Abstract

The standard cognitive explanation for the emergence of human communication is that it rests largely on the expression and attribution of communicative intentions which are, in turn, enabled by complex metarepresentations of mental states. This complexity is at odds with the limited metapsychological abilities of infants. But mentalistic metarepresentations are neither necessary nor sufficient in explaining communication. Coded ostensive signals (e.g., eye contact) and established channels (e.g., speech) allow that communicative episodes be identified through decoding rather than metarepresentational inferences. Thus, some metarepresentations may be unnecessary. However, metapsychology is also insufficient for explaining communication: the logic of instrumental actions permits interpreting their effect as following from intentions, yet the effect of communicative actions is often unavailable for inferring meaning. Moreover, current evidence for the developmental trajectory of communication and mental state attribution does not support the emergence of the former from the latter. My proposal is that our primitive concept of communication targets, instead, representational action. When we communicate, we typically convey a propositional content that is detached from our acts—a property absent in ordinary goal-directed actions. This view additionally raises the possibility that metarepresentational capacities evolved for representing external, communicative representations and were only later exapted for other purposes.