There are cases of emotion that we readily describe as 'sharing emotions with other people.' How should we understand such cases? Joel Krueger has proposed the Joint Ownership Thesis : the view that two or more people can literally share the same emotional episode. His view is partly inspired by his reading of Merleau-Ponty -- arguably Merleau-Ponty advocates a version of JOT in his "The child's relations with others." My critical analysis demonstrates that JOT is flawed in several respects: 1) It involves a vague account of joint subjects; 2) It relies on a confusion between phenomenological and ontological levels of analysis. When these are clearly distinguished, Krueger's phenomenological analysis contradicts JOT understood as an ontological claim; 3) It relies on a highly problematic coupling-constitution inference; 4) It relies on a shift from the claim that the child and the caregiver jointly _realize_ an emotion, to the claim about joint _ownership_, which is a _non sequitur_. I argue that we can reach a better understanding of the phenomenon of shared emotions by bringing in another level of analysis: that of _social relationships_. I propose that shared emotions are a special case of _social-relational emotions_, typically arising within and/or giving rise to communal relationships.