Nature did not construct human beings to stand alone. . . . Those who have never known the deep intimacy and intense companionship of happy mutual love have missed the best thing that life has to give. Love is something far more than desire for sexual intercourse; it is the principal means of escape from loneliness which afflicts most men and women throughout the greater part of their lives. (Russell 1929, 122–123)
To shed light on Bertrand Russell's proposition that love is the principle means to escape from loneliness, this entry will examine the links between loneliness and the family. In thinking about loneliness in a family and life cycle perspective, several questions come to mind. What is the relationship between marriage and loneliness? Is loneliness passed from parents to their children and, if so, how? From birth to death are there predictable fluctuations in loneliness due to parents and their children's life stages? Is it true, as is frequently depicted, that the loss of intimate relationships leads to loneliness? Since the mid-1970s social scientists have published a growing number of studies addressing these questions (Ernst and Cacioppo 1999).
Concept and Prevalence
Contemporary social scientists have defined loneliness as the unpleasant experience that occurs when a person's network of social relationships is deficient in some important way, either quantitatively or qualitatively (Peplau and Perlman 1982, p. 4). According to this conceptualization, loneliness stems from a discrepancy between the level of social contact a person needs or desires and the amount she or he has. The deficits can be in the person's intimate relationships, as Russell's quote implies, leading to emotional loneliness or in the individual's broader network of relationships leading to social loneliness (Weiss 1973). In either case, loneliness is a subjective experience—people can be alone without being lonely or lonely in a crowd.
Loneliness is widely prevalent. Although loneliness appears to occur in virtually all societies, its intensity varies by culture. In an eighteen-country survey (Stack 1998), the United States was in the top quarter of countries in terms of average levels of loneliness. Perhaps this in part reflects the individualistic, competitive nature of life in the United States. Individuals in European social democracies such as the Netherlands and Denmark were least lonely. Sociologists have associated national differences in loneliness with differences in social integration. The Dutch, for example, are socially well-integrated in terms of having more people in their social networks, such as being involved in civic organizations and volunteer work and receiving emotional support.
Loneliness and Marriage
One cultural universal found in a multinational study (Stack 1998) was that married men and women are less lonely than their unmarried counterparts. Cohabitation also buffered individuals from loneliness but not as much as marriage. When the unmarried are categorized into subgroups (never married, separated or divorced, widowed), the results vary somewhat by study. The general tendency appears to be for single people to be less lonely than the divorced or widowed (Perlman 1988, Table 3). In at least one Dutch study, single parents were also a group high in loneliness. Overall, loneliness seems to be more a reaction to the loss of a marital relationship rather than a response to its absence.
Differences in loneliness as a function of marital status can be explained either in terms of selection or what marital relationships provide. If selection is operating it means that the people who marry are different and would avoid loneliness even in the absence of getting married. This explanation is difficult to definitively test, although it is challenged to some extent by the relatively low levels of loneliness among never married respondents. The second view implies that the more the marital relationships provide, the less lonely the partners should be. Consistent with this explanation, low marital satisfaction is associated with greater loneliness. Similarly, compared with individuals who confide in their spouses, married individuals who talk most openly about the joys and sorrows of their lives with somebody besides their spouse are more prone to being lonely. One can conclude from the evidence that when marriages are working well, they provide partners with ingredients that buffer them from loneliness.
Parents, Children, and Loneliness
Social scientists frequently debate questions of heredity versus the environment. In the origins of loneliness, both appear to have a role. Consistent with there being an inherited component to loneliness, in a 2000 study (McGuire and Clifford 2000) both siblings and twins had some similarity in their levels of loneliness, but the similarity was greater for identical twins than for either fraternal twins or singleton siblings.
Researchers have also checked for an association between parents and their children in the likelihood of being lonely. Working with older parents (85 or older) and their mid-life children, M. V. Long and Peter Martin (2000) did not find evidence of intergenerational similarity. In contrast, J. Lobdell and Daniel Perlman (1986) administered questionnaires to 130 female undergraduates and their parents. As expected, they demonstrated that the parents' loneliness scores were modestly correlated with those of their daughters. Of course, such an association could be explained by either genetic or environmental factors.
