Attachment Styles in Adults
John Bowlby noted that attachment patterns correlate with patterns of social and play behavior with other adults (not just the mother). This correlation is observed “in the second and subsequent years of life,” even though available scientific data at the time suggested it persisted up to ages 5–6. Today, attachment styles are also linked to adult romantic relationships, suicide risk, depression, and abusive behavior in adults, which is the focus of this research (12, 18–21). In a review of major longitudinal studies on attachment classifications, Fonagy and colleagues concluded that 68–75% of adults retain the same attachment style they had in childhood (21).
Before discussing the connection between attachment and adult abusiveness, it’s important to describe the nature of attachment styles and their correlation with what I call chronic anger in close relationships, or the “anger script.” The empirical study of specific differences in responses to attachment and separation from the attachment figure began with Mary Ainsworth. These reactions were first recorded in the “Strange Situation” experiment, where a child was separated from a parent. Based on observations, infants were categorized into three attachment types: secure, anxious-avoidant, and anxious-ambivalent.
Infants with secure attachment were happy when their mother returned, reached out to her, and clung to her. They were easily comforted and differed from other groups in how often they sought emotional contact and comfort from their mother during distress. About 62–75% of American middle-class children fall into this category (18). Caregivers of these infants are responsive, accurately interpret, and appropriately respond to their needs, providing a predictable and controlled environment that helps regulate the infant’s arousal and sense of efficacy (22). This trait, sometimes called “attunement,” means the parent mirrors the infant’s emotional state (e.g., through facial expressions and sounds). Allan Schore, in his seminal work “Affect Regulation and the Origin of the Self,” argued that proper maternal attunement during critical periods of neurological development is essential for healthy child development. Attachment processes influence the physical development of neurostructures that regulate emotions (23). Not only the development but also the integrated communication of these neurostructures—the essence of brain function—depends on healthy attachment.
The second style, anxious-avoidant attachment (also called dismissive), gives the impression of independence. These infants explore their environment without relying on their mother and don’t check to see if she’s nearby (unlike securely attached children). When separated, they show no signs of anxiety, and when the mother returns, they push her away or avoid her. They only interact with the mother when things are going well; in distress, they don’t signal or seek physical contact. Many of these behaviors persist into age six. Anxious-avoidant children don’t pay attention to their mother upon her return, physically distance themselves, feel awkward discussing separation, and turn away from family members in photos (17). This style is characterized by minimal affect or distress in the mother’s presence and avoidance of the attachment figure in situations where children with secure attachment would seek closeness. These infants focus on their environment and actively ignore the parent.
Robert Karen describes this group: “The avoidant child acts in the opposite way (compared to the ambivalent child). He gets angry and distances himself (but the degree of attachment does not decrease). His pleas for attention are rejected, which hurts, and further appeals to the mother seem unbearable. The child seems to say: ‘Who needs you—I can handle it myself.’ Often, this attitude is accompanied by a sense of grandiosity: I’m great, I don’t need anyone… Bowlby believes that the avoidant attachment style underlies narcissistic personality traits—one of the most common psychiatric deviations of our time (p. 50).”
About 32% of Ainsworth’s sample fell into this category. Mothers of anxious-avoidant children are usually unresponsive, rarely react to the child, provide little stimulation, and are averse to physical contact. They reject requests for comfort and reassurance and talk a lot to drown out emotional expressions. If this verbalization doesn’t soften the infant’s affect, it’s followed by “sadistic misattunement” (i.e., expressing displaced, distorted feelings). Such mothers often don’t remember details of their own childhood or idealize their parental relationships, despite having memories of rejection. Psychologist Kim Bartholomew observed that anxious-avoidant children’s behavior may seem like a lack of need or desire for contact, but there’s strong evidence to the contrary. For example, their heart rate increases during separation, even though they show no outward distress (27). Bartholomew concludes that their focus on inanimate objects is a form of displacement reflected in their behavior. Moreover, while anxious-avoidant children rarely show aggression in the “Strange Situation,” they may act aggressively toward their mother at home (28, 29).
