Obviously the utility of the concept depends on how it is defined, but I have always understood it in the following two senses: 1. The "self-organization" of n

Obviously the utility of the concept depends on how it is defined, but I have always understood it in the following two senses: 1. The "self-organization" of novel entities. The paradigmatic example is the glider in Conway's Life automata. Gliders have properties their substrate entities (the individual life cells, perhaps) do not: direction, motion, information transmission, speed, and a whole logic of destruction/creation (from which, e.g. Turing machines can be constructed). Are gliders as ontologically "real" as the substrate cells and update rules that comprise them? I don't know, but clearly they are novel in some sense, and the term "emergent" seems to naturally apply here. 2. Conflation with the notion of "downward causality". Downward causality is a concept that arises mostly in philosophy of mind. The basic idea is that it seems to be the case that conscious entities control their bodies (including, to some extent, the neurons in their brains - for example if I want to imagine a pink elephant presumably "I" am causing some neurons in my head to fire differently than they were before I was envisioning a pink elephant). And yet, conscious entities as we believe are comprised of their bodies (including, to some extent, the neurons in their brains - presumably their is some necessary subset of neural activity in my brain comprising my conscious volition and the like). Downward causation is the idea that from certain patterns of activity arise sub-patterns that control or have causal influence on the underlying activity. And in this context the term 'emergent' seems appropriate.