Childhood Emotional Abuse and the Attachment System Across the Life Cycle: What Theory and Research Tell Us

Abstract

The purpose of this article is to review the theoretical and empirical literature regarding the normative development of the attachment system from infancy through adulthood, and then discuss deviations from the normal developmental pathways that occur in response to emotionally abusive parenting (e.g., strong rejection, intrusive or controlling, hostile, or frightening behavior). A theoretical model grounded in attachment theory is presented describing the development of maladaptive interaction patterns in adult romantic relationships. The model proposes that early emotional abuse engenders insecure attachment, which impairs emotional regulation, fosters negative views of self and others that support maladaptive coping responses, interferes with social functioning and the capacity for intimate adult attachments, contributes to poor mental health, and consequently shapes the quality of romantic relationships.

KEYWORDS:

• attachment development
• childhood emotional abuse
• romantic relationships

A large body of literature documents the harmful effects of physical abuse, sexual abuse, and neglect. Scholarly interest in emotional abuse ¹ as a distinct form of maltreatment emerged later and has grown slowly. Although frequently co-occurring, the latest findings indicate that relative to other forms of maltreatment (i.e., physical abuse, sexual abuse, physical neglect), emotional abuse has specific and independent consequences (Briere & Runtz, 1990; Gross & Keller, 1992; Mullen, Martin, Anderson, Romans, & Herbison, 1996; Nicholas & Bieber, 1996) and might, in fact, be the most pervasive and damaging type of abuse (Hart, Binggeli, & Brassard, 1997; McGee & Wolfe, 1991). Despite recent progress in the field, rarely has the extant research been driven by an overarching developmental theory (Wright, 2007). Attachment theory (Bowlby, 1969, 1973, 1980) is uniquely suited to fill this gap because it is built on the salient themes of early parent–child interactions and close interpersonal relationships throughout the life span. Moreover, attachment theory is now one of the most sophisticated and extensively researched theories of human development, which provides a comprehensive account of both normal and abnormal development. The purpose of this article is to examine the phenomenon of childhood emotional abuse from the framework of attachment theory and construct a theoretical model of development that accounts for the impact of childhood emotional abuse on romantic relationship functioning in adulthood.

Figure 1 presents a conceptual model grounded in attachment theory describing the intergenerational cycle of childhood emotional abuse in the absence of intervention. The model proposes that emotional abuse by attachment figures in infancy and early childhood contributes to the development of insecure attachment organization, which impairs emotional regulation and fosters negative internal working models of self and others that initiate and sustain negative coping responses. As the child matures, these deficits create substantial vulnerability for maladaptive interpersonal schemas that interfere with social functioning, contributing to poor peer relations in childhood and adolescence (see Taussig & Culhane, 2010), and later insecurity in the adult romantic attachment system. Maladaptive coping, social problems, and insecure attachment, in turn, are risk factors for psychological distress at all ages (see Riggs & Kaminski, 2010). These pathways of influence culminate in romantic relationship dysfunction that may be characterized by problems with intimacy (see Carbone, 2010; Dodge Reyome, in press; Dodge Reyome, Ward, & Witkiewitz, in press), caregiving and caretaking (see Dodge Reyome, in press; Dodge Reyome et al., in press), sexual functioning (see Zurbriggen, Gobin, & Freyd, in press), conflict resolution (see Berzenski & Yates, in press; Dodge Reyome, in press), or relational aggression (see Berzenski & Yates, in press; Riggs & Kaminski, 2010). Without intervention at some point, the cycle may repeat itself in the next generation when adult attachment insecurity, emotional disturbance, and poor marital relations negatively affect parenting behavior, potentially resulting in emotionally abusive behavior toward offspring.

FIGURE 1 Hypothesized hierarchical regression models for relationship adjustment, psychological aggression, and psychological victimization.

This article reviews the theoretical and empirical literature that supports the proposed model, describing the ways childhood emotional abuse can undermine the adaptive development of the attachment system and lead to difficulties in intimate relationships in adulthood. The first major section describes the basic tenets of attachment theory as it applies to early development, reviews attachment research that can inform ongoing efforts to define emotional abuse, and then explores the potential influence of emotional abuse on attachment organization in infancy and toddlerhood. The subsequent section discusses the continuity of the attachment system and examines the effects of emotional abuse in childhood and adolescence. Next, the literature regarding adult attachment and mental health is reviewed and the expected impact of emotional abuse on romantic relationship functioning is described. Finally, the conclusion considers the implications for future research and intervention.

EARLY DEVELOPMENT

Attachment in Infancy

Drawing on biological, evolutionary, and ethological theories, Bowlby (1980, 1982, 1988) postulated the existence of an attachment behavioral control system similar to other physiological systems in its homeostatic organization, but distinct from systems managing other behaviors such as feeding or sexual conduct. The biological function of the attachment bond between a parent and a child is protection because the attainment of proximity to a more powerful other increases the chances of individual survival (Bowlby, 1973). Although not always in evidence, the attachment system actively monitors personal safety and the accessibility of the attachment figure at some level on a continuing basis. Most visible in children when tired, ill, or afraid, the attachment system is activated whenever the individual perceives a threat to security, stimulating behaviors designed to achieve proximity to an attachment figure, especially in stressful circumstances (Bowlby, 1980). In optimal circumstances, attachment behavior is terminated when the attachment figure restores a sense of security through sensitive responding and physical or emotional accessibility.

Attachment relationships are hypothesized to exist not only within a behavioral system governing infant–parent interaction, but also within a conceptual system governing significant social relations throughout the life span. Bowlby (1969, 1973, 1980) theorized that in the last quarter of the first year of life, humans develop internal working models (IWMs) to represent the emotions and expectations arising out of early dyadic patterns of communication between infant and caregiver. Unlike simple cognitive schemas, IWMs are thought to include affective, defensive, and descriptive cognitive components (Bretherton, 1985; Main, Kaplan, & Cassidy, 1985). Parental styles of structuring and guiding interaction direct the general organization and structure of children's memories in a process of dialogic coconstruction in which both partners take an active role (Bretherton, 1993). Optimally, children can freely express both positive and negative affect and receive appropriate caregiver responses, which directly attend to the child's communications by soothing and comforting negative states or sharing and enhancing positive states. In this way, children become aware of their internal emotional experiences and learn to control and manipulate their behavioral responses in interactions with the caregiver. Well-coordinated regulatory patterns derived from these early transactions contribute to the capacity for self-regulation and provide the basis for an integrated sense of self as a coherent whole (E. A. Carlson & Sroufe, 1995; Sroufe & Fleeson, 1986).

Through affective and communicative attunement between caregiver and child, developing processes of perception, cognition, behavior, and language merge with existing patterns of regulating and expressing emotion and maintaining organization (Bretherton, 1985, 1993). Rules for evaluating behavior, thoughts, and feelings eventually come to operate automatically outside of awareness (Bowlby, 1980) and may be associated with the development of neural networks governing the types and degree of stimuli needed to activate (i.e., perceived threat) and deactivate (i.e., felt security) the attachment system (Fraley, Davis, & Shaver, 1998). The consolidation of IWMs forms the basis for conceptual representations of self, others, and the world, which enable individuals to predict and interpret others' behaviors and to plan their own response.

In support of the theoretical link between sensitive and responsive caregiving and the development of secure attachment patterns, research has established that differences in caregivers' behaviors and adult attachment representations measured by the Adult Attachment Interview (AAI; George, Kaplan, & Main, 1985) are associated with discrete patterns of infant attachment behavior in the Strange Situation laboratory procedure (Ainsworth, Blehar, Waters, & Wall, 1978). Table 1 provides a convenient overview of the various infant and adult attachment classifications and related parenting behavior, which are described in the following sections. Adult caregivers classified as secure on the AAI tend to demonstrate sensitive and responsive caregiving and have secure infants, who in the lab demonstrate active exploration in the caregiver's presence, normal distress at separation, active greeting, and subsequent calming after the caregiver's return. In contrast, three distinct patterns of insecure adult attachment representations predict three different nonoptimal caregiving styles and three types of insecure attachment in infants (Adam, Gunnar, & Tanaka, 2004; Lyons-Ruth, Bronfman, & Parsons, 1999; Main & Hesse, 1990; Main et al., 1985; van IJzendoorn, 1995). Presumably, insecure attachment strategies are adopted by infants because they are adaptive in relation to their attachment figure's behavioral responses. For example, caregivers classified as dismissing tend to reject their infants' bids for attention in the home and have infants who show avoidant (e.g., indifferent, overly self-reliant) attachment behaviors that are effective in keeping the attachment figure from turning away from the infant. Caregivers classified as preoccupied are either unpredictable or intrusive toward the infant in the home and generally have infants who exhibit ambivalent (e.g., clingy, anxious, angry) attachment behaviors that are effective in gaining the caregiver's attention. A fourth infant classification can be assigned in conjunction with one of the three primary categories and is characterized by contradictory infant behaviors with no coherent strategy, but rather a mixture of the other attachment strategies; research evidence indicates that this disorganized attachment pattern is associated with child maltreatment, parental lack of resolution to trauma or loss, maternal frightened or frightening behavior, maternal psychopathology, and excessive alcohol use (see Lyons-Ruth & Jacobvitz, 1999, for a review).

