Attempts to engineer a generally intelligent artificial agent have yet to meet with success, largely due to the (intercontext) frame problem. Given that humans are able to solve this problem on a daily basis, one strategy for making progress in AI is to look for disanalogies between humans and computers that might account for the difference. It has become popular to appeal to the emotions as the means by which the frame problem is solved in human agents. The purpose of this paper is to evaluate the tenability of this proposal, with a primary focus on Dylan Evans’ search hypothesis and Antonio Damasio’s somatic marker hypothesis. I will argue that while the emotions plausibly help solve the intracontext frame problem, they do not function to solve or help solve the intercontext frame problem, as they are themselves subject to contextual variability.