social Climate indices. Palo Alto. California. Social Ecology Laboratory. Stanford University and the Veterans Administration Medicine Center. 1981.
5. Holahan CJ. Moos R: Social support and psychological distress: A longitudinal analysis. J Abnorm Psychol 1981; 90:365-370.
6. Wetzel J: Dependence upon unsustaining environments as an antecedent variable of depression. Dissertation Abstracts International 1976; 37:5361a.
7. Wetzel J: Depression and dependence upon unsustaining environments. Clinical Social Work Journal 1978; 6:75r89.
8. Wetzel J, Redmond F: A person environment study of depression. Social Science Review 1980; 54:363-375.
9. Weissman MM, Myers JX: Affective disorders in a U.S. urban community. Arch Gen Psychiatry 1978: 35:1304-1311.
10. Lowenthal MF: Social isolation and mental illness in old age. in Neugarten BL (ed): Middle Age and Aging. Chicago, The Uoiv of Chicago Press, 1968.
11. Brown GW. Harris T: Social Origins of Depression. New York, Free Press, 1978.
12 Belle D, Corfman E: Depression and low-income, female-headed families, in Families Today, Department of Health, Education, and Welfare publication No. (ADM) 78-815, 1979. vol I. pp 79-898.
13. Lowenthal MF, Haven C: interaction and adaptation: Intimacy as a critical variable, in Neugartcn BL (ed): Middle Age and Aging. Chicago, Univ of Chicago Press. 1968.
14. Lewis JM. Beavers WR. Gossett JT, et al: No Single Thread: Psychological Health in Family Systems. New York, Brunner/ Mazel Inc. 1976.
15. Lewis JM: How's Your Family? New York, Brunner/ Mazel Inc. 1979.
16. Lewis JM: The family in the matrix of health and illness, in Hofling CH. Lewis JM feds): The Family: Treatment and Evaluation. New York, Brunner/ Mazel Ine, 1980.
17. Crabtree L: Affective psychosis and the family. Presented at Timberlawn Psychiatric Hospital. Dallas. August 13, 1981.
18. Vaughn CE. Leff JP: The influence of family and social factors on the course of psychiatric illness. BrJ Psychiatry 1976; 129:125-137.
19. Leff JP, Vaughn CE: The interaction of life events and relatives' expressed emotion in schizophrenia and depressive neurosis, Br J Psychiatry 1980; 136:146-153.
20. Davenport YG, Adland ML, Gold PW, et al: Manic-depressive illness: Psychodynamic features of multigenerational families. Am J Orthopsychiatry 1979; 49:24-25.
21. Bonime W: The psychodynamics of neurotic depression, in Arieti S (ed): American Handbook of Psychiatry. New York, Basic Books, 1966. vol 3.
22. Feldman LB: Depression and marital interaction. Family Process 1976; 15:389-395.
23- Coyne JC: Toward an interactional description of depression. Psychiatry 1976; 39:28-40.
24. Hinchliffe M. Hooper D. Roberts FJ. et al: A study of the interaction between depressed patients and their spouses. BrJ Psychiatry 1975; 126:164-172.
25. Hooper D, Roberts FJ, Hinchliffe M K. et al: The melancholy marriage: An inquiry into the interaction of depression. Part I. Introduction. Br J Med Psychol 1977; 50:113-124.
26. Hinchliffe MK. Vaughan PW, Hooper D. et al: The melancholy marriage: An inquiry into the interaction of depression. Part 2. Expressiveness. Br J Med Psychol 1977; 50:135-142.
27. Hinchliffe MK. Vaughan PW, Hooper D. et al: The melancholy marriage: An inquiry into the interaction of depression. Part 3. Responsiveness. BrJ Med Psychol 1978; 51:1-13.
28. Hinchliffe MK. Hooper D, Roberts FJ. et al: The melancholy marriage: An inquiry into the interaction of depression. Part 4. Disruptions. Br J Med Psychol 1978;51:15-25.
29. Lewis JM: The family, stress, and coping, in Stress and Coping, No. I. Philadelphia. Smith Kline & French, 1981.