Loneliness within a nomological net: An evolutionary perspective

Abstract

Loneliness is characterized by feelings of social pain and isolation and has both heritable and unshared environmental underpinnings. An evolutionary theory of loneliness is outlined, and four studies replicate and extend prior research on the characteristics of lonely individuals. Studies 1 and 2 indicate that loneliness and depressed affect are related but separable constructs. Study 3 confirms that lonely, relative to nonlonely, young adults are higher in anxiety, anger, negative mood, and fear of negative evaluation, and lower in optimism, social skills, social support, positive mood, surgency, emotional stability, conscientiousness, agreeableness, shyness, and sociability. The set of six personality factors associated with loneliness (surgency, emotional stability, agreeableness, conscientiousness, shyness, and sociability) do not explain the associations between loneliness and negative mood, anxiety, anger, optimism (pessimism), self-esteem, and social support, as each association remained statistically significant even after statistically controlling for these personality factors. Study 4 used hypnosis to experimentally manipulate loneliness to determine whether there were associated changes in the participants’ personality and socioemotional characteristics. Results confirmed that loneliness can influence the participants’ personality ratings and socioemotional states.

Introduction

…by the 1970’s, it was dawning on many evolutionary biologists, especially those interested in behaviour, that evolution by natural selection was not much about competition between species, not even mostly about competition between individuals, but was about competition between genes using individuals and occasionally societies as their temporary vehicles. (Ridley, 2000, p. 127).

Loneliness has been described as a complex set of feelings that occurs when intimate and social needs are not adequately met and that “drives” individuals to seek the fulfillment of these needs (Russell et al., 1980, Weiss, 1973). Estimates by Peplau and Perlman (1982) suggest that at any one time approximately 20% of the population complains of feelings of loneliness. Why are people plagued by this plight? In the present paper, we outline an evolutionary model to address this question in which loneliness reflects the operation of selfish genes using individuals to insure their perpetuation.

Darwin’s (1959) insight was that organisms compete for resources, and those that had some advantage in a habitat would be more likely to transmit this advantage to future generations via their offspring. Building on Darwin’s seminal insight, Dawkins (1990) describes how genes evolve and function to perpetuate themselves. According to Dawkins (1990), the world of the selfish gene is generally one of savage competition, ruthless exploitation, and deceit. Even apparently collective actions that foster survival, such as fish swimming in schools, can be explained in terms of individual self-interest. When sardines are predated, they form a dynamic fish ball as a last means of defense. The emergence and unfolding of this collective action can be explained by a single, selfish rule: Swim to the middle.

The genetic constitution of species characterized by brief periods of dependency is reducible to the reproductive success of individual members of the species. Simply stated, if an organism survives to reproduce, the genes of the organism are more likely to be included in the gene pool of the species. The genetic constitution of Homo sapiens, however, derives not simply from an individual’s reproductive success but more critically from the success of one’s offspring to reproduce (Cacioppo & Hawkley, 2003). Humans are not particularly strong, fast, or stealthy relative to other species. They lost their canine teeth thousands of years ago and they never had the safety offered by natural armor or flight. It is the ability to think and use tools, to employ and detect deceit, and to communicate, work together, and form alliances that makes Homo sapiens such a formidable species.

The importance of these capacities produced a selective pressure for larger brains with greater associative, reasoning, integrative, and communicative power (Calvin, 2004). Larger brains were achieved in part by greater folding of the cortices to squash more brain matter into a limited cranial space. Larger brains were also achieved by modifications of the pelvic skeletal structure in women, which although less optimal for walking, increased the likelihood that the mother and infant would survive childbirth. The cranial skull itself evolved to be incomplete at birth. Even though the hole at the top of the skull places a newborn at greater risk for injury or death from a blow, it improves the chances of successful transport through the birth canal. And the human infant is born to a prolonged period of abject dependency. It appears part of the solution in the negotiation between a larger brain and successful transport through the birth canal is that much of development and expansion of the human brain occurs after birth.

