A number of theoretical orientations and research directions (e.g., temperament and genetics) may assist in the investigation of socio-emotional development, its continuity during the preschool years and the early determinants of mental health and psychopathology (e.g., Bernstein, Borchardt, & Perwien, 1996). However, attachment theory is widely recognized as providing important theoretical insights into these processes (Main, 1999; Sroufe, Carlson, Levy, & Egeland, 1999) and in providing an important bridge between the biological and the psychological domains (Balbernie, 2001; Holmes, 1993; Siegel, 2001).
Attachment theory comprises a set of propositions that attempt to explain the way in which biological dispositions and early experiences predispose the individual to psychological health or pathology (Bowlby, 1969, 1973; Sroufe et al., 1999). The attachment system is considered to be a basic, inborn, and biologically adaptive motivational system that drives the infant to seek proximity to the primary attachment figure (or caregiver) in cases of danger or need (Bowlby, 1969, 1973). The quality of early attachment is determined by the interactions between attachment figures and the infant. In particular, the accessibility and responsiveness of attachment figures to the infant's emotional signals are considered pivotal for the organization and regulation of the infant's emotional experience (Bowlby, 1969). In the psychobiological literature, several processes have been implicated in the mediation of the attachment bond (e.g., Insel, 1997; Schore, 2001). For instance, Schore (2001) defined attachment as the regulation of biological synchronicity between organisms within species. According to this view, several processes are fundamental in the mediation of the attachment bond such as attunement and the interactive mutual training of physiological rhythms within the caregiver-child dyad (Schore, 2001).
According to Bowlby (1969, 1973), the emotional bond between an infant and their main caregiver affects their later social, psychological and biological capacities through the construction of internal representations or internal working models (IWMs). IWMs are internalized representations of the "self" and "other" based on a child's interactions with their main caregivers. It is hypothesized that these representations contain information about whether the caregiver is perceived as a person who responds to calls for support or protection (IWMs of other), and whether the self is worthy of receiving help from anyone, in particular the caregiver (IWMs of self). For instance, Bowlby (1969) maintained that a child experiencing their parents as emotionally available, responsive, and supportive will construct a self model as being lovable and competent. Conversely, experiences of rejection, emotional unavailability, and lack of support will lead to the construction of an unlovable, unworthy, and incompetent self model.
Ainsworth's (1969) "Strange Situation" research paradigm enabled the classification of children's attachment behaviors into three categories: secure; insecure-avoidant; and insecure-ambivalent. Children in each of these categories were found to differ in the strategies they used to regulate their distress. For example, following separation-reunion, some insecure children showed ambivalence and resistance when approaching the parents (anxious-ambivalent infants), whereas others withdrew from parents (anxious-avoidant infants). On the other hand, securely attached children sought proximity following separation and reunion, and were easier to settle and return to playing than infants in the other classification groups. The parental pattern associated with such infants was sensitivity and responsiveness to the infant's signals and communications. Later an additional group of children that were initially considered to be "unclassifiable", was categorized as "disorganized" (Main, 1991). These infants were shown to typically exhibit stereotyped behaviors in the caregiver's presence (for example, seeking the caregivers and then avoiding them) as well as states of disassociation (Lyon-Roth & Jacobvitz, 1999).
Ainsworth's classification demonstrated how different attachment experiences may lead to individual differences in the expression of attachment behaviors. A number of studies have found the expected relationships between attachment classifications and perceptions of self (e.g., Cassidy, Kirsh, Scolton & Parke, 1996; Jacobsen & Hofmann, 1997; Verschueren & Marcoen, 1999). Attachment classifications have also been reliably associated with vulnerability or resilience to childhood psychopathology (See Greenberg, 1999).
Thus, IWMs have been conceptualized as a set of conscious or tacit expectations and attitudes with respect to one's attachment figures and oneself. These eventually become the basic components of an individual's self-worth and subsequently they regulate the individual's expectation of others, such as teachers and close friends (Bretherton & Munholland, 1999; Guidano & Liotti, 1983; Howes, 1999). Such expectations further contribute to the individual's perception of human nature and the world as being more or less trustworthy and controllable (Bretherton & Munholland, 1999; Catlin & Epstein, 1992). In short, child-caregiver interactions are believed to be fundamental in the development of an individual's view of self, other, and of the world.
According to attachment theory, IWMs influence the child's perceptions of the self and world by predisposing the child to select and interpret incoming information in a way consistent with his IWMs of self and others. That is, children extract and detect regularities in the environment, generating and testing hypotheses consistent with their expectations of the environment (Guidano & Liotti, 1983; Sroufe et al., 1999). Indeed, findings suggest that children who differ in their attachment classifications also interpret their environment differently. For example, differences between secure and anxiously attached children have been found in their reactions to cartoons depicting potential social conflict (Suess, Grossmann & Sroufe, 1992) and memories of affective-cognitive stimuli (Belsky, Spritz, & Crnic, 1996).
Further, research has revealed that attachment classifications are relatively stable over time (Hamilton, 2000; Main & Cassidy, 1988; Verschueren, Buyck, & Marcoen, 2001; Wartner, Grossmann, Fremmer-Bombik, & Suess, 1994; Waters, Merrick, Treboux, Crowell, & Albersheim, 2000) and that children with positive "self" and "other" models display increased self-confidence, develop better social skills, have a more positive view of their peers, and have closer, more supportive friendships than their insecurely attached peers (Cassidy et al., 1996; Jacobsen & Hofmann, 1997; Verschueren & Marcoen, 1999). In fact, a recent meta-analysis of 27 longitudinal studies examining attachment related internal representations suggested attachment security is moderately stable across the first 19 years of life (Fraley, 2002). Attachment related anxiety (i.e., separation anxiety) was also linked with the later development of anxiety disorders such as panic disorder (de Ruiter & Van Ijzendoorn, 1992; Manicavasagar, Silove, Curtis, & Wagner, 2000; Silove, Manicavasagar, Curtis, & Blaszczynski, 1996; Thomson, 1986). Finally, it has been proposed that early attachment experiences influence early brain development (Schore, 2001). For instance, severe attachment failure has been linked with impairments in the early development of the right hemisphere of the brain stress coping systems and maladaptive infant mental health (Schore, 2001a).