To explore possible psychosocial factors, Lobdell and Perlman also had the university students in their study rate their parents' marriages and childrearing practices. Lonely students depicted their parents as having relatively little positive involvement with them. This is one of several studies showing the cold, remote picture of parent-child relations reported by lonely young adults. They also saw their parents as having lower than average marital satisfaction. This finding compliments other studies showing that children whose parents divorce are at risk for loneliness, especially if the divorce occurs early in the child's life. These findings can be interpreted within an environmental framework. In sum, the work on the origins of loneliness suggests that both genetic and family factors each play a role in levels of loneliness, although nonfamilial environmental influences are likely also critical.
The parental contribution to children's loneliness is not simply a one-time input. Instead, loneliness bidirectionally intertwines with parent-child relations over the life-cycle. A first noteworthy lifespan phenomenon is that in the transition to parenthood, women who are lonely during their pregnancy are at higher risk for postpartum depression.
In infancy, children are highly dependent upon their parents and caretakers. As they get older, peer relations become more important. Along with this shift comes a shift in what type of relations are most closely linked with loneliness. In the middle elementary years, it is the quality of children's relationships with their mothers. In late adolescence, it is the quality of university students' relationships with their peers.
Concerning more mature children, Pauline Bart (1980) has analyzed how children's leaving home affects middle-aged mothers. She concluded that women who adopt the traditional role of being homemakers devoted to their children are prone to experience greater loneliness and depression when their children leave home than are women less invested in a maternal, homemaker role.
For many people, one perceived benefit of having children in the first place is the notion that they will provide comfort and support in old age. As far as loneliness goes, there is evidence challenging this view. Tanya Koropeckyj-Cox (1998) looked at older adults with and without children. Contrary to common belief, the results didn't show a clear advantage of having children. A second line of research has examined whether family or friends are more strongly associated with avoiding loneliness in old age. Martin Pinquart and Silvia Sorensen's (2001) meta-analysis, a technique for statistically combining the results of several studies, shows the primary role of friends as opposed to family members in buffering seniors from loneliness.
Relationship Endings and Loneliness
Having examined separation and loss in parent-child relationships, what happens when these phenomena occur in romantic relations? As young adult dating relationships end, presumably both partners experienced a decline in the social aspects of their lives. But in many couples, one person initiates the breakup whereas the other is "left behind." Charles Hill, Zick Rubin, and Letitia Peplau (1976) found that the initiators suffered significantly less loneliness than the partners who were spurned. Perhaps having control over such life changes helps reduce the distressing effects of loosing a partner.
After their young adult dating experiences, many individuals marry and eventually end those unions via divorce. In one study (Woodward, Zabel, and Decosta 1980) fifty-nine divorced persons were asked when, and under what circumstances, they felt lonely. For these respondents, the period of greatest loneliness occurred before (rather than after) the divorce decree became final. Both ex-husbands and ex-wives felt lonely when they felt out of place at a particular social event or excluded by others. For ex-wives, loneliness was also triggered when (1) they wanted to join an activity but were unable to do so; (2) they had no one with whom to share decision-making responsibilities and daily tasks; (3) they felt stigmatized by being divorced; and (4) they had financial problems.
A University of Tulsa study involving seventy-four men and women compared the divorce experiences of lonely versus nonlonely individuals. Lonely individuals blamed more of the marriage's problems on their former spouse. They also had more difficulties in their relationships with their ex-partners. They argued more over childrearing, felt less affection, and had less friendly interactions. In terms of adjusting to separation, lonely respondents drank more, experienced greater depression, and felt more cut-off from their friends. They spent more time with their children and were less likely to become romantically involved with a new partner.
For many North Americans, marriage lasts "till death do us part." If relationships end via death of a spouse, U.S. Census data show a 5 to 1 sex ratio with women predominantly being the individual left widowed. Helena Lopata (1969) has identified several ways that widows miss their husbands. For example, when their spouse dies, women lose a) a partner who made them feel important; b) a companion with whom they shared activities; c) an escort to public encounters as well as a partner in couple-based socializing; and d) a financial provider who enabled them to participate in more costly activities and enjoy a more expensive lifestyle. With such losses, it is not surprising that loneliness is a major problem in bereavement.