TABLE 1 Features Associated With Attachment Classification

Infant/Adult Classification	Secure/Autonomous	Avoidant/Dismissing	Ambivalent/Preoccupied	Disorganized/Unresolved or Fearful
Infant
Strange Situation (SS) behavior	Explores readilyExpresses distress at departure of caregiverSoothed upon return of parent	Extreme autonomyAvoids closeness and contactAttachment deactivated	Extreme distress in absence of parent	Disorganized/disorientedNo coherent strategyConflicting, disjointed behavior
			Demanding or resistant
			Attachment overactivated
Adult
Adult Attachment Interview (AAI)	Coherent, autonomous, collaborative	Incoherence: Violates discourse rules	Incoherence: Violates discourse rules	Unresolved: Assigned along with primary classification
	Few discourse violations	IdealizationLack of recallDerogation	Involving angerPassivityConfusion	Lapses in reasoning or discourse re: loss or trauma
				Lack of resolution to trauma
Parenting behavior (as associated with AAI & SS)	Sensitive, responsive when needed or wanted	Avoids emotional or physical closenessRejecting	Inconsistently availableRole reversalBlurred boundaries	Frightened/frighteningHostile/helplessMaltreatment by some
	Maintains contact
Self-reported romantic attachment style	Low attachment avoidance and attachment anxiety	High attachment avoidance: Distrust of others	High attachment anxiety: Fear of abandonment	Fearful: Not determined by trauma, but empirically related
	Positive self and other models	Positive self-model, negative other model	Negative self-model, positive other model	High attachment avoidance and attachment anxiety
				Negative self and other models
Organizational/relational strategy	Flexible/adaptiveBalance between exploration and attachmentHealthy connection to others	Minimize emotion and attachment behaviorDefensively turn awayOverly self-reliantDistant, disengaged	Exaggerate emotion and attachment behaviorInternalizationOverly dependent on othersEnmeshment	Unresolved: Related to underlying primary strategy, but generally disorganized/disoriented
				Fearful: Approach/avoidance

Emotionally Abusive Caregiving

Whereas the differentiation among physical, sexual, and emotional abuse is fairly straightforward, it is much more difficult to discriminate among imperfect parenting, parental mistakes, and emotionally abusive parental behavior (Messman-Moore & Coates, 2007; Wright, 2007). Several research groups (e.g., Bifulco, Brown, & Harris, 1994; Hart, Brassard, Binggeli, & Davidson, 2002; Thompson & Kaplan, 1996) have identified a wide range of caregiver behaviors comprising emotional abuse, but the lack of consensus makes it difficult to compare and generalize across research findings. Because parental behavior lies along a continuum, determining where acceptable behavior becomes emotionally abusive is clearly complicated and subject to varied opinions (Messman-Moore & Coates, 2007; Wright, 2007). Attachment theory and research might provide clues regarding how to distinguish emotionally abusive parenting and its effects on children.

The attachment literature identifies insensitive, rejecting, and intrusive and role-reversing parenting as key contributors to insecure attachment (Ainsworth et al., 1978; van IJzendoorn, 1995). However, Collins and Feeney (2004) described avoidant and ambivalent infant attachment patterns as “good enough” strategies to obtain sufficient amounts of security and contact with others. Additionally, although theory and research suggest that insecure attachment creates a vulnerability for psychopathology (see Carlson & Sroufe, 1995, for a review), many insecure children and adults do not demonstrate interpersonal or psychological problems, and avoidant or dismissing individuals have been noted to exhibit an absence of problems (e.g., Riggs & Jacobvitz, 2002). Furthermore, Solomon and George (1999) argued that reparative actions by caregivers (e.g., hugs, apologies, explanations) following negative parenting behaviors will minimize the deleterious effects of these interactions.

Meta-analytic findings indicate parenting that is somewhat inconsistent and self-preoccupied or somewhat rejecting and dismissing is associated with ambivalent and resistant or avoidant infant behavior, but not disorganized attachment behavior (van IJzendoorn, 1995). In laboratory assessments where avoidant and ambivalent infants have established a coherent, albeit nonoptimal strategy for obtaining and maintaining proximity to caregivers, disorganized infants demonstrate disorientation, simultaneous approach avoidance, freezing behaviors, and other odd actions that do not seem to have a clearly identifiable purpose or pattern (Main & Morgan, 1996; Main & Solomon, 1986). Main and Hesse (1990) theorized that this disorganization arises when caregivers display frightening or frightened behaviors, thus presenting their children with the irresolvable paradox of wanting to seek security from their attachment figures when threatened and afraid, but also fearing the attachment figure who is “the source of and solution to their alarm” (p. 163). Evidence linking disorganized infant attachment to physical maltreatment and parental frightened or frightening behavior supports this hypothesis (Carlson, Cicchetti, Barnett, & Braunwald, 1989; Lyons-Ruth, Bronfman, & Parsons, 1999; Lyons-Ruth, Connell, Grunebaum, & Botein, 1990; Schuengel, Bakermans-Kranenburg, & van IJzendoorn, 1999).

Lyons-Ruth and her colleagues (e.g., Lyons-Ruth, Alpern, & Repacholi, 1993; Lyons-Ruth, Bronfman, & Atwood, 1999; Lyons-Ruth, Easterbrooks, & Cibelli, 1997; Lyons-Ruth, Melnick, Bronfman, Sherry, & Llanas, 2004; Lyons-Ruth, Repacholi, McLeod, & Silva, 1991) have conducted a 20+-year longitudinal study with a low-income sample of community and clinical parent–infant dyads, many of them demonstrating disorganized attachment behaviors. Building on earlier work linking parents' frightening and/or frightened behaviors to infant disorganization (Lyons-Ruth, Bronfman, & Parsons, 1999; Main & Hesse, 1990; Schuengel et al., 1999), this research group identified two profiles of problematic parenting that are associated with two subcategories of disorganized infant classification. One parenting profile, called helpless-fearful regarding attachment, is characterized by nonhostile and ostensibly responsive behavior with high levels of parental withdrawal, fearfulness, and inhibition. Helpless-fearful parenting is related to a disorganized attachment pattern termed disorganized-secure or disorganized-approach because infants still approach their parents and achieve physical contact despite demonstrating disorganized behaviors. This pattern is relatively difficult to identify and appears to be more common in low-risk, middle-class samples. The second parenting profile, called hostile-self-referential regarding attachment, is characterized by excessive role reversal, frightening and negative and intrusive behavior, as well as behavior that both elicits and rejects the infant's attachment-related emotions and behaviors. Of the five primary parenting scales coded, affective communication errors (e.g., simultaneous conflicting signals, nonresponse, or mismatched response) were the most strongly related to infant disorganization. This contradictory combination of both rejecting and attention-seeking behaviors by parents was associated with a second disorganized pattern in infants called disorganized-insecure or disorganized-avoid-resist because their disorganized behaviors were mixed with other behaviors resembling the avoidant and ambivalent or resistant groups. Despite clear group differences in behavioral manifestation, the evidence indicates that both subgroups of disorganized infants are at risk for similar negative outcomes, including conflict, apprehension, and dysphoria in reunions with caregivers, as well as highly aggressive behavior with peers by ages 5 and 7 (Lyons-Ruth, 1996; Lyons-Ruth et al., 1993; Lyons-Ruth et al., 1997).

Similar to frightening parenting described in earlier work, the hostile-self-referential parenting pattern appears to fit clearly within the bounds of what typically would be considered emotional abuse in the literature. However, parenting behaviors coded as frightened or helpless-fearful are often subtle and nonabusive (Hesse & Main, 2006; Lyons-Ruth et al., 2004). Consequently, despite similar outcomes, it is less clear where the frightened or helpless patterns fall on the parenting continuum, although they might be a form of emotional neglect in that they deprive children of needed support and positive interactions. Theoretically, in the context of both types of hostile or helpless (or frightened or frightening) relationships, insecurely attached children experience fearful arousal that has no resolution because they receive no comfort or soothing to reduce their anxiety (Hesse & Main, 2006; Lyons-Ruth et al., 2004).

Based on the available theory and research, the proposed model in Figure 1 construes only extreme and relentless rejection (e.g., antipathy, degradation), intrusiveness (e.g., controlling, guilting, role-reversing behavior), or both, which is frightening to the child and not followed by parental efforts to reconcile, as constituting emotional abuse. As suggested by the work of Lyons-Ruth, it might be the incongruous mixture of the two types of extreme parenting, especially simultaneous conflicting signals, that represents the most harmful mode of emotional abuse. The model assumes that emotionally abusive parenting will foster the development of insecure attachment in children. Although all three types of insecure attachment are possible consequences, the more severe the abuse, the more likely it is that the infant will demonstrate a disorganized-avoid-resist pattern.

Early Effects of Emotional Abuse

Figure 1 depicts two primary components of insecure attachment organization that are affected by emotional abuse: the capacity for emotional regulation and IMWs. As a consequence of “fright without solution” (Hesse & Main, 2006) caused by frightened or frightening or hostile or helpless parental behavior, insecure children experience a disregulation of arousal caused by chronic activation of the fight-or-flight mechanisms and consequently do not learn to regulate emotion effectively (Lyons-Ruth et al., 2004). In the earliest stages of development, emotions are diffuse and infants have little control over their behavioral expression, relying instead on the dyadic regulation provided by their caregivers' responses to behavioral signals (Sroufe & Fleeson, 1986). As normal development progresses, diffuse emotions become more and more refined and an increasing number of positive and negative emotions are experienced. In contrast, when children experience emotionally abusive parenting, the inability to regulate more complex emotions sets in motion a series of developmental failures in their emotional growth. O'Hagan (2006) argued that children who suffer emotional abuse experience only a limited number of emotions, which are dominated by negative rather than positive affect. It also seems likely that emotional abuse interferes with children's understanding of emotions such that they have difficulty identifying emotions in themselves or others (i.e., they lack empathy and ability to recognize social cues). As a result, emotional expression may become distorted so that these children often function at the outermost limits of the emotional spectrum (O'Hagan, 2006) and conceivably demonstrate conflicting emotional and behavioral responses consistent with attachment disorganization.