The fact that humans are born in modal litter sizes of one and to a long period of abject dependency changes the effects of the operation of the “selfish gene.” If infants do not elicit nurturance and protection from caregivers, or if caregivers are not motivated to provide such care over an extended period of time, the infants perish along with the genetic legacy of the parents. The selfish gene therefore had to add to its repertoire the evolution of social connection and care. Ancestors who were inclined to form social connections, communicate and work together, share food and defense, and retaliate in the face of violations of reciprocity norms had a selective advantage to survive. In conditions of hardship, hunter-gatherers who had a genetic disposition to experience social pain from social separation (i.e., loneliness) may have been more likely to return to share their food, shelter, or defense with their family and allies to diminish the pain of loneliness. Individuals with no such feelings of loneliness when separated from others may have roamed the earth better nourished than those who felt distressed by social separation, but the abandoned offspring—and the genetic predisposition of the parents—would have been less likely to survive.

The present model has features in common with Bowlby’s (1969/1982) emphasis on infant attachment but the need to form attachments represents only a part of the story. Social pain in response to disconnections, the reward felt when those connections are reinstated, and individual differences in sensitivity to social pain and reward are also fundamental to the evolutionary model. Physical pain evolved as a reflex to protect the organism from tissue damage. Physical pain is influenced by situational factors but there are also stable individual differences in pain sensitivity. According to our evolutionary model (Cacioppo & Hawkley, 2003) loneliness is thought to operate in part through social pain, which co-opts the physical pain system (Eisenberger, Lieberman, & Williams, 2003), and social reward, which co-opts the appetitive system (Rilling et al., 2002), to protect the gene if not the individual. Specifically, the social pain of loneliness and the social reward of connecting with others motivate the person to repair and maintain social connections even when his or her immediate self-interests are not served by the sharing of resources or defense. In this way, loneliness, like punitive altruism (Boyd et al., 2003, Fehr and Gächter, 2002), promotes adaptive collective behavior to preserve the selfish molecules known as genes (Dawkins, 1990). Moreover, loneliness like physical pain and reward is posited to be influenced by situational factors but there are also stable individual differences in sensitivity.

Recent work by Eisenberger et al. (2003) on social pain and by Rilling et al. (2002) on social reward is in accord with the notion that loneliness may have evolved to use the physical pain and reward systems to extend its protective function to include others. In a recent functional magnetic resonance imaging (fMRI) study in which participants were excluded from or included in a social situation (i.e., a ball tossing game), for instance, Eisenberger et al. (2003) found neural activation localized in a dorsal portion of the anterior cingulate cortex (ACC) that is implicated in the affective component of the pain response. The authors argue that the similarity in activation of the dorsal ACC to physical pain (e.g., Rainville, Duncan, Price, Carrier, & Bushnell, 1997) and to social pain suggests that the experiences of physical and social pain may share a common neuroanatomical basis. Furthermore, Eisenberger and her colleagues suggest that “Because of the adaptive value of mammalian social bonds, the social attachment system… may have piggybacked onto the physical pain system to promote survival” (p. 291).

According to our evolutionary model, a person made to feel lonely also feels unsafe and activates an anachronistic survival mechanism that heightens sensitivity to threats from all sides. Lonely individuals are not simply unhappy, they have heightened sensitivity to threats and attacks. Defensive behaviors such as preventative rejection of others may help fend off treachery, rejections or attacks. It is this ingrained, self-protective focus on threats that drives their anxiety and caustic interactions and allows them to minimize the short term damage of negative interactions but at the cost of potentially self-defeating hostility, fault finding, and blaming (Cacioppo and Hawkley, 2005, Rotenberg, 1994). Consistent with this model, prior research has shown loneliness to be related to a constellation of socioemotional states and traits that include poor mood, anxiety, anger, optimism, low social support, dysphoria, and lower self esteem (e.g., see reviews by Berscheid and Reis, 1998, Duck et al., 1994, Ernst and Cacioppo, 1999, Rook, 1984, Shaver and Brennan, 1991).

If the motivation to form and maintain social bonds has evolutionary origins, one might expect significant genetic contributions to loneliness. The first investigation of the heritability of loneliness was published by McGuire and Clifford (2000). In their first study, 69 biologically related sibling pairs and 64 unrelated pairs in adoptive families in the Colorado Adoption Project completed an 8-item loneliness scale when they were 9, 10, 11, and 12 years of age. In a second study, 22 monozygotic (MZ) twins, 40 dizygotic (DZ) twins, and 80 full-siblings 8–14 years of age completed a 16-item scale to assess loneliness in relation to their schoolmates. Results revealed significant genetic (h² = 55 and 48%, respectively, in Studies 1 and 2) and unshared environmental contributions to individual differences in loneliness.