Robert Hansson and his associates (1986) found a general tendency for greater loneliness to be associated with a maladaptive orientation toward widowhood. Prior to the death of their husbands, the lonely widows engaged in less behavioral rehearsal (e.g., finding jobs, getting around on their own) for widowhood and instead engaged in more rumination about the negative consequences of their spouse's impending death. At the time of their spouse's death, subsequently lonely widows experienced more negative emotions and felt less prepared to cope. Lonely widows were also less likely to engage in social comparison with widowed friends.
If a spouse dies unexpectedly, loneliness is especially pronounced. To overcome loneliness, widows typically turn to informal supports (e.g., friends, children, and siblings) as opposed to formal organizations or professionals (e.g., their church, psychotherapists). In widowhood as in other transitions, time heals: Feelings of loneliness are greatest shortly after the loss of a spouse but decline over the months and years. As widows continue their lives, the quality of their closest friendship is more likely to be associated with their experiences of loneliness than is the quantity or quality of their quality of their closest kin relationship (Essex and Nam 1987).
In sum, the findings from contemporary social science research indicate that married individuals are less likely to be lonely. However, the picture is more complex than Russell's simple suggestion that love, at least as provided by marital and kin relations, provides a surefire escape from loneliness. At some ages and positions in life, kin relationships appear to be a less important aspect of the loneliness equation than friendships or other factors. Parents not only protect their children from being lonely but also they contribute to it. If siblings are close, they tend to be less lonely (Ponzetti and James 1997). Throughout adulthood, unsatisfying marriages and the endings of intimate relationships are associated with greater loneliness. Thus, it is not simply relationships but what happens in them that counts.
bart, p. (1980). "loneliness of the long-distance mother." in the anatomy of loneliness, ed. j. hartog, j. r. audy, and y. a. cohen. new york: international universities press.
ernst, j. m., and cacioppo, j. t. (1999). "lonely hearts: psychological perspectives on loneliness." applied and preventive psychology 8:1–22.
essex, m. j., and nam, s. (1987). "marital status and loneliness among older women: the differential importance of close family and friends." journal of marriage and the family 49:93–106.
hansson, r. o.; jones, w. h.; carpenter, b. n.; and remondet, j. h. (1986). "loneliness and adjustment to old age." international journal of aging and human development 24:41–53.
hill, c. t.; rubin, z.; and peplau, l. a. (1976). "breakups before marriage: the end of 103 affairs." journal of social issues 32:147–168.
koropeckyj-cox, t. (1998). "loneliness and depression in middle and old age: are the childless more vulnerable?" journal of gerontology: series b: psychological sciences and social sciences 53b:s303–s312.
lobdell, j., and perlman, d. (1986). "the intergenerational transmission of loneliness: a study of college females and their parents." journal of marriage and the family 48:589–595.
long, m. v., and martin, p. (2002). "personality, relationship closeness, and loneliness of oldest old adults and their children." journal of gerontology: series b: psychological sciences and social sciences 55b: p311–p319.
lopata, h. z. (1969). "loneliness: forms and components." social problems 17:248–261.
mcguire, s., and clifford, j. (2000). "genetic and environmental contributions to loneliness in children." psychological science 11:487–491
peplau, l. a., and perlman, d., eds. (1982). loneliness: asourcebook of current theory, research, and therapy. new york: wiley.
perlman, d. (1988). "loneliness: a life span, developmental perspective." in families and social networks, ed. r. m. milardo. newbury park, ca.: sage.
pinquart, m., and soerensen, s. (2001). "influences on loneliness in older adults: a meta-analysis." basic and applied social psychology 23:245–266.
ponzetti, j. j., jr., and james, c. (1997). "loneliness and sibling relationships." journal of social behavior and personality 12:103–112
russell, b. (1929). marriage and morals. new york: liveright.
stack, s. (1998). "marriage, family and loneliness: a cross-national study." sociological perspectives 41: 415–432.
weiss, r. s. (1973). loneliness: the experience of emotional and social isolation. cambridge, ma: mit press.
woodward, j. c.; zabel, j.; and decosta, c. (1980). "loneliness and divorce." journal of divorce 4:73–82.
Social scientists agree that loneliness stems from the subjective experience of deficiencies in social relationships and that these deficiencies are unpleasant, aversive, and exceptionally common. Objectively deficient social relationships (i.e., social isolation) do not necessarily correspond with feeling lonely. Thus, it is common to appear alone but not feel lonely, and to feel lonely within a seemingly rich social relationship network. This potential paradox highlights an important distinction between quantitative and qualitative aspects of social relationships.
