People who study visual communication commonly assume that the ability to recognize the content of still or moving images requires prior familiarity with a set of representational conventions. This article argues against this notion. Drawing on recent theoretical developments regarding visual cognition, the article contends that an image's lack of fidelity to visual reality need not impede interpretation by an inexperienced viewer, so long as the image is able to satisfy certain minimal informational requirements of, real‐world vision. An outline of these requirements serves as the basis for a set of predictions about inexperienced viewers’ability to make sense of a variety of representational conventions, and these predictions are checked against available empirical evidence. Since there is in fact considerable research with a bearing on the issues examined here, the article's aim is to provide a theoretical synthesis and explanation of the findings of this research.