Mitchell, Susan - Integrating Knowledge: Complexity, Science and Policy Chapter 2

10/03/2008

Unpublished Manuscript

This chapter mainly deals with the problem of reduction of complex systems. Author argues for a language that 'tracks the diversity' in holistic behavior versus behavior of the parts. There are many kinds of reduction, from Kim's ontological reductionism to epistemic and methodological reduction. Author takes the position that a reduction that entails 'a "nothing more than" account' will miss 'important features' of the behavior of complex systems. It is 'too meager a conceptual framework'. [This appears to be a stance against epistemic reduction.]

Author first discusses how materialism gives prima facie reasons for ontological reduction. However 'reductionism actually makes a stronger claim, namely that there is some basic level of description that corresponds to the basic fundamental level of matter'. Mill and CD Broad took up some of Aristotle's discussion on "emergence" and put forward an epistemic account of non-reducibility e.g. wetness of water. Once most of these examples were shown to actually be reducible, it seemed reductionism had no barriers. Emergence again appeared in the 1970s when discussing neurological complexity. It is in this context that Kim makes the strong case for ontological reductionism.

Author discusses Kim's reductionism and points out the bold assumption he makes: that 'every material object has a unique complete microstructural description'. Author believes this condition won't be satisfied with type-instantiation (multiple realizeability) rather than token-instantiation. Further, to believe that our descriptions of the world can be unified into one language mistakenly believes that our language can successfully map onto reality perfectly in every case, avoiding e.g. bad grammars and generalities.

Author's most lengthy response is as follows: complex systems aren't all just aggregations of simple parts. There are feedback systems, chaotic systems, 'non-linear' dynamics, etc. that are missed on 'any other account of reduction'. Author describes the flight of groups of starling birds and of foraging bees in a colony. One of the issues here is predictability: an account of the individual bird or bee will not predict how the group behaves.