The aim of this paper is to provide a cognitive semantic analysis of multimodal viewpoint phenomena by conducting a case study of political campaign posters from the United Kingdom’s Conservative and Labour parties. It provides a qualitative account of a selection of posters within the framework of Discourse Viewpoint Space (Dancygier and Vandelanotte 2016), specifically exploring discourse viewpoint configurations in the posters with a special focus on their texts (e.g., lexical choices and style), visual images, and the viewpoints of different participants (i.e., current speaker/addressee and represented speaker/addressee). We collected 87 posters from the Advertising Archives (http://www.advertisingarchives.co.uk) in August 2018. We categorized the selected posters into three types: (1) those explicitly encoding the speaker, (2) those explicitly encoding the addressee, and (3) those explicitly encoding both. We take a detailed look at each type, focusing on whether the viewpoint of the current interlocutor is aligned with that of the represented interlocutor, assuming that the current speaker is equivalent to those who design the posters, and the current addressee to those who view the posters. Based on this inductive functional taxonomy, this paper discusses how the multimodal posters’ political messages are constructed and construed in terms of the levels of hierarchical viewpoint networks, which include local perspectives and more comprehensive discourse viewpoints, as well as viewpoint alignment between the participants. The study shows that the viewpoint configurations of the posters are key to the articulation of the intended messages such that they fit the values of the two political parties. The study also supports the claim that multiple viewpoints are intrinsic to meaning construction, and to the conceptualization of multimodal expressions (Sweetser 2012). (Hankuk University of Foreign Studies)