We follow a theoretical approach to define the concept of reading visualizations. In the past, researchers often assessed readability based on the cognitive processes at work during an individual's engagement with a visual representation. The commonly used term "reading" in these studies, however, often lacks consistency: sometimes it refers solely to the extraction of textual information, while in other instances it is limited to the interpretation of visual signals such as patterns, color gradients, or object sizes. We argue that there exists a gap in the literature for a comprehensive, unifying definition of reading that would potentially broaden the horizons of design spaces and analytical frameworks in our field. To address this issue, we discuss models of reading text and how they can potentially relate to visualization reading.