Theoretical perspectives on loneliness differ concerning the nature of loneliness, whether loneliness stems from internal or situational causes, and where such causes occur developmentally. Psychoanalysts view loneliness as a pathological result of internal factors rooted in childhood. Sociologists view loneliness as a normative event stemming from societal influences that occur throughout development. From cognitive perspectives, loneliness occurs when people perceive discrepancies between their desired and actual patterns of relationships, with desired patterns stemming from previous relationships and comparisons of one’s own relationships to those of similar others (i.e., social comparison). Although different perspectives contribute uniquely to our understanding, the cognitive perspective serves as the dominant model for studying loneliness.
The measurement of loneliness depends inherently on theoretical conceptualizations of the construct. Unidimensional views of loneliness posit a common core of experience that varies in intensity regardless of the antecedents or causes of feeling lonely. Many unidimensional scales exist, but the UCLA Loneliness Scale is the most commonly used. This measure assesses loneliness via self-report without using the terms lonely or loneliness, thereby reducing social desirability influences. Extensive psychometric work indicates that it is a reliable and valid measure of loneliness.
Multidimensional measures involve assessments of perceived quality and quantity of social interaction across multiple domains such as romantic, friendship, family, and community relationships. Multidimensional views of loneliness have become more common in the research literature, spurring intensive psychometric work on these types of scales. Two multidimensional views of loneliness have received considerable research attention. The first breaks down loneliness into distinct components that reflect stable enduring traits or transitory states tied strongly to situation/context. Studies examining these components using specialized scales typically yield test-retest correlations that are higher (indicating greater stability) for measures assessing trait loneliness than for measures assessing state loneliness, suggesting that this distinction is valid. A second multidimensional view involves a distinction between social loneliness that results from a lack of relationships that provide a sense of belonging (e.g., friendships) and emotional loneliness that occurs when people lack relationships that foster deep connection or feelings of attachment (e.g., romantic relationships). Many scales designed to measure social and emotional loneliness exist, and research findings suggest that this distinction is also a valid one.
Studies examining associations between personality characteristics and loneliness consistently show that extroverted people report less loneliness, whereas highly neurotic people often feel lonely. Low self-esteem, shyness, and pessimism also correspond to higher levels of loneliness. It remains unclear whether these personality traits lead to loneliness by limiting social contact and preventing the formation and maintenance of quality relationships, whether feeling lonely biases self-assessment of personality, or whether personality predisposes one to develop few relationships, and the subsequent lack of relationships reciprocally influences one’s personality. There remains a need for continued research that delineates causal paths among personality characteristics and loneliness.
Most children understand that being alone does not necessarily mean one is lonely, and that people can feel lonely when they do not appear to be alone. Adolescents experience more loneliness than do other individuals due to necessary restructuring of social groups to include friendships and other social relationships outside of the family during transitions to elementary, junior high, and high school. Family environments also influence the development of relationships in that children surrounded by parental conflict exhibit social anxiety and avoidance that contribute to loneliness.
Young adults face many contextual events, such as moving out of the home or going away to college, that require the development of new social ties. During these transitions, interpersonal difficulties may hinder the development of new social relationships, leaving one feeling lonely. Shy people who maintained a few high-quality relationships during high school may suddenly feel very lonely away at college when their shyness interferes with opportunities to make new friends. Although these types of life transitions often foster feelings of temporary loneliness that subside over time for most people, some individuals remain chronically lonely, suggesting that interpersonal difficulties or personality characteristics contribute to feelings of loneliness across the life span.
During adulthood, contextual transitions including college graduations and the establishment of careers present challenges to existing social networks and the need to form new relationships. New obstacles include individualistic or competitive work environments that make the formation and maintenance of satisfying relationships difficult. As adults, people place less emphasis on friendships than on intimate or romantic relationships. Although intimate relationships provide protection from loneliness, relationship quality is vital as adults in strained or unsatisfying relationships often report feeling lonely.
