Emergence

Emergence occurs when the components of a system produce more complicated behavior than expected. For instance, collective intelligence is one example of emergence. Mathematically, emergence occurs when the whole of the system (i.e. its whole behavior) exceeds the sum of its parts (Luce, 1999).

However, emergence does not mean only that the system surprises. Holland (1998) describes a number of different characteristics that must be present before a system can be said to exhibit emergence:

• Recognizable, recurring patterns
• Rule-governed system (recurring patterns can be related to law-like regularities)
  Dynamic system (system varies over time)
• Adaptive system (system can learn)
• Persistent patterns, changing components

Emergence is a key concept in cognitive science, because one of the advantages of connectionist models and of behavior-based robots is that they usually lead to emergence, and thus can provide simpler-than-expected theories of complex behaviors.