Although most research on autistic behavior has considered autism categorically, the increasingly apparent genetic and phenotypic complexities of autism are prompting a more dimensional approach to this area. The long-standing interest in a less categorical approach is made clear from a review of literature. The accumulating empirical support for viewing autism-related phenomena as separable and fractionable is outlined and includes data indicating that many of the behaviors occur in isolation in family members and the general population, are not highly correlated within individuals, and appear to be inherited separately. However, it is emphasized that some of the most common and characteristic phenomena observed in individuals diagnosed with autism do not run in their families. It is suggested that these novel, "emergent," phenomena may arise in the individual from interacting configurations of co-occurring traits or from the interaction of genetic and biological factors underlying the traits. A number of autism-related phenomena including intellectual disability, seizures, persistence of primitive reflexes, stereotypies, self-injurious behavior, savant abilities, and morphological abnormalities, among others, are discussed as potentially being emergent. It is concluded that consideration of the role of emergence in autistic behavior and related phenomena should complement a reductionist approach and might help illuminate the components and complexities of autism.