Emotional abuse also profoundly influences children's developing understanding of self and others (Liem & Boudewyn, 1999). In optimal circumstances, infants whose caregivers respond contingently and appropriately to infant signals will develop a secure attachment, characterized by positive IWMs of self as capable of affecting the environment or caregiver and worthy of care, as well as positive IWMs of others as reliable and supportive (Bartholomew & Horowitz, 1991). As a result, secure infants will feel safe enough to explore the environment but will actively seek comfort and protection from attachment figures when they are fatigued, sick, injured, or afraid. Conversely, infants whose caregivers respond in an inconsistent or intrusive manner will likely develop negative self-models as incompetent, unworthy, and powerless (see Dodge Reyome et al., in press). Attachment theory suggests that negative IWMs of self may contribute to self-blame, overdependence on the attachment figure, limited exploration, and subsequent ineffective use of social support. Other infants whose caregivers respond in a rejecting manner are likely to develop negative IWMs of others as unresponsive, inconsiderate, rejecting, or harsh (Liem & Boudewyn, 1999), making it less likely that children will trust other people or turn to others for help.

As depicted in Figure 1, the proposed theoretical model predicts that extreme forms of insecure, emotionally abusive parenting will intensify and strengthen negative internal models of the self (e.g., inept, damaged, contemptible, disgusting) or others (e.g., cruel, betraying, unforgiving, hateful). As a consequence of maltreatment, children lack a “sense of self-efficacy and capacity to manipulate the environment in service of emotional and instrumental coping” (Liem & Boudewyn, 1999, p. 1143). Children who experience a mixture of inconsistency or intrusion and rejection from their parents are likely to be doubly disadvantaged by negative IWMs of both self and others. The model further proposes that emotionally abusive parenting may engender dissociated mental representations that create a high risk for maladaptive social relations and psychopathology (see Carbone, 2010, this 2010; Riggs & Kaminski, 2010). According to Bowlby (1973, 1980), defensive exclusion occurs when individuals deactivate or block normal attachment responses or separate attachment information from consciousness to minimize conflict and anxiety. Experiences kept from awareness in this way might include strong rejection or punishment for attachment behaviors and feelings, direct or implied threats of abandonment, more severe abuse, and other events that are too painful to think about or that parents would prefer children not recognize (Bowlby, 1980, 1988). Defensive exclusion processes that are unconsciously activated and closed to environmental feedback may create distortions in IWMs, resulting in dissociation between conscious and unconscious working models and the formation of multiple incompatible belief systems (Bowlby, 1980). Bowlby (1973) suggested that these incongruent representations play a critical role in maladaptive adjustment and psychopathology.

CHILDHOOD AND ADOLESCENCE

Continuity of Attachment

Although the proposed model assumes that emotional abuse begins at birth, there are clearly situations when children might not experience emotional abuse until later ages when parental treatment shifts from acceptable to abusive or a new abusive caregiver enters their lives. The impact of this later emotional abuse will vary, depending to some degree on the children's past attachment experiences, their current attachment network, and other circumstances. Bowlby (1988) and Ainsworth (1989) believed that the attachment pattern formed in early childhood becomes an essential component of a person's overall personality structure with considerable influence on personality traits and individual ways of being and reacting to the world. Several longitudinal studies have demonstrated stability of attachment classification from infancy through 6 years of age and up to adolescence and early adulthood (see Grossmann, Grossmann, & Waters, 2005, for a review of major longitudinal studies following infants into adulthood). Yet, the attachment organization developed in childhood does not predetermine an individual's attachment style in later life, nor does it necessarily forecast a particular prognosis for lifetime adjustment and mental health. Several studies have found lawful discontinuity of attachment from infancy to adolescence through early adulthood in relation to disruptive life events, such as death, divorce, and maltreatment (Allen, McElhaney, Kuperminc, & Jodl, 2004; Cozzarelli, Karafa, Collins, & Tagler, 2003; Waters, Merrick, Treboux, Crowell, & Albersheim, 2000).

Collins and Sroufe (1999) proposed a transactional model whereby early attachment experiences continue to affect children's functioning and later relationships, but children also contribute to their own development and environments by responding and engaging with others according to previous patterns of adaptation and expectations. Specifically, IWMs of self and others formed in childhood influence immediate emotional reactions, the cognitive processes that maintain and amplify or lessen emotional responses, and the resulting behavior. According to the transactional model, experience-based variations in acquiring specific relational skills (e.g., empathy, social perception, communication, negotiation, conflict resolution) and interpersonal cognitions or biases likely mediate the link between early attachments and later relationships.

Theoretically, continuity in attachment patterns can be explained at least partially by the impact of early IWMs on cognitive-affective processes, which direct attention to specific stimuli, create biases in memory encoding and retrieval, guide expectations regarding the availability of others, and influence attributions regarding the behavior of others in ambiguous situations (Cassidy, 2000). For example, research indicates that secure children make realistic or benevolent attributions regarding the intentions of peers, whereas avoidant children showed more unrealistic or hostile and negative biases (Suess, Grossmann, & Sroufe, 1992). These cognitive-affective processes influence the individual's decisions, emotional responses, and behaviors. In particular, secure individuals tend to be open and flexible in emotional expression, but insecure children may either minimize (avoidant) or heighten (ambivalent) the expression of negative emotions, which then elicits the kind of treatment the individual receives from others and typically reinforces the IWMs in a self-perpetuating cycle across new contexts and developmental stages (Cassidy, 1994).

Consistent with this idea, Sroufe and Fleeson (1986) reported findings that avoidant preschoolers appeared to elicit punitive reactions from teachers, whereas ambivalent or resistant children were rated as more emotionally immature by teachers and thus received more nurturance and tolerance from them. Importantly, these researchers noted that in different peer dyads children often displayed opposite or complementary behavior (e.g., child was rejected or conversely rejecting) depending on the attachment strategy of the dyadic partner, suggesting that the whole relationship is internalized and carried forward into new relationships in such a way that the particular role enacted will depend on the other partner's responses. For example, having learned both roles in the victim–victimizer dichotomy, in peer interactions, a maltreated child may assume the role of victimizer, bullying other children who appear weak and easy targets, or conversely the same child may assume the role of victim and be bullied by other children who are more dominant and aggressive.

Effects of Emotional Abuse in Childhood and Adolescence

Yates (2007) conceptualized emotional abuse as a form of chronic relational adversity that disrupts development across social, emotional, self, cognitive, and biological domains. Thus, childhood emotional abuse, like other forms of early maltreatment, poses a serious challenge to the formation and maintenance of secure attachments, causing significant interference with the adaptive development of emotional regulation and IWMs of self and others (Cicchetti & Toth, 2000; Collins, Guichard, Ford, & Feeney, 2004). Although emotional abuse at any age is likely to have detrimental effects, the transactional model of Collins and Sroufe (1999) suggests that emotional abuse in infancy may be more damaging than emotional abuse occurring later in development because early experiences interact with later experiences to determine present-day outcomes. Although a secure attachment established in infancy might somewhat protect children from the full impact of later emotional abuse, insecure infant attachment could create a greater vulnerability to maladaptive responses to emotional abuse occurring at older ages. However, the child's current attachment network is also critical, such that the presence of a secure relationship with a supportive adult might counteract the negative impact of emotional abuse from another caregiver.

Consistent with the proposed model in Figure 1, O'Hagan (2006) described two prototypes of emotionally abused children who possess coping strategies similar to the behaviors expected for the avoidant and ambivalent attachment groups. Specifically, like the behaviors expected of avoidantly attached children, some emotionally abused children might exhibit the extreme of “dominant silent, invisible negative emotion” because they have learned that the natural expression of emotions provokes hostile reactions from the caregiver and strive to inhibit any observable indications of affect. According to O'Hagan, the constant suppression will eventually become automatic and weaken the capacity to feel positive affect, leaving their emotional world ruled by indifference, apathy, dejection, and unspoken hatred. At the other extreme, similar to the exaggerated emotional coping of ambivalently attached children, some emotionally abused children exhibit “dominant audible and visible negative emotions,” likely modeled after a caregiver's highly visible emotional displays, which can be loud, aggressive, erratic, unrestrained, and destructive (O'Hagan, 2006).

Although some emotionally abused children might demonstrate ambivalent and avoidant attachment strategies, theoretically, if the abuse begins at birth, it seems likely that these children start out as disorganized infants, who by definition do not demonstrate a predominant, coherent strategy for seeking and maintaining proximity to the attachment figure. However, as noted in Figure 1, by the time they reach school age, disorganized children show more organized coping strategies that have been described as “controlling” toward the parent (Main & Cassidy, 1988; Wartner, Grossmann, Fremmer-Bombik, & Suess, 1994). Rather than seeking comfort and protection for their own needs, this controlling strategy appears to be designed to maintain engagement with the caregiver by gratifying the caregiver's self-preoccupation (Lyons-Ruth, Bronfman, & Parsons, 1999; Main & Cassidy, 1988) and might take the form of caretaking of the parent by the child or a hostile and punitive stance toward the parent. Lyons-Ruth et al. (2004) speculated that these two types of controlling strategies are developmental extensions of the two subcategories of disorganized infant attachment. Cross-sectional studies have reported that both controlling strategies are associated with children's externalizing symptoms (Greenberg, Speltz, DeKlyen, & Endriga, 1991; Moss, Parent, Gosselin, Rousseau, & St-Laurent, 1996; Moss, Rousseau, Parent, St-Laurent, & Saintonge, 1998; Solomon, George, & De Jong, 1995; Speltz, Greenberg, & DeKlyen, 1990). In adolescence, preliminary evidence suggests that behavioral disorders and cognitive distortions are related to controlling behaviors with mothers and disorganized behaviors with fathers (Hilburn-Cobb, 2004).

