The combination of the verbal and the visual track in television news discourse poses a considerable analytical challenge. In the viewers’ minds the co-habitation of these two semiotic channels triggers a complex network of inferential processes, based on expectations of coherence and relevance, with which they make sense of the representation of the world offered in the news. Through the analysis of a number of news items, this article considers the cognitive processes which viewers may activate when extracting meaning from the multimedial messages contained in television news. The analysis of news items from two British television networks offered by the authors traces the possible meanings that, it is assumed, become available to a potential, ‘idealised’ or ‘implied’ viewer, who accesses the information with some social and cultural knowledge of contemporary Britain. Building on existing studies, the article proposes a model for the classification of verbal–visual relations.