In the centuries-long history of decision-making research, emotion’s role in choice has only been investigated relatively recently. Early theories of decision-making, which conceived of emotions as byproducts of choice, have evolved over time to emphasize emotion’s integral role in choice. These theoretical advances have ushered in an explosion of empirical research on the modulatory influence of emotion in value-based decision-making. And yet, the field currently lacks a mechanistic understanding of where value comes from. While reinforcement learning models have been instrumental in revealing the computational principles by which people update the value of choice options based on feedback, these models shed little insight into the psychological mechanisms underpinning evaluative processes. We synthesize recent research on emotion and decision-making to propose how the field might close this gap: endogenous affective processes help transform choice options and decision outcomes into meaningful value signals. While the precise relation between affect and value—at a computational, algorithmic, and implementational level—remains to be described, such an endeavor will be facilitated by promising advancements in modeling and neuroimaging techniques.