IWMs of self and others provide the basis for expectations of others and guide responses, thus affecting individual and interpersonal outcomes. Consistent with the idea that emotional abuse negatively impacts IWMs and the ability to regulate affect, research suggests that emotional abuse places children at risk for poor self-concept and disorders of emotional regulation and impulse control. For example, studies with school-age children and adolescents indicate that emotional abuse is associated with low self-esteem and high levels of aggression (see Dodge Reyome et al., in press; Herrenkohl, Egolf, & Herrenkohl, 1997), depression (see Dodge Reyome et al., in press; Kaufman, 1991; Riggs & Kaminski, 2010), depressive attributional style (Cerezo & Frias, 1994), delinquency (Vissing, Straus, Gelles, & Harrop, 1991), emotional instability, and externalizing and internalizing problems (see Dodge Reyome et al., in press; Riggs & Kaminski, 2010; Wenar & Kerig, 2000).

As suggested by evidence linking emotional abuse to children's interpersonal problems with peers (see Taussig & Culhane, 2010; Vissing et al., 1991), the proposed model assumes that the attachment insecurity resulting from emotionally abusive parenting also contributes to poor social functioning. It is in the context of early attachment relationships that children begin to develop the skills needed to build future social relationships, such as self-awareness, empathy, negotiation, and conflict resolution. Across various developmental stages, longitudinal research suggests that security of attachment influences many aspects of interpersonal relationships, including effectiveness in peer groups, reciprocity in relationships, empathy, problem solving and conflict resolution, and establishing close and intimate relationships with peers (see Carlson & Sroufe, 1995, for a review; see also Kestenbaum, Farber, & Sroufe, 1989; Suess et al., 1992). Undoubtedly, emotional abuse in the attachment relationship substantially increases the probability of developing insecure attachment, which is empirically linked to low empathy and reciprocity, hostility or aggression and impulsivity, exploitation or ridicule by peers, social withdrawal or exclusion from group activities, and general patterns of unrelatedness and isolation or complete fusion and dependence where intimacy was achieved at the expense of autonomy (Elicker, Englund, & Sroufe, 1992; Erickson, Egeland, & Sroufe, 1985; Grossmann & Grossmann, 1991; Lewis, Feiring, McGuffog, & Jaskir, 1984; Lyons-Ruth et al., 1993; Renken, Egeland, Marvinney, Sroufe, & Mangelsdorf, 1989; Sroufe, Carlson, & Shulman, 1993; Sroufe & Fleeson, 1986; Troy & Sroufe, 1987; Urban, Carlson, Egeland, & Sroufe, 1991).

ADULT DEVELOPMENT AND FUNCTIONING

Childhood emotional abuse continues to influence individual functioning in adulthood. In retrospective studies, adults who acknowledged emotional abuse in childhood reported low self-esteem, interpersonal sensitivity, hostility, and aggression, as well as symptoms of depression, dissociation, suicidality, and somatization (Bifulco, Moran, Baines, Bunn, & Stanford, 2002; Briere & Runtz, 1988, 1990; Ferguson & Dacey, 1997; Gross & Keller, 1992; Morimoto & Sharma, 2004; Mullen et al., 1996; Nicholas & Bieber, 1996; Spertus, Yehuda, Wong, Halligan, & Seremetis, 2003; Stone, 1993). Moreover, evidence indicates that childhood emotional abuse is significantly related to poor interpersonal skills and relationship dysfunction in adulthood (see Berzenski & Yates, in press; Carbone, 2010; Davis, Petretic-Jackson, & Ting, 2001; Dodge Reyome, in press; Dodge Reyome et al., in press; Messman-Moore & Coates, 2007; Paradis & Boucher, in press; Riggs & Kaminski, this 2010; Varia & Abidin, 1999). The theoretical model presented in Figure 1 proposes that these negative outcomes are largely due to the deleterious impact of emotional abuse on the emergent attachment system, leading to emotional dysregulation, negative IWMs, cognitive biases and distortions, maladaptive coping strategies, and insecure adult attachment style. The model further suggests that all of these factors are likely mediators between early childhood emotional abuse and mental health, which in turn might mediate their associations with romantic relationship quality (DiLillo, Lewis, & Di Loreto-Colgan, 2007; Perry, DiLillo, & Peugh, 2007). The following two sections review the relevant correlates of adult attachment classifications, with a special emphasis on coping patterns and mental health, then explore the potential impact of childhood emotional abuse on romantic relationship functioning.

Adult Attachment, Coping, and Mental Health

In the late 1980s, two different lines of adult attachment research emerged distinguished by conceptual and methodological differences (see Riggs et al., 2007, for a review). The developmental line of adult attachment research focuses on early childhood experience with parents and its impact on caregiving and infant attachment behavior, and it has generally employed the AAI (George et al., 1985) to assess adults' current mental representations with respect to early parent–child attachments. The social or personality line of adult attachment research studied adult relationships with romantic partners, which are assumed to be outgrowths of early attachment bonds, and typically measures adult romantic attachment using self-report instruments assumed to assess consciously held attitudes regarding close relationships (Bartholomew, 1997; Simpson & Rholes, 2002). Both lines of research identify adult attachment classifications analogous to the infant attachment patterns described in earlier research (Ainsworth et al., 1978; Crittenden, 1985; Main & Solomon, 1986) and are relevant to the proposed model—the latter in association with romantic relationship functioning and the former in relation to the potential for intergenerational transmission of emotionally abusive parenting behaviors. Consequently, both literatures are reviewed here and both forms of adult attachment (self-reported romantic attachment vs. AAI current representations of parent–child attachment) are included in Table 1 and the model depicted in Figure 1.

Conceptually, there are obvious similarities between parent–child attachment and adult romantic attachment, but there are also important differences. Both types of attachment bonds consist of four defining features: secure base, proximity maintenance, safe haven, and separation distress (Hazan & Zeifman, 1999). However, adult romantic relationships generally involve three systems: the attachment system, the caregiving system, and the sexual system (Ainsworth, 1989; Shaver, Hazan, & Bradshaw, 1988). Whereas infant–caregiver attachment is predominately asymmetrical in terms of caregiving, which is exclusively the adult's role, adult romantic attachment is bidirectional and reciprocal because each partner uses the other as a source of security and also serves as a provider of security. According to Hazan and Zeifman (1999), the two types of attachment bonds also differ in terms of distinct motivations for physical contact and evolutionary function; that is, infants seek proximity to reduce distress and promote safety and survival, whereas adults seek proximity initially based on sexual attraction and the propagation of the species, and later for mutual comfort and protection.

Bartholomew (1990; Bartholomew & Horowitz, 1991) theorized that IWMs of self and others formed in childhood are related to the degree of abandonment anxiety, interpersonal avoidance, or both that individuals experience throughout their lives. Having established trust and open communication in past relationships, individuals classified as secure experience low levels of attachment anxiety and avoidance and are likely to remain open and flexible in their interpretation of and response to new information. Consistent with this conceptualization, Main et al. (1985) hypothesized that language provides the means to organize experiences in the form of a coherent life narrative, or secure working model of attachment, which meaningfully integrates self-/other-awareness and contributes to ongoing self-interpretation. In interviews using the AAI, secure attachment manifests in coherent, autonomous, and cooperative discourse regarding childhood attachment experiences and the ability to provide an integrated, credible account of early experiences and their meaning (Main & Goldwyn, 1998).

When the attachment system is activated after a threat is perceived, the positive mindset of secure individuals influences the adaptive mobilization of internal resources and social support during the stress and coping process (Mikulincer & Shaver, 2003). Consequently, a secure attachment organization with positive models of self and others should act as a buffer against maladaptive responses to stress, allowing individuals to pursue and effectively utilize support from friends, family, or mental health practitioners (Riggs & Jacobvitz, 2002). Research findings generally suggest that secure individuals perceive stressors as less threatening, employ more constructive coping methods, and experience higher self-efficacy about coping (see Mikulincer & Florian, 1998, for a review). As a result, secure individuals are less likely to show symptoms of emotional disturbance (e.g., Kobak & Sceery, 1988; Riggs & Jacobvitz, 2002) and are more likely to demonstrate effective coping strategies, such as a healthy degree of self-disclosure and help-seeking (Lopez, Melendez, Sauer, Berger, & Wyssmann, 1998; Mikulincer & Nachshon, 1991; Riggs, Jacobvitz, & Hazen, 2002). Researchers have also reported that secure attachment patterns in adults are associated with personality traits indicative of self-confidence, self-acceptance and self-esteem, identity and integration, psychological well-being and resiliency, empathy, agreeableness, warmth, intimacy and balance in friendships, extroversion, and engagement with the social world (e.g., Bartholomew & Horowitz, 1991; Benoit, Zeanah, & Barton, 1989; Diehl, Elnick, Bourbeau, & Labouvie-Vief, 1998; Onishi, Gjerde, & Block, 2001; Shaver & Brennan, 1992; Treboux, Crowell, & Colon-Downs, 1992).

In contrast, theory and research suggest that insecure individuals are at risk for the development of interpersonal and emotional problems due to significant interference with cognitive or affective aspects of mental representations, which limits their ability to respond flexibly in unfavorable situations (Carlson & Sroufe, 1995). Cognitive distortions triggered by present-day events resembling earlier anxiety-provoking interactions may promote the formation of multiple models, which are inflexible and fairly impermeable to new information (Bowlby, 1980, 1988). Among individuals who were emotionally abused as children, significant life stressors can stimulate high levels of anxious arousal, resulting in an escalation of defenses, maladaptive behaviors, vulnerability to cognitive disorganization and the formation of dissociated contradictory mental models, and ultimately an emotional disorder. Theoretically, the particular manifestation of emotional and behavioral disturbance will depend to a large extent on the predominant attachment strategy previously adopted by the individual.

