Achieving progress in understanding the cause, nature, and treatment of autism requires an integration of concepts, approaches, and empirical findings from genetic, cognitive neuroscience, animal, and clinical studies. The need for such integration has been a fundamental tenet of the discipline of developmental psychopathology from its inception. It is likely that the discovery of autism susceptibility genes will depend on the development of dimensional measures of broader phenotype autism traits. It is argued that knowledge of the cognitive neuroscience of social and language behavior will provide a useful framework for defining such measures. In this article, the current state of knowledge of the cognitive neuroscience of social and language impairments in autism is reviewed. Following from this, six candidate broader phenotype autism traits are proposed: (a) face processing, including structural encoding of facial features and face movements, such as eye gaze; (b) social affiliation or sensitivity to social reward, pertaining to the social motivational impairments found in autism; (c) motor imitation ability, particularly imitation of body actions; (d) memory, specifically those aspects of memory mediated by the medial temporal lobe-prefrontal circuits; (e) executive function, especially planning and flexibility; and (f) Language ability, particularly those aspects of language that overlap with specific language impairment, namely, phonological processing.