Elderly individuals face a number of challenges to maintaining their social networks, including the loss of relationships with coworkers through retirement, reduced contact with adult children, and the deaths of spouses or friends. Decreased functional mobility, cognitive impairment, and physical illness strain existing relationships and impede the establishment of new relationships. Despite these challenges, the elderly are less lonely on average than are college students, and increases in loneliness occur only among individuals eighty years and older. Married men and women report less loneliness than do elderly widows and widowers and men and women who are divorced or never married. Spouses in elderly couples provide functional support in addition to companionship and emotional connection, suggesting multiple ways in which marriage serves to protect against feelings of loneliness. For those without spouses, friendships with similar others provide more protection against loneliness than do relationships with adult children and neighbors.
Consistent links between loneliness, life satisfaction, and anxiety exist, and loneliness is associated with depression independently of age, gender, physical health, cognitive impairment, network size, and social activity involvement.
In addition, loneliness influences well-being and feelings of hopelessness independently of associations with social isolation and perceived social support. Loneliness also relates to physical health, as evidenced by its consistent associations with alcohol abuse, admission of the elderly to nursing homes, suicide, and mortality.
Although loneliness uniquely influences physical health, potential causes for these connections have received varying degrees of support. One view posits that loneliness affects health through maladaptive behaviors including smoking, drinking, poor exercise habits, and substandard dietary practices; however, lonely and nonlonely people rarely differ in the frequency of such behaviors. Alternative models argue that loneliness influences physical reactions to stress including cardiovascular activation, cortisol production, immunocompetence deficiencies, and sleep disruptions that link directly to development of cardiovascular disease, susceptibility to disease and infection, and diminished restorative processes that maintain overall resilience. Emerging findings provide initial support for these links, suggesting promising avenues for future research.
Strategies for coping with loneliness include changing actual relationships, expectations about relationships, or reducing the importance of relationships. Attempts to change one’s social relationships are active coping strategies wherein feelings of loneliness motivate people to form new relationship ties. Changing expectations about social relationships involves cognitive restructuring of how people view the social relationships of others. Attempts to reduce the importance of social relationships or engage in diversionary activities are passive coping strategies that often do little to alleviate loneliness.
Researchers have begun to explore the success of intervention programs in reducing feelings of loneliness, and promising findings have emerged. Social skills training for children and young adults provide the tools needed to effectively initiate, develop, and maintain satisfying social relationships. College orientation, mentoring, and buddy-pairing programs provide social contact with similar others in hopes of fostering relationship development during a transitional period when loneliness is quite common. Finally, interventions that effectively reduce loneliness among older adults target specific groups (e.g., divorcées or widows) and provide social contact opportunities with similar others as well as information that is useful for maintaining established social relationships.
Brashears, Matthew E., Miller McPherson, and Lynn Smith-Lovin. 2006. Social Isolation in America: Changes in Core Discussion Networks over Two Decades. American Sociological Review 71 (June): 353–375.
Peplau, Leticia A., and Daniel Perlman, eds. 1982. Loneliness: A Sourcebook of Current Theory, Research, and Therapy. New York: Wiley-Interscience.
Russell, Daniel W. 1996. The UCLA Loneliness Scale (Version 3): Reliability, Validity, and Factor Structure. Journal of Personality Assessment 66 (1): 20–40.
Russell, Daniel W., Carolyn E. Cutrona, Arlene de la Mora, and Robert B. Wallace. 1997. Loneliness and Nursing Home Admission among Rural Older Adults. Psychology and Aging 12 (4): 574–589.
Steptoe, Andrew, Natalie Owen, Sabine R. Kunz-Ebrecht, and Lena Brydon. 2004. Loneliness and Neuroendocrine, Cardiovascular, and Inflammatory Stress Responses in Middle-Aged Men and Women. Psychoneuroendocrinology 29: 593–611.
Weiss, Robert S. 1973. Loneliness: The Experience of Emotional and Social Isolation. Cambridge, MA: MIT Press.
W. Todd Abraham
Daniel W. Russell
Definition and theory
Loneliness is an affective emotional condition experienced when a person feels apart from familiar social supports. It is a psychosocial condition that is differentially experienced within different cultures. Most developments in this condition have been observed in the United States and Europe. Among those not in institutions, studies show that less than 20 percent of older persons experience loneliness. There is some evidence to suggest that loneliness increases with increasing totality of institutionalization. Over the life span, loneliness seems to vary curvilinearily by age. It is highest among adolescents, declines into late middle age, then increases again with advancing older age.