The adoption studies of McGuire and Clifford (2000), however, can be explained in terms of an alternative model that attributes the resemblance in loneliness between relatives to shared environment. Heritability estimates of complex traits such as loneliness may also change across the lifespan, as the frequency, duration, and range of exposure to environmental influences accrues. We, therefore, extended the work of McGuire and Clifford (2000) using data from the Netherlands Twin Register Study (Boomsma, Willemse, Dolan, Hawkley, & Cacioppo, 2005). Data on loneliness from 7665 young adult and adult Dutch twins (average age 24 years) were analyzed with genetic structural equation models, which provide estimates of the shared environmental and unique environmental contributions as well as the genetic contributions. The estimate of genetic contributions to variation in loneliness in adults was 47%, with the remaining variance explained by unique environmental factors. Thus, the heritability estimates in adults were similar to that found previously in children. No evidence was found for sex or age differences in genetic architecture or for nonadditive genetic effects.

In sum, loneliness is posited to be heritable for several reasons. First, the separation from those with whom an individual has formed connections produces social pain (loneliness) that, like physical pain, motivates the person to repair the social disconnections even when his or her immediate self-interests are not served by the sharing of resources or defense. Furthermore, social acceptance and affirmation is posited to be rewarding. That is, social pain leads to negativity, lowered feelings of control, depressed affect, and lower self-esteem, whereas social cooperation and acceptance leads to increased feelings of control, positive affect, and higher self esteem. Thus, loneliness and depressed affect should be closely related. In addition, though, those who feel socially isolated also feel unsafe, which in turn activates a constellation of socioemotional processes including threat vigilance, stress, mood, anxiety, anger, and low social support. Depressed affect does not trigger the same constellation (e.g., Kremen, 1996). Thus, loneliness and depressed affect are conceived as related but separable constructs. In Studies 1 and 2, we attempted to evaluate whether the two constructs could be statistically disentangled. A common factor analysis of the items in a depressed affect inventory and a loneliness measure was conducted using data collected from two samples—a study of 2525 young adults (Study 1), and a population-based sample of 229 English-speaking Blacks/African Americans (37 males, 44 females), non-black Hispanics (33 males, 33 females), and non-Hispanic Caucasians (39 males, 43 females) between the ages of 50 and 68 years (M = 57.5, SD = 4.5) from the Chicago Health, Aging, and Social Relations Study (CHASRS).

An assumption built into studies that compare lonely versus nonlonely groups or that limit analyses to linear correlation/regression is that individuals average in loneliness are as similar to individuals low in loneliness as they are to those high in loneliness. However, prior research suggests that loneliness scores are positively skewed with most individuals reporting low loneliness (e.g., Russell et al., 1980; see Shaver & Brennan, 1991). Given this distribution, it is possible that individuals average in loneliness, whose loneliness ratings are more similar to those of others in the high frequency portion of the distribution (i.e., those low in loneliness), may be more similar to individuals low than high in loneliness on related socioemotional scales and interpersonal behaviors, as well. Such a result is entirely consistent with the emphasis in prior research on individuals high in loneliness as a psychological problem. If there is a threshold for feelings of loneliness, then the association between loneliness and states such as anxiety and depressed affect may be characterized by quadratic as well as linear components. Such a step-function might be expected if loneliness reflects an inhibited temperament (Cole et al., 1997, Kagan et al., 1994). Therefore, we tested for quadratic as well as linear effects in Study 3. Another aim of Study 3 was to investigate whether the association between loneliness and the socioemotional states of mood, anxiety, anger, optimism, fear of negative evaluation, social skills, social support, dysphoria, and self esteem was secondary to one or more personality dimensions such as surgency (extraversion), emotional stability (neuroticism), or shyness.

Finally, the evolutionary model casts loneliness as capable of influencing a constellation of socioemotional states. If those who feel socially isolated not only feel unhappy but also unsafe, then changes in loneliness should activate a constellation of socioemotional processes. To investigate this question, we used hypnosis to manipulate feelings of loneliness in a within-participants design, and we measured the effects of this manipulation on states and traits investigated in Study 3. By doing so, we were able to investigate the possible causal role of loneliness in social and emotional processes.