Towards a scientific definition of animal emotions: Integrating innate, appraisal, and network mechanisms

Abstract

This paper introduces a mechanistic framework for understanding animal emotions, which is designed for biologists studying animal behavior and welfare. Researchers often examine emotions—short-term valenced experiences—through behavioral, somatic, and cognitive indicators. However, proposed indicators are often ambivalent (emerge in contexts with opposing emotional valence) or undetermined (arise in both affective and non-affective processes). To ground hypothesis formulation regarding animal emotions on a better foundation, the paper advocates for building on what we know regarding the mechanisms of human emotions—the behavioral rules that transform sensory input into motor output during emotional episodes. In particular, it integrates key assumptions from three dominant psychological theories of emotion—innate, appraisal, and network theories—into a single framework and argues that this can serve as a common ground to transfer insights from human to animal emotion research. Additionally, the paper tackles the question of how emotions relate to closely linked processes such as decision-making, distinguishing between parallel architecture models—where emotions and decision-making processes interact but remain distinct—and unified models—where affective states are conceived as integral to goal-oriented processes. Finally, we discuss how our mechanistic proposal can help us address four key questions in animal emotion research: Do animals experience emotions? If so, which animals experience emotions? Which emotions do they experience? And how do these emotions compare to human emotions? The paper concludes by emphasizing the need for further empirical research on the mechanisms of animal emotions and their distinction from other processes.