The more avoidance a child shows upon reunion with the mother, the more anger and dependent behavior they display in the following weeks. This supports Bowlby’s idea that anger is protest behavior aimed at increasing closeness with the mother. Thus, anger expressed by avoidant children toward the mother in less stressful situations (usually in response to rejection or unemotional treatment) is evidence that they are not indifferent. During separation, avoidant infants react with anger toward the mother, but expressing anger in such situations is risky, as it may reduce closeness, so anger is suppressed and replaced by “cold,” detached avoidance (22, 30). Angry impulses are expressed in less stressful situations. Children facing chronic rejection feel especially strong anger but avoid expressing it. Again, in less stressful circumstances, anger is expressed indirectly. Bartholomew argues that chronic avoidance may mask a strong, unresolved approach-avoidance pattern: threats trigger a tendency to seek closeness with the attachment figure, who rejects physical contact, leading to withdrawal and an even greater need for attachment. This self-reinforcing feedback leads to chronic avoidance (presumably with a chronically unmet need for attachment). Thus, anger is central to the anxious-avoidant attachment style. This raises the question: can a specific pattern indicate the emotional origin of a distancing style where anger is suppressed? Gayla Margolin found similarities in communication styles between physically abusive and distancing couples (31). Both show low assertiveness and a tendency to avoid conflict.
The third category, anxious-ambivalent attachment (also known as preoccupied or anxious), is characterized by clinging to the mother and resisting independent exploration. These infants become highly agitated during separation, often crying intensely. They seek contact when the mother returns but arch away from her in anger and are hard to comfort. This pattern arises because anger and fear of abandonment merge. Mothers of these infants are typically inconsistent and the least confident in early caregiving (30). Later, this category was split into anxious-ambivalent and preoccupied. The first group experiences constant anxiety in close relationships; the second is torn by ambivalent impulses. Karen describes the behavior style of such children: “The ambivalent child (about 10% of middle-class American children) desperately tries to influence [his mother]. He clings to the fact that sometimes he does get a response from her. He knows she’ll react—sometimes out of guilt—if he makes enough noise. So he constantly tries to hold on to her or punish her for being unavailable. He is extremely dependent on her and on attempts to change her (11, p. 50).”
These ambivalent children are surprisingly similar to physically abusive men (32, 33). The intensity of their behavior and need to influence women resemble descriptions of abusive personalities and the power motive that Winter considers fundamental to the “Don Juan”—the sexually promiscuous man addicted to repeated patterns of conquest and abandonment (34, 35).
In a longitudinal study, Alan Sroufe showed that third graders with anxious attachment had the poorest social skills and clear psychiatric disorders (29). People with anxious attachment have a reduced ability to form a social support network and seek help when needed. Thus, a lifelong pattern of isolation may be based on early experiences that taught them not to rely on others for support. Aggressive men are typically unsociable and withdrawn. They also develop certain expectations of the opposite sex. For example, with an absent father and a demanding (but unavailable) mother, a boy learns not to expect emotional support from men, and that women seem to offer support but are too demanding and untrustworthy. As a result, he withdraws and isolates himself, left alone with simmering anger.
Failed Attachment: Developmental Psychopathology
Intuitively, we might think that if an infant is regularly abused by a parent, attachment to that person may not form. However, evidence shows that such circumstances create the strongest bonds—characterized by both closeness and suppressed anger. As Bowlby wrote, “We may assume that attack from any source causes fear and withdrawal. It is surprising that in this situation, the attack comes from the very haven of safety, and this, of course, contributes to the formation of conflicting tendencies. From this one threat or signal, the child receives two messages: ‘stay away from me’ and ‘seek safety’” (4, p. 209).
