Some young children are not merely insecurely attached and at risk for later problems; by virtue of the severity of their attachment disturbances they are already disordered. This article reviews and critiques the approaches of the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) and the International Classification of Diseases (10th ed.; ICD-10; World Health Organization, 1992) to attachment disorders and finds that they have not made use of findings from developmental research on attachment in developing their criteria. An alternative system of classifying attachment disorders that is compatible with the major findings from developmental research on infant-caregiver attachment is presented. Finally, many areas in need of empirical contributions are indicated.