338 L.A. Sroufe Table 1. Developmental principles illustrated by childhood disorders Disorder AD/HD Autism (PDD) Anorexia Conduct Disorder Depression Observation Changing symptom picture from childhood to adolescence The profound handicap in all social and cognitive areas of functioning Predicted by early perfectionism but not food finickiness The strong persistence of symptoms over time The etiological role of early loss things, that developmentalists have seri- ously underestimated the complexity of so- cial knowledge during middle childhood, a discovery we were led to by first noting an atypical reaction or, rather, noting behav- ior children defined as seriously atypical. More generally, our work on boundary maintenance (between parents and chil- dren, between siblings, and between peers) and coherence of relationships and family systems (Sroufe & Fleeson, 1988) was all inspired by observation of seductive care and other examples of boundary viola- tion (e.g., Sroufe, Jacobvitz, Mangelsdorf, DeAngelo, & Ward, 1985; Sroufe & Ward, 1980). Disorder and Principles of Development Study of disordered behavior itself illus- trates the laws of development, just as does the study of normal behavior (e.g., Robins & Rutter, 1990). Disordered development is characterized by both continuity and transformation or qualitative change over time. Links between early and later disor- dered behavior may be complex, just as are links between early and later normal devel- opment (e.g., secure attachment and later peer competence). In short, development shows the same integrated, coherent nature for normal and abnormal alike. In Table 1, I have listed a few of the disorders com- mon to children and adolescence, along with one developmental principle illus trated by each. Developmental Principle Developmental transformation, coherence within change Integrated nature of development Complexity of developmental links The child as active in creating his environment (internal working models and positive feedback loops) Latent, sleeper effects The many labels for what is now called Attention Deficit/Hyperactivity Disorder (AD/HD) arose in part because of the developmental complexity of this phe- nomenon. The centrality of hyperactivity in the picture during early and middle child- hood for many children led to the earlier labels, Hyperactive Child Syndrome and Hyperkinetic-Impulse Disorder. But cen- tral though it was, hyperactivity is precisely the symptom that most likely fades in ado- lescence, while problems of impulsivity and attention remain. Thus, the emphasis on attention deficits for the last decade. Lon- gitudinal data made it clear that there was considerable continuity in this disorder, but continuity characterized by coherent transformation (Weiss & Hechtman, 1986). It is largely the same children that show this changing array of problem behaviors over time, so AD/HD is well thought of as a developmental pathway rather than sepa- rate disorders at different ages. What was for decades called "autism" or Early Childhood Autism now is subsumed under the label "Pervasive Developmental Disorder," so well does this disorder illus- trate the integrated nature of development (Rutter, 1985). It is not possible to have the profound cognitive and perceptual deficits of the autistic child and not also have seri- ous social impairments. Also, even though most researchers view some type of cogni- tive impairment as central, lack of playful babbling and social responsiveness in in- fancy probably are the most readily identi- Considering normal and abnormal together fied early signs of this disorder. While au- tism graphically illustrates the general developmental principle of integration, so too does every disorder. AD/HD children, of course, have academic problems, and difficulties with both parents and peers, along with the motility and attentional atypicalities. For example, AD/HD chil- dren exhibit more controlling, dominating and negative or aggressive behaviors than do control children (Clark, Cheyne, Cun- ningham, & Siegel, 1988; Cunningham, Siegel, & Offord, 1985). Anorexia nervosa provides an example of the complex linkages between earlier and later behavior. Its developmental origins seem to lie in early perfectionism and rigid compliance, and it has no discernable con- nection to finickiness regarding food, age of weaning, and the like (Wenar, 1990). Given the changing meaning of behavior with development, as well as its hierarchi- cal nature, such complexity is understand able (Sroufe & Rutter, 1984). Linkages are understandable, not capricious or random, but the coherence of behavior over time lies more at the level of meaning than simple behavioral identity. As Wenar has sug- gested, adolescent refusal to take in food and early perfectionism may both be re- lated to the issue of self-control. Again, other disorders also illustrate this complex- ity. I mentioned the absence of babbling in autistics, which could be interpreted as an infantile deficit in imposing meaning on ex- perience. Early predictors of AD/HD also are not obvious (Jacobvitz & Sroufe, 1987). Activity level in the first 2 years does not predict later AD/HD at all. By the pre- school period, distractibility is predictive, but dysfunction in the caregiver-child dyad is predictive even in infancy. While change is always possible, individ- uals nonetheless manifest remarkable con- tinuity across broad spans of life (Robins & Rutter, 1990). Given cognitive develop ment, education, new experiences, and changing social partners as well as sur roundings, why should this be so? Conduct Disorders, the most stable of any common disorder, may provide an important clue to 339 the whole issue of continuity in individual behavior. A likely explanation for the inor- dinate stability of antisocial problems has to do with the feedback these children elicit from others, that is, the kind of social envi- ronment they create. Hostile, aggressive children, for example, have been shown to be disliked and treated negatively by teach- ers (Sroufe & Fleeson, 1988), to be actively disliked by peers (Asher & Wheeler, 1985; Rubin, LeMare, & Lollis, 1990), to elicit more negative and less positive behaviors from them (Pancake, 1985), and to con- strue ambiguous social situations as imply- ing hostility (Dodge & Frame, 1982). Thus, they engage in behaviors that alienate oth- ers and elicit negative reactions, which serve to reinforce negative social attitudes and behavior in an ongoing cyclical man- ner. The disorder leads to conditions that foster the disorder, which is the essence of maladaptation. Other disorders, including autism, lead to withdrawal from required social inputs. Still others lead to cycles of failure to cope and anxiety, such as may underlie hyperactivity. But what is true for Conduct Disorders is true for other disor- ders and for all development. The child ac- tively creates his or her own environment, increasingly so with advancing develop- ment. Many disorders illustrate latent effects of much earlier experiences. The relation between early significant loss and later de- pression is a notable example (Brown, Har- ris, & Bifulco, 1986). Early intrusive behav- iors by caregivers and later hyperactivity is another (Jacobvitz & Sroufe, 1987). Such latent effects are not limited to pathogens but apply to normal development as well, where they are viewed in terms of protec- tive factors or buffers. Work on attach- ment discussed in a later section illustrates this nicely. Infants who have a secure at- tachment relationship with the primary caregiver later have more positive relation- ships with teachers and more close friend- ships with peers, affects that are assessed years later (Elicker et al., in press; Sroufe & Fleeson, 1988). Work with adults (e.g., Main & Hesse, 1990) and observational