<p>Prinz argues that emotions represent relational properties that are matters of concern in the organism-environmental relationship, such as danger or loss. While I agree that emotions represent relational properties, I argue that his account fails to provide an adequate mechanism for how this representation occurs. Drawing on constructionism about emotions and the predictive processing framework, I propose an alternative account in which emotions are constructed through the brain’s domain-general predictive system. On this view, emotion concepts operate as structured, hierarchical priors that integrate interoceptive and exteroceptive signals, playing a constitutive role in determining emotional content. By embedding emotion construction within the brain’s predictive architecture, my account offers a novel neurocomputational framework for understanding how emotions represent evaluative significance in a context-sensitive and dynamically embodied way.</p>