Emergence

An emergent property belongs to a system and not to its parts. Wetness is not in a water molecule, a traffic jam is not in any car, and a flock is not in any bird. Each part follows a local rule and sees only its neighbours, yet a coherent pattern appears at a larger scale with nobody in charge of it. The Stanford Encyclopedia of Philosophy separates weak emergence, where the pattern is fully fixed by the parts but you can only find it by running the system, from strong emergence, where the whole is claimed to have causal powers of its own. The weak kind is everywhere in science; the strong kind is still argued over.

The mechanism is local interaction repeated many times. Ballerini and colleagues measured it in starlings: each bird adjusts to a fixed number of nearest neighbours, six or seven, whatever the distance to them, and that one rule keeps a flock of thousands cohesive. Conway's Game of Life does the same with four rules on a grid. The practical lesson is that studying one part harder will not reveal the pattern; you have to model the interaction. Markets, cities, ant colonies and the capabilities of large neural networks are all cases where the interesting behaviour sits at the level of the whole.

Calling something emergent can be a way of saying you do not understand it, and that stops the inquiry. The starling study is the model of the opposite: the team filmed real flocks, reconstructed positions in three dimensions and measured the rule. Treat a reported emergent jump in a system's behaviour with care too, since whether something looks like a sudden jump or a smooth climb often depends on the metric chosen, so check the underlying quantity. The philosophical claim of strong emergence, that a whole has causal powers its parts lack, is a separate position and still contested; the weak kind is all that the flock, the traffic jam and the market need. Do not let the word carry more than the evidence does.