The potential utility of a distinction between ‘automatic (or spontaneous) and implicit’ versus ‘controlled and explicit’ processes in theory of mind (ToM) is undercut by the fact that the terms can be employed to describe different but related distinctions within cognitive systems serving that function. These include distinctions in the underlying cognitive systems, processes, or representations involved in ToM, distinctions among methodologies or task procedures used to measure ToM, and distinctions among behavioural signatures evaluated as evidence for the engagement of ToM. We propose an approach in which rather than continued dispute over whether or not ToM ‘is’ or ‘is not’ automatic, researchers focus instead on discovering what the range of stimulus conditions and task contexts are that give rise to various signatures of the ToM system. These input–output relations will constrain theorizing about the kinds of representations employed, the types of processing operating over those representations, and the overall architecture of ToM mechanisms.