Among preoccupied individuals who have experienced inconsistent, controlling, or role-reversing parenting, negative beliefs about the self engender anxiety regarding abandonment, but generally positive views of others favor help-seeking (Bartholomew, 1990). In stressful circumstances, uncertainty about the responsiveness of the attachment figure increases anxiety levels and sustains the activity of the attachment system, leading to emotion-focused coping strategies (Mikulincer & Shaver, 2003). Consequently, preoccupied individuals appear to employ a hyperactivating style, whereby they exaggerate emotions and become overinvolved in the availability (or lack thereof) of others (Dozier, Stovall, & Albus, 1999), which might manifest itself in AAI discourse as passive dependence, confusion, or anger toward parental attachment figures (Main & Goldwyn, 1998). Although individuals with preoccupied attachment style might seek social support when threatened, they can exaggerate the danger and demand assistance with very little justification (i.e., their reaction is out of proportion to the stimulus) or exhibit clingy overdependence and ineptitude. Studies have linked AAI preoccupied attachment to increased psychological distress, a high incidence of mood disturbance and anxiety symptoms, and borderline personality disorder (Cole-Detke & Kobak, 1996; Fonagy et al., 1996; Kobak & Sceery, 1988; Patrick, Hobson, Castle, Howard, & Maughan, 1994; Pianta, Egeland, & Adam, 1996; Rosenstein & Horowitz, 1996). Self-reported preoccupied attachment has been empirically associated with negative self-view, high neuroticism, low self-control and tolerance, interpersonal dependence and reliance, as well as histrionic, dependent, and borderline personality traits (Allen, Coyne, & Huntoon, 1998; Bartholomew & Horowitz, 1991; Brennan & Shaver, 1998; Collins & Read, 1990; Diehl et al., 1998; Onishi et al., 2001; Shaver & Brennan, 1992).

Individuals with dismissing attachment strategies are likely to have childhood histories of avoidant attachment marked by parental rejection. As a result, they develop an intra- and interpersonal strategy intended to deactivate or inhibit natural attachment responses with the goals of avoiding negative emotional states and achieving distance, control, and independence (Shaver & Mikulincer, 2002). To maintain IWMs of the self as invulnerable and others as untrustworthy or weak, dismissing individuals tend to use avoidant coping strategies involving denial, externalization, detachment, withdrawal, perfectionism, anger, and possibly narcissism (Bowlby, 1979; Carlson & Sroufe, 1995; Crittenden, 1995; Dozier et al., 1999; Rosenstein & Horowitz, 1996). When threatened, people with dismissing attachment might experience distress and anxiety at a physiological level, but fail to consciously acknowledge these feelings, instead denying the impact of stressors and withdrawing from others to preserve their self-reliance and inhibit any distress-related emotions (Mikulincer & Florian, 1998; Mikulincer & Shaver, 2003; Simpson, Rholes, & Nelligan, 1992). Researchers have consistently reported that the dismissing category is associated with hostility, interpersonal avoidance, loneliness, distrust, low emotional expressiveness, low agreeableness, weak ego resiliency, low levels of self-disclosure and intimacy, and interpersonal difficulty at work (Bartholomew & Horowitz, 1991; N. Collins & Read, 1990; Hazan & Shaver, 1987, 1990; Kobak & Sceery, 1988; Onishi et al., 2001; Shaver & Brennan, 1992). On the AAI, dismissing attachment is associated with a lack of memory for childhood events, and idealization or derogation of attachment figures. With respect to mental disorders, recent studies have provided empirical support linking the AAI dismissing classification to conduct disorder, antisocial personality or criminality, substance abuse, and bipolar disorder in clinical populations (e.g., Allen, Hauser, & Borman-Spurrell, 1996; Cole-Detke & Kobak, 1996; Fonagy et al., 1996; Rosenstein & Horowitz, 1996; Tyrell & Dozier, 1997). Studies utilizing self-report instruments have found that dismissing attachment is associated with somatization and repressive tendencies (Mikulincer, Florian, & Weller, 1993; Onishi et al., 2001).

In the developmental literature utilizing the AAI, a fourth classification is the unresolved category, which parallels and is associated with having an infant classified as disorganized (van IJzendoorn, Schuengel, & Bakermans-Kranenburg, 1999). Unresolved adults typically demonstrate a predominant secure, dismissing, or preoccupied strategy, but by definition, have experienced either a traumatic loss or physical or sexual abuse, which is unintegrated and contributes to lapses in discourse or reasoning during discussions of the trauma (Main & Goldwyn, 1998). The unresolved classification has been empirically linked to emotional distress and substance abuse, psychiatric hospitalization, suicidal ideation, dissociative disorders, posttraumatic stress disorder (PTSD), and borderline and schizotypal personality disorders (Adam, Sheldon-Keller, & West, 1996; Alexander, 1992; Allen et al., 1996; Dozier et al., 1999; Fonagy et al., 1996; Fonagy et al., 1995; Patrick et al., 1994; Riggs & Jacobvitz, 2002; Riggs et al., 2007; Stovall-McClough & Cloitre, 2006; van IJzendoorn & Bakermans-Kranenburg, 1996).

Given that disorganized children might be at greater risk for subsequent loss, trauma, or both (Lyons-Ruth et al., 2004), the theoretical model in Figure 1 suggests that early emotional abuse might contribute indirectly to a lack of resolution of these experiences on the AAI in adulthood. Further, the coding of lapses in reasoning and discourse, which are the hallmark of the AAI unresolved classification, might tap into dissociative mechanisms (Riggs et al., 2007), which have been found to be associated with early emotional abuse (Briere & Runtz, 1988; Ferguson & Dacey, 1997). However, a fifth AAI category called cannot classify is sometimes found in high-risk samples and is associated with a history of psychiatric disorder, marital violence, PTSD symptoms, and sexual or physical abuse (Allen et al., 1996; Holtzworth-Munroe, Stuart, & Hutchinson, 1997; Riggs et al., 2007; Stalker & Davies, 1995). The cannot classify category is characterized by an unusual mixture of secure, preoccupied, and dismissing features and represents a “global breakdown in the organization and maintenance of a singular [attachment] strategy” (Hesse, 1996, p. 4). Although further clarification and refinement of this category is needed, the combination of conflicting attachment strategies associated with the cannot classify category is reminiscent of the hostile-self-referential parenting profile and infant disorganized-avoid-resist pattern, suggesting that emotionally abused children might be at risk for this type of global breakdown in the adult attachment system. Based on empirical associations between parents' AAI attachment classifications and their infants' attachment patterns (van IJzendoorn, 1995), the model presented in Figure 1 proposes that insecure states of mind regarding childhood attachment experiences, especially unresolved and cannot classify states, represent a crucial link connecting adult attachment to emotionally abusive parenting, thus perpetuating the cycle of emotional abuse.

In the social and personality literature, a fourth self-report attachment category labeled fearful is characterized by both negative self and others models (Bartholomew, 1990). This attachment pattern is prevalent among adults reporting a childhood history of physical or sexual trauma (Alexander, 1993; Riggs et al., 2007; Shaver & Clark, 1994) and can also characterize survivors of severe emotional abuse. Simpson and Rholes (2002) theorized that fearful individuals are unable to determine the viability of seeking help from a romantic attachment figure and consequently experience strong internal conflicts that lead to behavioral and emotional disruption. As a result, like disorganized infants and parents with a hostile caregiving profile, fearful adults can exhibit a mixture of preoccupied and dismissing behaviors in their close relationships. These chaotic coping strategies might generate a confusing array of approach and avoidance behaviors (Simpson & Rholes, 2002) or might cause fearful individuals overwhelmed by stress to completely shut down, refuse to take action, and withdraw (Bartholomew & Horowitz, 1991). Similar to preoccupied attachment, fearful individuals lack self-worth but desire and ruminate about close relationships; like dismissing attachment, they distrust others, causing them to avoid intimacy for fear of being hurt.

Not surprisingly, empirical findings suggest that the fearful style is associated with the poorest adjustment of the four self-report prototypes. Studies have documented strong statistical relationships between the fearful style and a variety of maladaptive personality traits, such as distrustful interpersonal distance, anxious self-preoccupation, hostility, and low levels of the following: self-disclosure, intimacy, reliance on others, self-confidence, assertiveness, sociability, empathy, responsibility, and sense of well-being or self-worth (Bartholomew & Horowitz, 1991; Diehl et al., 1998; Onishi et al., 2001). Theoretical predictions that contradictory mental models of self and others contribute to psychopathological symptoms are supported by significant associations between the fearful style and neurotic defenses, depression, anxiety expressed in body symptoms, substance abuse, and dissociation, as well as paranoid, schizoid, schizotypal, avoidant, self-defeating, borderline, narcissistic, and obsessive-compulsive personality traits (Anderson & Alexander, 1996; Brennan & Shaver, 1995, 1998; Brennan, Shaver, & Tobey, 1991; Carnelley, Pietromonaco, & Jaffe, 1994; Diehl et al., 1998; Levy, Meehan, Weber, Reynosa, & Clarkin, 2005; Onishi et al., 2001; Riggs et al., 2007).

