Imitation abnormalities are often documented in young children with Autism Spectrum Disorder (ASD), however the relevance of imitation to early development and early detection of ASD remains unclear. Recent studies that investigated whether imitation at 12 months distinguishes children who will subsequently receive an ASD diagnosis from other high-risk groups have reported conflicting results. The purpose of this note is to provide a framework to interpret these conflicting findings, which is based on the often-overlooked distinction between the "propensity" to imitate and the "ability" to imitate. We argue that this distinction can critically inform understanding of early imitative behaviour in ASD and the development of early detection procedures.