In this article, I evaluate the merits of existing empirical and philosophical theories of collective emotions in accounting for certain established functions of these emotions in the emergence, maintenance, and development of social groups. The empirical theories in focus are aggregative theories, ritualistic theories, and intergroup emotions theory, whereas the philosophical theories are Margaret Gilbert's plural subject view and Hans Bernhard Schmid's phenomenological account. All of these approaches offer important insights into the functions of collective emotions in social dynamics. However, I argue that none of the existing theories offers a satisfying explanation for all established functions of collective emotions in social groups. Therefore, I offer a new typology that distinguishes between collective emotions of different kinds in terms of their divergent degrees of collectivity. In particular, I argue that collective emotions of different kinds have dissimilar functions in social groups, and that more collective emotions serve the emergence, maintenance, and development of social groups more effectively than less collective emotions.