Re: Reductionism - was: Young but distant gallaxies

Posted by Kenneth Lloyd on
URL: http://friam.383.s1.nabble.com/Young-but-distant-gallaxies-tp839193p1075545.html

Steve,
 
"The fact this seems to work ..." that whole line was a touch of sarcasm.  It appears experimentally verified in my work, but people still like to argue with me about it.  Is it possible to argue a phenomenon out of existence?
 
Of course, I can be wrong, but someone will have to prove it by experimental counter-example - not just words.  That doesn't seem to stop people from trying.
 
Re: the question about application to non-probabilistic models - good question!  I'd need to run an example.  Got one? 
 
By all means check out Inverse theory (Tarantola, Mosegaard, Scales).  Powerful stuff.  Scales, ea. has a very accessible book on the web "Introduction to Geophysical Inverse Theory"
 
http://acoustics.mines.edu/~jscales/gp605/snapshot.pdf
 
 
Ken
 


From: [hidden email] [mailto:[hidden email]] On Behalf Of Steve Smith
Sent: Monday, September 08, 2008 9:17 AM
To: The Friday Morning Applied Complexity Coffee Group
Subject: Re: [FRIAM] Reductionism - was: Young but distant gallaxies

Ken -
 
Reductionism has its place in the analytical phase at equilibrium.  Analysis is normally a study of integrable, often linear systems, but it can be accomplished on non-linear, feed-forward systems as well.
Well said... 
The synthesis phase puts information re: complex behavior and emergence back into the integrated mix and may be "analyzed" in non-linear, recurrent networks.
It is the synthesis/analysis duality that always (often) gets lost in arguments about Reductionism.  There are very many useful things (e.g. linear and near-equilibrium systems) to be studied analytically, but there are many *more* interesting and often useful things (non linear, far-from-equilibrium, complex systems with emergent behaviour) which also beg for synthesis.
This is actually a probabilistic inversion of analysis as described in Inverse Theory.
I'll have to look this up.
 
Bayesian refinement cycles (forward <-> inverse) are applied to new information as one progresses through the DANSR cycle. This refines the effect of new information on prior information - which I hope folks see is not simply additive - and which may be entirely disruptive (see evolution of science itself) .
Do find this applies as well in non-probabalistic models?
 
The fact this seems to work for complex systems is philosophically uninteresting, and may ignored - so the discussion can continue.
"seems to work" sends up red flags, as does "philosophically uninteresting".  I could use some refinement on what you mean here.  
Final point: Descartes ultimately rejected the concept of zero because of historical religious orthodoxy - so he personally never applied it to the continuum extension of negative numbers. All his original Cartesian coordinates started with 1 on a finite bottom, left-hand boundary - according to Zero, The Biography of a Dangerous Idea, by Charles Seife.
And didn't Shakespeare dramatize this in his famous work "Much Ado about Nothing"?  (bad literary pun, sorry).


============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org


============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org