Getting the Picture

In the early 20th century, Gestalt psychologists seriously challenged prevailing notions regarding human perception. They showed that there is a difference between seeing the pixels that make up a picture and understanding what a picture represents. We have all had that “aha” moment, for instance, where a scene suddenly becomes clear (e.g. “oh, it’s a smiley face”). The more general point is that people may have all of the information needed to draw a conclusion yet---in contrast to standard economic models---they fail to connect the dots. We build a model that conceptualizes this idea. An agent’s task is to learn whether a picture possesses some feature (such as whether it depicts a smiley face). They have a knowledge set consisting of “codewords” that they think apply to the picture. This set initially contains codewords for each pixel’s color, but no codewords describing the larger picture. The agent adds to their knowledge set by loading existing codewords into working memory and drawing conclusions. Importantly, the agent has limited working memory, which bounds their ability to draw conclusions. We show that the model captures a number of important phenomena, such as multi-stable perception, and provides a useful conceptualization of narratives as “big-picture statements.” We explore several applications, including to the politics of persuasion.