Impact of Emotional Abuse on Romantic Relationships

By interfering with the development and functioning of the attachment system, early maltreatment by parents creates deficits in areas of interpersonal functioning, which contribute to poor romantic relationship functioning and marital dissatisfaction (Carbone, 2010; Davis et al., 2001; DiLillo, 2001; DiLillo et al., 2007; Dodge Reyome, in press; Dodge Reyome et al., in press; Paradis & Boucher, in press; Perry et al., 2007). Although there is a small, but growing body of literature on the impact of childhood emotional abuse on adult romantic relationships, and a rather large body of literature on the links between adult attachment style and romantic relationships, relatively little work has been done to integrate these literatures. Other than one study reporting that nonclinical adults who acknowledged childhood emotional abuse were more likely to endorse dismissing or preoccupied attachment styles (Varia & Abidin, 1999), research on direct associations between early emotional abuse and adult attachment is lacking (see Riggs & Kaminski, 2010). Consequently, the following discussion considers preliminary work on childhood emotional abuse and adult relationships, but primarily concentrates on reviewing attachment theory and research that supports the proposed developmental model illustrated in Figure 1.

Although longitudinal evidence has recently emerged linking infant attachment classification to emotional expression and conflict resolution behavior with romantic partners in early adulthood (Roisman, Collins, Sroufe, & Egeland, 2005; Simpson, Collins, Tran, & Haydon, 2007), the bulk of the relevant literature focuses on links between adult attachment and various aspects of romantic relationships. At a general level, there is ample and consistent evidence that compared to insecure attachment, secure adult attachment is related to better marital functioning and satisfaction (e.g., Banse, 2004; Cohn, Cowan, Cowan, & Pearson, 1992; Eiden, Teti, & Corns, 1995; Fuller & Fincham, 1995), as well as healthier emotional regulation and less withdrawal during marital interactions (Paley, Cox, Burchinal, & Payne, 1999). More focused research has demonstrated that adult attachment style is significantly associated with fundamental components of romantic relationships involving multiple behavioral systems that interact to influence dyadic functioning and satisfaction, including the capacity for intimacy, partner caretaking and support, sexual behavior, conflict management, and relational aggression (Collins & Read, 1990; Dutton, Saunders, Starzomski, & Bartholomew, 1994; Senchak & Leonard, 1992; Shi, 2003; Simpson, 1990; Simpson, Rholes, & Phillips, 1996; Simpson, Winterheld, Rholes, & Orina, 2007; Wekerle & Wolfe, 1998). The component features of romantic relationships that are most likely to characterize the relationships of adults with different attachment styles are discussed next and summarized in Table 2.

TABLE 2 The Adult Attachment System and Romantic Relationship Components

Components	Secure Attachment	Insecure Attachment
		Preoccupied (High Anxiety)	Fearful (High Anxiety and Avoidance)	Dismissing (High Avoidance)
Intimacy	• Value intimacy, closeness• Balance of intimacy and autonomy• Self-disclosure, reciprocity• Tolerates intense emotions	• Desire intimacy, but anxious regarding abandonment• Enmeshment, dependency• Overdisclosure• Exaggerated emotions (e.g., jealousy)	• Desire but fear of intimacy• Variable fusion and distance• Hypersensitive to rejection• Low disclosure• Unable to tolerate intense emotion	• Distrust, avoid intimacy• Disengagement, compulsive self-reliance• Low disclosure; blame others• Inhibition of emotions, withdrawal
Care
Careseeking	• Seek support when stressed• Perceive support as helpful	• Excessive seeking of care and reassurance• Feel unlovable, unworthy• Negative support strategies• Compulsive caregiving• Low reciprocity• Low responsiveness	• Approach/avoidance: Want care, but fear being hurt• Perceive support as harmful• Negative support strategies• Controlling style• Contradictory caregiving: Rejecting, smothering	• Unlikely to seek support• Perceive support as unhelpful, noncaring• Instrumental care• Discomfort and rejection of partner's attachment needs
Caregiving	• Availability• Act as secure base• Openness and flexibility• Attunement and responsivity
Sex	• Later onset• Enjoy physical contact• Prefer normative behaviors• Unlikely to have sex outside of relationship	• Earlier onset with goal of gaining acceptance• More willing to engage in risky sexual behavior• Use sex to seek care	• May seek sex then abruptly withdraw• Averse to intimate touch• Strong negative emotions after sex	• Earlier onset with goal of increasing status and prestige• Unlikely to use touch to express emotion• Casual attitude, promiscuity
Conflict	• Optimistic attributions• Constructive communication• Compromise• Problem-solving strategies	• Self-blame• High obliging, guilt• Hostility, attacking• Controlling: Coercion, threat	• Perceive conflict as threat: Fight–flight activation• Extreme attachment behaviors• Intimacy anger: Anxiety may lead to relational aggression	• Hostile misattributions

Intimacy

Retrospective research findings suggest that early emotional abuse impairs the ability to develop intimacy with others later in life (Carbone, 2010; Davis et al., 2001). According to Collins and Sroufe (1999), to achieve intimacy a person must be (a) oriented to value and seek closeness, (b) able to tolerate intense emotions that are inherently involved in close relationships, and (c) capable of self-disclosure, reciprocity, and sensitivity to and concern for others' feelings and well-being. Theoretically, early caregiving relationships contribute to the capacity for intimacy by providing the initial context for learning reciprocity and empathy and also via their influence on IWMs that guide relationship expectations and the development of self-worth and self-efficacy (Collins & Sroufe, 1999). Building on previous attachment experiences, as children venture outside the home, the capacity for intimacy evolves slowly through a series of transformations in peer interactions.

Secure children clearly have an advantage in this developmental sequence because they have developed the social competence and skills that will help them succeed in interpersonal relationships at each developmental stage. For example, research indicates that secure infants later display more empathy than insecure infants (Elicker et al., 1992; Kestenbaum et al., 1989; Sroufe, 1983), and older secure children are more likely to demonstrate the ability to communicate, respond empathically, and resolve conflict (see Collins & Sroufe, 1999, for review; see also Kerns, 1994). Other findings suggest that secure adolescents are more popular with their peers, demonstrate the highest mastery of social skills, have higher self-esteem, and are more oriented toward emotional depth that comes with romantic involvement (see Collins & Sroufe, 1999, for review). As adults, because they tend to have positive cognitions about relational events (e.g., arguments) and provide optimistic explanations for ambiguous relational events to minimize the potential of negative impact (Collins, 1996; Collins & Feeney, 2004), secure individuals are more likely to achieve high levels of intimacy in their romantic relationships without losing personal autonomy and typically are comfortable with self-disclosure and responsive when their romantic partners self-disclose to them (Collins & Feeney, 2004; Grabill & Kerns, 2000; Mikulincer & Nachshon, 1991).

In contrast, children who experience an emotionally abusive attachment relationship are at a clear disadvantage in interpersonal contexts due to insecure IWMs of attachment and problems with emotional regulation. Abuse-related emotional and cognitive deficits, such as negatively biased information processing or perceptual vigilance, contribute to the development of maladaptive schema and rigid interpersonal beliefs, which repeat throughout life (Carbone, 2010; Rogosch, Cicchetti, & Aber, 1995; Young, Klosko, & Weishaar, 2003). Given the demands for intimacy, trust, and conflict management, romantic relationships are likely to be central contexts for acting out these maladaptive interpersonal schemas (Crawford & Wright, 2007). In particular, experiences of childhood emotional abuse may contribute to schema related to interpersonal disconnection and rejection, which have significant long-term effects because they interfere with the ability to form secure, fulfilling relationships with partners (Messman-Moore & Coates, 2007). Preliminary evidence suggests that early emotional abuse is related to later maladaptive schemas of mistrust and abuse, emotional deprivation, emotional inhibition, self-sacrifice, defectiveness and shame, and abandonment (Cecero, Nelson, & Gillie, 2004; Crawford & Wright, 2007; Dodge Reyome, in press; Dodge Reyome et al., in press; Messman-Moore & Coates, 2007).

Not only do insecure adults experience difficulty managing the powerful emotions activated in close relationships, but the biased interpretations and maladaptive coping strategies associated with negative models of self or others create an imbalance in the search for connection and maintenance of autonomy, which are necessary for a healthy intimate relationship (Blatt & Levy, 2003). For example, preoccupied individuals desire intimacy, but findings suggest they tend to become enmeshed, losing their autonomy in romantic relationships and might self-disclose too easily and too much to gain approval of others (Collins & Feeney, 2004; Shaver, Schachner, & Mikulincer, 2005). A high level of anxiety in relationships and doubt that they can be understood, validated, and loved by a romantic partner (Grabill & Kerns, 2000) can result in relational dissatisfaction, negative attributions for partner's transgressions, and romantic jealousy (Collins, 1996; Collins & Feeney, 2004; Guerrero, 1998). On the other hand, dismissing adults generally avoid intimate relationships and deny need for closeness (Collins & Feeney, 2004; Feeney & Noller, 1990). Research has shown that dismissing individuals use distancing strategies when experiencing relational distress, demonstrate low self-disclosure, and respond negatively when others self-disclose to them (Bradford, Feeney, & Campbell, 2002; Collins & Feeney, 2004; Feeney, 1999; Fraley & Shaver, 1999).

With negative IWMs of both self and others, fearful adults are liable to experience the greatest difficulties in intimate relationships. Initiated in the context of defective or emotionally abusive relationships with parents, high levels of anxiety about abandonment combined with a deep-seated mistrust of others likely engender an intense fear of intimacy (Wood, 2007). Like preoccupied adults, they fervently desire intimacy but feel unworthy of love and are hypersensitive to rejection. Moreover, similar to dismissing adults, fearful individuals view their romantic partners as unreliable and use distancing strategies during times of distress (Brennan, Wu, & Love, 1998; Collins & Feeney, 2004). As a result, their emotional desire for intimacy is usually overridden by their negative cognitions regarding self and others (Collins & Feeney, 2004). If fearful adults do enter into an intimate relationship, theory suggests that emotionally abused individuals whose history is marked by disorganized attachment in infancy and controlling attachment strategies in childhood might demonstrate a controlling and overdominating interpersonal style, which could prompt jealous rages and attempts to isolate romantic partners. In addition, dissociated mental models triggered by painful relational events reminiscent of earlier attachment experiences might initiate the fight-or-flight response, engendering paradoxical approach-avoidance behaviors that confuse and bewilder partners, one minute appearing madly in love and desiring complete enmeshment and the next minute disconnecting from and disparaging the very same partner who was adored moments before.

