I will propose and explore the hypothesis that there are exactly three basic ways of representing processes in the world (i.e., naive mental process ontologies) which are conceptually distinct and alternatively implemented as fundamental mind sets. They are used for action understanding as well as control. Each of these ontologies is characterized by a particular causal scheme for representing processes and actions. First, the “material” mental process ontology is characterized by proximate causality, which means physical forward causality. The active units are mentally conceived as material objects (e.g., a falling stone or blowing air). Second, the “teleo-functional” ontology is characterized by ultimate causality, which means that processes and actions are understood in terms of “function”, “purpose”, or “goal”. The active units are conceived as goal-directed agents (e.g., a lion hunting for an antelope).  Third, the “personal” ontology is characterized by free agency causality, which means that processes and actions are understood in terms of “reasons”. The active units are conceived as persons (e.g., a student considering the practical and ethical implications of a job offered to her). These three ontologies are incommensurable because the underlying causality schemes are: None of these three causality concepts can be defined or derived in terms of the other ones. Based on collaborative work with Ruediger Seitz (University of Düsseldorf), I will argue that specific cortical relay nodes in the medial frontal cortex subserve process and action representation on these three levels.