In the classic study by H. Harlow and M. Harlow, “evil surrogate mothers” blew unpleasant air at baby rhesus monkeys, poked them with needles, dropped them, or shook them so hard their teeth chattered. None of this prevented the monkeys from forming attachments; in fact, the process intensified. The authors concluded, “Instead of producing neurosis in experimental conditions, we found a technique for strengthening attachment to the mother” (36, p. 206). Essentially, Harlow recreated child abuse in the lab, alternating comfort with harmful behavior. To their surprise, attachment did not weaken but grew stronger.
The effects of maltreatment on the attachment system in human infants were studied by Patricia Crittenden and Dante Cicchetti (21, 37–40). Most studies involved families under social services’ supervision for child abuse. Abused children typically show anxious-ambivalent attachment and anger, as well as excessive closeness as compulsive compliance. The authors see this behavior as a way to maintain attachment while denying constant anger at the abusive “parent.” In other words, abused children may become hyper-controlling, abusive spouses, with the anxious-ambivalent pattern becoming the raw material for a cyclical personality. Filled with rage they cannot express, holding negative representations of themselves and women/men, yet cyclically drawn to the opposite sex by a “push-pull” pattern, anxious-ambivalent children are the prototype of the abusive adult.
A child of abusive parents is forced to form attachment to someone who both cares for and harms them. The rage felt in contact with such a parent is suppressed and remains unexpressed until similar close relationships form in adulthood. Meanwhile, as the child becomes a teenager, they focus on building a public persona. The inner chaos remains latent until close adult attachment triggers the emotional script from the original attachment experience (41–43). Abusive men and women who experienced physical abuse in their family of origin are at risk for developing anxious-ambivalent attachment (44, 45). Previously, it was thought that such “transgenerational transmission” of violence was mainly due to behavioral modeling, but attachment theory and research show that more is at play: the formation of faulty internal schemas, a damaged self-image, and expectations of attachment to others, mixed with fear and rage. There is also a lack of neural structures and integration that could help experience healthy emotions during conflict. This is how the foundations of abusiveness are laid. Childhood abuse is not just learned behavior patterns. It leads to an anxious-ambivalent style of connection, resulting in excessive demands and anger in adult romantic relationships—exactly how abused women describe their husbands, and these traits match the pathology of borderline personality disorder.
Adult Attachment and Problematic Behavior
Intriguing facts about the stability of attachment style can be found in the research of psychologist Robert Silverman (46). He argued that “the powerful unconscious desire for unity with the ‘good mother of early childhood’ and the satisfaction of this desire may promote adaptation.” Silverman and colleagues exposed subjects to subliminal stimulation (lasting 4 milliseconds) with the phrase “Mommy and I are one.” This stimulation had a positive effect on a range of behavioral problems from schizophrenia to smoking. Silverman described this technique as “activation of symbiotic fantasies (fantasies of unity),” where “representations of self and other merge and blend,” as the mother is perceived early in life as comforting, protective, and nurturing (46, p. 1297). The positive effect was stronger in men than women, suggesting gender differences. Silverman hypothesized that daughters have less need to differentiate from their mothers than sons. Experiments with “Daddy and I are one” and “My beloved and I are one” had positive effects on schizophrenia and anxiety disorders in female subjects. Silverman suggested that activating fantasies of unity reduces anxiety and satisfies dependency needs, and he confirmed this with empirical results. Although Silverman and colleagues studied groups with various adult problems, they did not record subjects’ attachment styles (40). Given that adults with insecure attachment have less experience of ideal maternal care than those with secure attachment, the positive effect of such subliminal stimulation may differ between groups.
The bridge from infant to adult attachment was finally built by the landmark work of social psychologists Cindy Hazan and Phillip Shaver, who transitioned from studying infant attachment to describing adult romantic attachment styles in their influential article “Romantic Love as Attachment Process” (18). They argued that adult romantic love has attachment properties derived from its infant forms:
“The stability of personality characteristics is primarily related to the stability of mental models created by a fairly stable family environment… We are ready to make the bolder claim that all important love relationships—especially with parents, as well as with lovers and spouses—are attachments in Bowlby’s sense. Every confirmed feature of attachment has a corresponding feature of love, and for most confirmed features of love, there is a confirmed or highly probable feature of attachment” (47, p. 73).
