Narrative conjunction's junction function: A theoretical model of "additive" inference in visual narratives

Visual narratives often depict images of individual characters without showing the larger scene, meaning that this whole spatial environment must be inferred from these component parts. However, few theoretical models of narrative or discourse have attempted to explain the generation of such “additive” inference. This paper explores the complex interactions between narrative structure and meaning within these types of discourse phenomena, situated within the model of Visual Narrative Grammar based on Jackendoff’s (2002) Parallel Architecture of linguistic structure. Narrative “Conjunction” repeats a single narrative category within a broader constituent, allowing for expansion of a sequence beyond the canonical narrative arc. These conjoined units then correspond to semantic structures in a variety of ways, allowing an “additive” inference of actions, scenes, characters, and/or semantic associative networks. This simple yet powerful architecture enables us to account for a large variety of phenomena in visual narratives and other discourse contexts, while providing a structure that can be tested in empirical research.