Theoretical conceptualizations of loneliness can be categorized as: 1) the social needs approach, with foundations in the social developmental approaches and the social support perspectives; 2) the behavioral-personality approach; and 3) the cognitive processes approach. Common to these approaches are three points. Loneliness is: 1) a subjective emotional experience that may be unrelated to actual social isolation, that is, aloneness; 2) an aversive psychological condition; and 3) caused by some form of social relationship deficit.
Measurement instruments to assess loneliness include scales developed by Russell, Peplau, and Cutrona (1980), van Tilburg and de Jong Gierveld (1999), and Vincenzi and Grabosky (1989).
The loneliness related to emotional isolation results from the absence of a person with whom one is emotionally connected. The loneliness experienced is a psychological state characterized by feelings of loss, distress, separation, and isolation. Loneliness resulting from social isolation is related to a person's perceived isolation from those around him or her. The emotional condition of loneliness in this regard is influenced by a deficit in the quantity of relationships, and/or the lack of relatedness to the social environment.
Loneliness and selected factors
Background issues related to loneliness include the following:
- Gender. Gender is a more consistent predictor of loneliness than is age. Studies show that either gender has no effect on loneliness, or that women are more lonely than men.
- Race and ethnicity. Race and ethnicity have not been systematically examined with regard to loneliness. Cross-racial, or cross-ethnic, comparisons of loneliness and its antecedents have not yet been conducted in a manner that lends any clarity to interpretation.
- Urban/rural residence. It is commonly held that urban elders are more lonely and isolated than their rural counterparts, though research has not consistently confirmed this stereotype.
- Health. The overall weight of the evidence points to a reasonably strong and consistent association between poorer physical and/or mental health, and greater loneliness.
Interpersonal relationships also factor into an individual's potential loneliness.
- Spouse. Results indicate greater loneliness in the absence of a mate. Severe loneliness appears to be unusual among married men, somewhat more prevalent among married women, and quite prevalent among unmarried individuals of either sex.
- Children. Studies of the relationship between adult children and loneliness show conflicting results. Most have found no association between frequency of contact with children and loneliness. The commitment in the relationship seems to be more important than the actual contact.
- Friends. Research shows that close friends exert a positive influence on the emotional well-being of older persons. Older persons who have contact with their friends, and especially those who are satisfied with these relationships, are less lonely.
The two essential aspects of loneliness, i.e., the loneliness associated with social isolation and/or with emotional isolation, shows that they can be experienced as an affective emotional experience in which a person feels apart from other persons and from familiar support networks. In turn, this can lead to a realization that social contacts are diminishing, lacking, or not at a level that are emotionally supportive or satisfying.
Larry C. Mullins
See also Depression; Social Support.
Andersson, L. "A Model of Estrangement—Including a Theoretical Understanding of Loneliness." Psychological Reports 58 (1986): 683.
Hall-Elston, C., and Mullins, L. "Social Relationships, Emotional Closeness and Loneliness among Older Meal Participants." Social Behavior and Personality 27 (1999): 503.
Johnson, D., and Mullins, L. "Growing Old and Lonely in Different Societies: Toward a Comparative Perspective." Journal of Cross-Cultural Gerontology 1 (1987): 257.
Marangoni, C., and Ickes, W. "Loneliness: A Theoretical Review with Implications for Measurement." Journal of Social Psychology 116 (1989): 269.
Peplau, L., and Perlman, D., ed. Loneliness: A Sourcebook of Current Theory, Research and Therapy. New York: Wiley, 1982.
Russell, D.; Peplau, L.; and Cutrona, C. "The Revised UCLA Loneliness Scale: Concurrent and Discriminant Validity Evidence." Journal of Personality and Social Psychology 39 (1980): 472.
Van Tilburg, T., and de Jong Gierveld, J. "Cesuurbepaling van de Eenzaamheidsschall [Cutting Scores on the De Jong Gierveld Loneliness Scale]." Tijdschrift voor Gerontologie en Geriatrie 30 (1999): 158.
Vencenzi, H., and Grabosky, F. "Measuring the Emotional/Social Aspects of Loneliness and Isolation." In Loneliness: Theory, Research, and Applications. Edited by M. Hojat and R. Crandall. Newberry Park, Calif.: Sage, 1989. Pages 257–270.
Weiss, R. Loneliness: The Experience of Emotional and Social Isolation. Cambridge, Mass.: MIT Press, 1973.