Reflections on optimality theory (and society)

From: John Collier <ag659@ncf.ca>
Date: Thu 21 Nov 2002 - 19:57:05 CET

Folks,

Sorry for rehashing an old issue, but I have had a chance to think about during recent travels. As you may recall, Pedro and others suggested a number of cases in which optimality has been proposed for a number of complex systems for which some version of "information" is relevant. I raised a number of caveats (that reduce to two) for optimality theory. As I as thinking more on this issue, several thing occurred to me that are important for issues of optimality, complexity and self-organization.

First, there was an interesting session at the Leuven meetings of the International Society for Histroy, Philsophy and Social Studies of Biology on optimality versus self-organization as a source of biological ordering. It was organized by Werner Callebaut. (Brief advertisement: next meetings of ISHPSSB are in Vienna in July 2003, sponsored by the Konrad Lorenz Institute: information at www.kli.ac.at). At that meeting the two approaches were opposed. When I asked the optimality people how they justified their optimality claims, they answered that it was justified by selection theory and the truth of natural selection. I argued that they needed proof of optimality (the hypothesis of optimality) must be tested in each case -- both the optimality and with respect to what. This was seen as superfluous by the optimality theorists, unless there was another competing hypothesis.

I have since wondered if self-organization can also lead to a sort of optimality. I have written about this in several places, but perhaps the most significant are two papers with Mark Burch on rhythmic entrainment, “Symmetry, Levels and Entrainment” John Collier and Mark Burch Proceedings of the International Society for Systems Sciences, 2000, and “Order From Rhythmic Entrainment and the Origin of Levels Through Dissipation” John Collier and Mark Burch, Symmetry: Culture and Science Order / Disorder, Proceedings of the Haifa Congress, 1998  Vol. 9, Nos. 24 (1998): 165-178. There are other things in “Information originates in symmetry breaking” Symmetry: Culture & Science 7 (1996): 247256, “Dealing With the Unexpected. Partial Proceedings of CASYS 2000: Fourth International Conference on Computing Anticipatory Systems, International Journal of Computing Anticipatory Systems, 10, pp. 2130, published by CHAOS 2001, and “Complexly Organised Dynamical Systems”, with C.A. Hooker, Open Systems and Information Dynamics, 6 (1999): 241-302. The basic idea is that self-organization minimizes effort required compared to all other possible ways of organizing. This is an optimality principle of sorts, but not like the usual ones because it is highly integrative over all aspects of a dynamical system, not just with respect to one aspect like selection. I should add that if this perspective is taken, my two caveats about optimality are defeated.

I had not realized that I might have been taken to be arguing against a position I have advocated. Given the use of "optimality" inmost optimality theory. I do not think that the self-organizational type is of the same sort, since it is not goal oriented, but adjusts goals also to minimize effort, which is an end in fact (teleomatic), but not in design (teleonomic or teleological). Since most optimality theory is phrased in design terms, I would say that it is, in general, off the track.

Incidentally, Mark Burch and I suggest social applications in our two papers. My objections to the use of game theory (and its implicit optimality assumptions) are very similar. In complex systems, designed optimality will fail.

John
Received on Thu Nov 21 19:58:14 2002