The frame problem, a long-standing issue in Artificial Intelligence (AI), revolves around determining the relevance of information in an ever-changing array of contexts, posing a formidable challenge in modeling human reasoning. The purpose of this paper is to explore the hypothesis that emotions are able to solve, or at least enable a substantial step towards a solution. I argue that, while emotions are integral to cognitive processes, they do not offer a solution to the frame problem, nor can they play a significant role in a solution. In the course of the argument, I’ll show why some previous claims that emotions could be helpful towards a solution rely on a problematic analysis of what the frame problem is. Ultimately, the conclusion is that no context-dependent feature can play a non-trivial role in solutions to the frame problem, and that is the case of emotions.
