Emergent behavior rules out the idea of free will from
intelligence. Johnson is saying that by having enough complexity in the rules,
intelligence just falls out from the interactions between these rules. The “uncoordinated
local actions” all add up to the complexities of observable intelligence. These
uncoordinated local actions could be the same as lines in a computer program or
thoughts in a person’s mind. He also extends this idea to things beyond just
instructions. Music is also just a mess of complex patterns. Things that seem
irrational and cultural are really just emergent behavior. Shannon worked on
the border between random noise and useful information. Though encrypted
messages would often seem to be random; they, in fact, held enough information
that they could be decoded plus the message itself. Like those encrypted,
intelligence might seem like random movement or thoughts when seen at the local
level, but when the aggregate is taken together they show how complexity can
build out of the noise. There are also different types of complexity: simple
systems, disorganized complexity, and grey area between the other two. Simple
systems are easy to predict the product of each individual part. Disorganized
complexity is when the problem seems too complex to deal with individual
particles in the system. By modelling the system statistically, the complex
interactions between the particles becomes worthless and the bigger picture can
be predicted with useful results. This is similar to intelligence in how
individual neurons (or ants) are basically worthless to understanding the whole
product of their intelligence, only by looking at the whole picture can the
intelligence be seen as a whole.
No comments:
Post a Comment