When Is Something Alive? Integrating Bottom-Up and Top-Down Perspectives on Aliveness

What it means for an individual entity to be alive remains conceptually unresolved, particularly for artificial systems that display increasingly life-like behavior. This paper distinguishes life as an evolutionary, population-level process from aliveness as an individual property, and develops a candidate condition for assessing when an entity, biological or artificial, should be considered alive. A bottom-up heuristic identifies structural and functional characteristics commonly associated with biological organisms, but applying this heuristic to an embodied artificial agent shows that such characteristics are not by themselves sufficiently discriminative. A complementary top-down perspective introduces self-referential regulation, understood as the capacity of an entity to internally represent and regulate its own continued existence through embodied causal processes. Integrating both perspectives yields a unified candidate condition for aliveness, proposed primarily for complex entities while leaving open how aliveness should be characterized in simpler ones.