The levels of analysis

My main goal in this paper is to propose a reformulation of foundational models in behavioral research, including Tinbergen’s (1963) well-known four levels of analysis (namely, ontogenetic, mechanistic, functional, and evolutionary questions) and Mayr’s (1961) dichotomy between proximate and ultimate causations. After critically evaluating these influential but problematic models, I present a three-level neo-Tinbergenian approach to behavior that considers the triadic integration of behavioral causes, structure, and consequences along a single temporal continuum. I then argue that object-directed play is a good candidate behavior to apply this new paradigm by presenting significant examples of the combined analysis of at least two of these three levels. Finally, I show how stone handling, a form of culturally-transmitted object play in macaques, is perfectly amenable to this unified three-level explanatory framework. My proposed approach fits recent theoretical and empirical advances in behavioral biology, has a heuristic value, and may provide numerous benefits to a range of behavioral scientists.