Caretaking and care-seeking

Relationships with peers and romantic partners are optimally reciprocal, with each partner giving and taking in a balanced fashion. Because they have received inadequate and abusive care from parents, emotionally abused children are likely to develop a warped conception of what a loving and caring relationship entails (see Carbone, 2010; Dodge Reyome et al., in press). On one side of the caregiving relationship is the ability to seek and accept care from others. Research indicates that secure adults are willing to ask their romantic partners for both emotional and instrumental support and believe the caretaking behaviors of their partners are motivated by altruistic rather than selfish concerns (Collins & Feeney, 2004; Knobloch, Solomon, & Cruz, 2001). In contrast, empirical findings suggest that insecure attachment is characterized by either extreme of help-seeking, with dismissing individuals being unlikely to seek support or ineffective in their efforts to do so (Collins & Feeney, 2000; Kunce & Shaver, 1994), and preoccupied individuals demonstrating excessive reassurance seeking (Collins & Feeney, 2004; Shaver et al., 2005). Among preoccupied adults, early strategies of heightening attachment behavior, although adaptive for the purpose of maintaining proximity to the attachment figure in childhood, could become maladaptive at older ages when immature dependency and clinging designed to increase care from others interferes with autonomous exploration and the development of self-efficacy. Simpson and Rholes (2002) theorized that fearful adults are unable to determine the viability of support seeking and consequently experience strong internal conflicts. Contradictory approach-avoidance tendencies might manifest in aborted or disguised bids for care and comfort from relationship partners. Although receipt of a partner's caregiving is usually related to relationship satisfaction (Carnelley, Pietromonaco, & Jaffe, 1996), consistent with previous research indicating that avoidant children tend to interpret the behavior of others more negatively (Suess et al., 1992), Collins and Feeney (2004) found that emerging adults who were dismissing tended to interpret their partners' support as unhelpful and noncaring, whereas fearful participants were likely to view their partners' support attempts as hurtful.

On the flip side, the ability to provide care to others in part develops from the experience of being cared for by an attachment figure. Emotionally abusive parents model “care” to their children in atypical and unacceptable ways, potentially leaving them with a distorted understanding of appropriate caregiving. Cassidy (2000) suggested that “giving care” in adult partnerships involves (a) being available in times of trouble, (b) serving as a secure base and being comfortable with a partner's exploration, (c) demonstrating openness and flexibility, and (d) recognizing when a person needs care (i.e., sensitive attunement) and subsequently providing love, acceptance, and respect (i.e., responsivity). The comfort level and ability to effectively provide these things to partners is related to adult attachment style. Research has shown that secure adults are more likely to provide care and support to their romantic partners (Carnelley et al., 1996; J. Feeney, 1996; Simpson, Rholes, Orina, & Grich, 2002), whereas insecure adults are less likely to be sensitive or supportive to their partners and more likely to be controlling (Feeney & Collins, 2001; Kunce & Shaver, 1994; Simpson et al., 2002). In addition, a high level of attachment avoidance (dismissing and fearful styles) is related to low compulsive caregiving (defensive behavior involving caring for others and prioritizing their needs while simultaneously rejecting care from others), whereas a high level of attachment anxiety (preoccupied and fearful styles) has been linked to poor overall caregiving, characterized by less reciprocity, low responsiveness, negative support strategies, or high levels of compulsive caregiving (Brennan, Wu, et al., 1998; Collins & Feeney, 2000; Feeney & Collins, 2001; Kunce & Shaver, 1994). Fearful adults, in particular, might display compulsive caregiving behaviors to cope with their fears and to compensate for a lack of intimate contact with their romantic partner (Collins & Feeney, 2004) or conversely demonstrate overly controlling and punitive behaviors designed to prevent the feared abandonment by the partner.

Sexual behavior

Early maltreatment has been identified as a risk factor for sexual problems in adulthood (Briere & Runtz, 1990; DiLillo et al., 2007; Mullen et al., 1996; Zurbriggen et al., in press). In particular, emotional abuse by a male parent is especially deleterious to female adult sexuality (Mullen et al., 1996). Despite being identified as one of three key behavioral systems involved in romantic relationships, surprisingly little research has been published about the links between adult attachment organization and the sexual system. However, the available research suggests that secure individuals are likely to become sexually active at later ages, enjoy physical contact that is both intimate and sexual with their partners, prefer “normative” sexual behaviors, experience positive emotions after sexual intercourse, and are unlikely to have sex outside their primary relationships (Brennan, Clark, & Shaver, 1998; Feeney, 1999). Contrariwise, insecure attachment is associated with an earlier age of onset and sexual involvement without commitment (Collins & Sroufe, 1999). Schachner and Shaver (2004) hypothesized that early sexual involvement for adolescents high in attachment avoidance might be motivated by the desire to increase one's status and prestige, whereas those with high levels of attachment anxiety might be motivated to engage in early sexual relations to reduce insecurity and foster intense intimacy lacking in other attachment relationships. Research suggests that dismissing individuals are not likely to use touch to express emotions but endorse more casual attitudes toward sex, including promiscuity, “one-night stands,” extra-relationship sex, and pleasurable sex without love (Brennan & Shaver, 1995; Brennan, Wu, et al., 1998; Collins & Feeney, 2004; Feeney, 1999; Feeney, Noller, & Patty, 1993; Hazan & Zeifman, 1994). Preoccupied individuals may be less discriminate about their sexual partners, more willing to engage in risky sexual behavior, and more likely to use sex to seek care and satisfy their needs for closeness and acceptance (Brennan, Wu, et al., 1998; Collins & Feeney, 2004). Findings also suggest that fearful individuals might be averse to intimate touch and experience strong negative emotions after sexual intercourse (Brennan, Wu, et al., 1998). Because fearful adults experience a high level of ambivalence regarding romantic relationships, both craving and fearing intimacy, they might actively seek out sexual relationships but then in response to overwhelming anxiety abruptly withdraw and end contact with the partner to protect themselves from potential pain and rejection.

Conflict resolution and relational aggression

Given the varied coping strategies, it should not be surprising that different attachment styles are associated with distinct conflict resolution approaches in close relationships. Research has demonstrated that during relationship conflict, secure individuals are more likely than insecure individuals to use problem-solving strategies, integrating and compromising behaviors, self-disclosure, mutual discussion, and constructive communication (Carnelley et al., 1994; Corcoran & Mallinckrodt, 2000; J. Feeney, 1994; J. Feeney, Noller, & Callan, 1994; Mikulincer & Nachshon, 1991; Pistole, 1989, 1993). Secure adults are also less likely than insecure adults to engage in withdrawal and verbal aggression (Creasey, Kershaw, & Boston, 1999; Senchak & Leonard, 1992).

In contrast, because insecure individuals suffer from social information processing deficits (e.g., failure to listen carefully, not validating partner's positive messages) related to negative expectations derived from IWMs, they may frequently misinterpret their partners' verbal and nonverbal messages (Creasey et al., 1999). Early maltreatment theoretically contributes to an understanding of close relationships as involving greater conflict and could create a predisposition to interpret even benign or neutral environmental cues as threats, which can then lead to intense emotional responses, distorted perceptions and interpretations of others' behavior, and subsequent preemptive retaliatory behaviors (Bugental, 1993). Consistent with this idea, several studies have demonstrated associations between insecure attachment and attacking or dominating and controlling behaviors, as well as fewer compromising behaviors during conflict with partners (Feeney & Collins, 2001; Marchand, 2004; Shi, 2003). In addition, research has found that dismissing adults tend to avoid or withdraw from conflict, demonstrate low levels of obliging behavior, and are the least likely to use compromising and integrating behavior (Corcoran & Mallinckrodt, 2000; Creasey et al., 1999; Shi, 2003). Preoccupied attachment, on the other hand, is related to high levels of obliging behavior and feelings of hurt and guilt after arguments (Pistole, 1989; Shi, 2003), but also has been linked to a lack of mutual negotiation, greater hostility, and the use of coercion or pressure, blame, threats, and physical and verbal aggression (Corcoran & Mallinckrodt, 2000; Feeney et al., 1994; Simpson et al., 1996). According to Collins and Feeney (2004), fearful adults are likely to interpret conflicts with romantic partners as relationship threatening. As a result, it is conceivable that they could respond to relational conflict with extreme forms of attachment behavior such as intense anger or threats of self-harm.

It is well documented that victims of child abuse are more likely to demonstrate relational aggression and perpetrate intimate partner violence in adolescence and adulthood (Berzenski & Yates, in press; Dodge Reyome, in press; Fang & Corso, 2007; Riggs & Kaminski, 2010; Wekerle & Wolfe, 1998; Zurbriggen et al., in press). Recent evidence specifically suggests that childhood emotional abuse contributes to higher interpersonal conflict, hostility, aggression, and partner victimization in later life (Crawford & Wright, 2007; Messman-Moore & Coates, 2007; Nicholas & Bieber, 1996; Wekerle et al., 2001). Cicchetti and Toth (1995, 2000) suggested that abuse-related schemas formed in childhood and activated in later intimate relationships place individuals at risk for continued abuse by a partner or for perpetrating abuse on a partner. Insecure attachment creates an inherent risk for relationship problems, but early abuse introduces the expectation of abusive or violent interactions with significant power differentials between the victim and victimizer roles (Cicchetti & Howes, 1991; Crittenden & Ainsworth, 1989; Dodge, Pettit, & Bates, 1997). In particular, IWMs of the self as bad, unworthy, or incapable might create a greater likelihood of tolerating poor treatment by others, whereas IWMs of others as abusive, hostile, or rejecting can decrease the likelihood of considering relational aggression and violence outside the range of normal human experience (Liem & Boudewyn, 1999). Once again, because of negative models of both self and others, fearful adults are likely to be the most vulnerable to highly conflictual and potentially abusive intimate relationships.

