Brightly coloured textual inserts, which often occupy a sizable part of the TV screen, have become a key feature in Japanese TV. This paper unpacks the contribution of such multimodal stimuli to inference, and the consequences this has for the interpretation process. Using data derived from a mixed-methods approach (i.e. eye-tracking and a multimodal content analysis), we evaluate the relationship between cognitive processing and communicative stimuli. We demonstrate how typographical features (colours and fonts) are used as highlighting stylistic devices by TV producers to manipulate the viewer comprehension process by guiding the audience to an intended interpretation. The results suggest how editorial choices regarding typographical features to trigger certain effects might be subsumed under the current view of style in relevance theory.