Using a self-report questionnaire on adult romantic attachment, they found that among over 700 adults, attachment styles were distributed almost identically to Ainsworth’s findings: 56% described their attachment as secure, 25% as anxious-avoidant, and 20% as anxious-ambivalent. Anxious-ambivalents saw love as obsession, a desire for reciprocity and unity, emotional highs and lows, and extreme sexual attraction and jealousy. They said they fell in love easily and often felt they were falling in love, but rarely found what they considered “real” love. They were more likely to doubt themselves and feel misunderstood by others.
Attachment history was assessed by respondents’ answers about how their parents generally treated them (and each other) in childhood. Anxious-ambivalent respondents described mothers as more intrusive and unfair (compared to those with secure attachment), and fathers as unfair and dangerous. The term intrusive is often used by Ainsworth and colleagues to describe mothers of anxious-ambivalent infants. Descriptions of mothers matching Ainsworth’s anxious-avoidant category indicated greater inconsistency. Hazan and Shaver’s study was a crucial first step in drawing parallels between early attachment and adult relationship types.
After this landmark study, research on adult attachment boomed. Although debates continue, four adult attachment styles are now commonly identified: secure, anxious-avoidant, anxious-preoccupied, and fearful-avoidant (48). Researchers are also interested in the so-called disorganized style (21, 49). Disorganized attachment is also characterized by strong conflicting emotions when the attachment system is activated. People with this style have semantic and syntactic difficulties in creating attachment narratives (writing about personal relationships) and tend to dissociate. While we found a link between the “fearful-avoidant style” and borderline personality organization (BPO), Fonagy and colleagues found a connection between disorganized attachment and borderline traits (21). They hypothesized that people with borderline personality histories have parents who respond to their high arousal with “withdrawal, communication errors, role confusion, negative, intrusive, or frightening behavior.” Thus, “these children perceive their own arousal as a warning that they will soon be abandoned… That is, we are dealing with a parent image who withdraws from the child in states of anxiety or rage, to which the child responds with a complementary dissociative style” (p. 113). There is still no clear research data to distinguish these two categories, except that dissociation is always present in disorganized attachment.
People with secure attachment have positive self-scripts and expectations for close relationships; they expect the best and are not anxious about intimacy. Anxious-avoidants, according to Hazan and Shaver, have “signed off” from close relationships. They are extremely independent: they need no one. The anxious-preoccupied type clings to relationships, worries about rejection, and will do anything to please and gain approval. The fearful-avoidant group is especially interesting, as they are equally drawn to and repelled by intimacy. As Kim Bartholomew writes, “They seek social contact and closeness but constantly experience distrust and fear of rejection in interpersonal relationships” (27, p. 176).
What attachment style and abusiveness have in common are the chronic feelings caused by insecure attachment, and how these feelings can transform into behavior in men who had abusive role models in childhood. Men with early attachment problems are more likely to feel anxious about regulating closeness. The arousal, anxiety, and anger they experience are rooted in deep anxiety about the original attachment figure.
Abusive men have an exaggerated need to control close relationships because their needs match their anxiety, and control is a behavior aimed at reducing anxiety/anger. These men try to reduce abandonment anxiety by excessively controlling their partner. Child psychologist Patricia Crittenden writes:
“Anxious attachment can manifest at any age. Some manifestations of anxious attachment in older children and adults resemble those in infancy: unreasonable anxiety and concern about the attachment figure’s whereabouts, unreasonable difficulty separating from them, chronic anger and rejection, inability to seek or use support when needed, or lack of feelings toward the attachment figure” (38, p. 131).
Crittenden further suggests that other attachment disorders are less common but are based on “traumatic or depriving separation from the attachment figure” (p. 133).