Attachment theorists have proposed that IWMs are self-perpetuating and persist because individuals tend to selectively enter relationships or caregiving environments that confirm their beliefs about self and others and for which their learned attachment strategies are most suitable (Bowlby, 1969, 1973; Collins & Feeney, 2004; Collins & Sroufe, 1999). This idea is supported by empirical findings indicating that secure children appear to be more attracted to other children with secure histories, whereas insecure children spend more time in isolation or interacting with adults or other insecure children (Elicker et al., 1992; Sroufe et al., 1993; Urban et al., 1991). In addition, when avoidant children were paired with another insecure child (either avoidant or ambivalent), there was a tendency for victimization to occur (Troy & Sroufe, 1987). Similarly, studies have also reported that secure adults tend to more often become romantically involved with other secure adults, whereas insecure adults more often become involved with other insecure adults (Collins & Read, 1990; Kirkpatrick & Davis, 1994). However, insecure men and women do not tend to pair with partners who share their attachment insecurities, but rather complement those insecurities and confirm their expectations. Pistole (1994) theorized that the common couple combination of dismissing and preoccupied attachment styles might contribute to the development of a negative pursuit-distance pattern when an avoidant partner's discomfort with closeness and distancing strategies trigger the preoccupied partner's fear of abandonment and subsequent frantic pursuit of connection, which in turn actually intensifies the dismissing partner's distancing reaction. Taking into account the notion that individuals internalize the whole relationship and can enact complementary roles (Collins & Sroufe, 1999; Sroufe & Fleeson, 1986), theory and research both suggest that individuals who have been emotionally abused in childhood might unconsciously look for romantic partners whose behaviors match their negative expectations and ultimately repeat the pattern of abuse by either perpetrating abuse on their partner or being victimized by their partner. Preliminary support for this hypothesis emerged in a recent study reporting that female college students who reported psychological aggression in their adult relationships were more likely to demonstrate high levels of attachment anxiety and report a preference for dating men with characteristics common to perpetrators of partner psychological aggression (Zayas & Shoda, 2007).

Dutton et al. (1994) proposed a model of intimate partner violence grounded in attachment theory. Dutton et al. conceptualized intimacy anger as a form of protest behavior directed at a romantic partner when a person anticipates separation or abandonment by the partner. In contrast to secure and dismissing adults, who have positive self-models and thus tend not to consciously experience insecurity in relationships, preoccupied and fearful adults have negative self-models that make them chronically anxious about abandonment and rejection, and consequently are likely to experience high levels of anger, jealousy, and emotional instability in relationships, which have been found to be strongly associated with verbal and physical abuse in romantic relationships (Dutton, 1994; Dutton & Starzomski, 1993). Dutton et al. proposed that the fearful classification might be most strongly associated with intimacy anger because the combination of negative IWMs of both self and others would lead to avoidance of intimacy despite an intense desire for closeness, resulting in chronic frustration of attachment needs followed by intense anger. Their findings with a sample of men in treatment for spousal assault supported this hypothesis, showing strong associations of fearful attachment with self-reported anger, jealousy, symptoms of trauma, and borderline personality disorder, as well wives' reports of their husbands' verbal abuse (i.e., dominance and isolation, psychological aggression). Similar, but weaker associations were reported for preoccupied attachment. Dutton et al. speculated that partner relational aggression develops as a consequence of attachment rage and subsequent efforts to dominate or isolate partners might be an attempt to minimize anxiety about abandonment. “The yoked dysphoric modes—anger and anxiety—thus may have a common origin in insecure attachment and operate to generate both abusing and controlling behaviors” (p. 1382).

CONCLUSION

Current theory and research suggest that emotional abuse in early caregiving relationships has a profound impact on the attachment system, which affects individual development at every phase of life and might reach further into the next generation. As depicted in Figure 1, the presented model proposes that childhood emotional abuse initially interferes with the ability to regulate emotions and contributes to insecure attachment characterized by negative IWMs of self, others, or both. In particular, hostile-self-referential parenting is associated with disorganized-avoid-resist infant attachment patterns and possibly later controlling strategies among school-age children and adolescents. These children demonstrate both internalizing and externalizing problems, as well as perceptual biases and social deficits in peer relationships, which interfere with the adaptive development of interpersonal competencies. By adulthood, insecure internal models of attachment comprise multiple maladaptive interpersonal schemas and are manifested in hyperactivating or deactivating coping patterns that create a vulnerability for emotional disruption and psychological disorder. For instance, the approach-avoidance paradox experienced by disorganized infants is likely to be repeated in adult attachment relationships that trigger early maladaptive interpersonal schema related to abusive experiences.

Theoretically, early emotional abuse contributes indirectly via these cognitive-affective factors to high levels of attachment anxiety and avoidance, the combination of which produces a fearful romantic attachment style, which is associated with the most severe forms of psychological disturbance and dysfunction in romantic relationships. In addition, adults who were emotionally abused as children might not be able to resolve early experiences of maltreatment from parents and consequently might demonstrate the lapses of reasoning and discourse that characterize unresolved attachment on the AAI. Lack of resolution to trauma, in turn, has been related to dissociation and other serious psychopathology, as well as frightened-frightening or hostile-helpless parenting, both of which are associated with disorganized attachment in the next generation of infants. Thus, the AAI unresolved classification could contribute directly and the self-report fearful attachment style could contribute more indirectly via romantic dysfunction or psychopathology to another cycle of disturbed parenting in a process of intergenerational transmission of emotional abuse.

More research is needed to provide empirical support for the proposed model. Attachment theory might provide a useful theoretical foundation to guide future studies investigating the long-term impact of childhood maltreatment, and to identify the differential effects of emotional abuse versus other types of child abuse and neglect. The recent research on distinct parenting profiles associated with different manifestations of disorganized infant attachment potentially can inform ongoing efforts to define childhood emotional abuse and possibly distinguish it from emotional neglect. Studies are sorely needed to examine connections between early emotional abuse and attachment organization at all levels of development and to clarify how the effects of emotional abuse differ from other forms of maltreatment, particularly in relation to later personality traits, as well as psychological and relational functioning. In particular, research testing the proposed model's pathways of influence might identify mediators or moderators in the relationship between childhood emotional abuse and romantic relationship functioning. Although retrospective studies are informative, confirmation for a true causal model must await data from longitudinal studies following emotionally abused infants and children into adulthood.

On a more positive note, the available evidence indicates that attachment patterns can shift in response to alterations in the environment affecting the caregiving relationship, making flexibility rather than determinism the general rule of attachment development (Erickson et al., 1985; Thompson, 1991). By tracing the developmental sequelae of childhood emotional abuse that culminates in dysfunctional interaction patterns in adult romantic relationships, it might be possible to identify specific areas to focus on for intervention at various stages of development. Fortin and Chamberland (1995) recommended that intervention efforts target (a) the alleviation of socioeconomic stress and provision of better opportunities; (b) reduction of family dysfunction by modifying parents' unrealistic expectations of child behavior, satisfying parents' psychological and social needs, and eliminating maladaptive behavioral patterns within the family; (c) promotion of the development of skills, awareness, and positive self-concept in parents and children; and (d) facilitation of social support from the extended family network, peers, and community members.

Clearly, the best hope for children goes beyond parenting skills training to include broader ecological considerations as well as more in-depth restructuring of parents' IWMs of attachment. Yet, IWMs function largely outside of awareness, and therefore are resistant to change unless a conflict between the model and reality becomes extremely apparent (Bretherton, Ridgeway, & Cassidy, 1990). Only repeated interactions that explicitly highlight a contradiction with established working models, such as those encountered in a secure relationship with an adult partner (Egeland, Jacobvitz, & Sroufe, 1988) or an insightful therapist (Bowlby, 1979; Bretherton, 1990), can provide opportunities for integrating dissociated mental models and fundamentally changing attachment patterns. In addition to intensive individual therapy, adults with a history of emotional abuse who report relationship problems might benefit from couples counseling that takes an attachment-based approach, such as emotion-focused therapy for couples, which has been adapted specifically for work with couples where one or both partners report a history of attachment trauma (Carbone, 2010; Johnson, 2002). Thus, although emotional abuse in childhood can jeopardize the functioning of the attachment system, it does not necessarily condemn someone permanently to unhappy romantic relationships. With appropriate support and intervention, it is possible to overcome this unfortunate historical disadvantage and find happiness and contentment with a loving partner.

Notes

¹ Some scholars have used the terms psychological abuse and emotional abuse interchangeably (e.g., Thompson & Kaplan, 1996), others have argued that they are distinct constructs (e.g., O'Hagan, 2006), and still others have maintained that emotional abuse is a type of psychological abuse distinguished from psychosocial neglect (e.g., Stein et al., 2000, as cited in Allen, 2001). Although these definitions and distinctions are important, they are beyond the scope of this article, so I have elected to use the term emotional abuse broadly to encompass theory and research regarding psychological abuse, emotional abuse, and emotional neglect that occurs in childhood.