04 September 2005

Emergence

Finished reading Emergence by Steven Johnson, an introduction to the young science of self-organized complex systems. (Did I use those buzzwords correctly?) I think it's an interesting field because it explores how nature's systems have properties quite different than modular man-made stuff:

Brains, trees, ant colonies, and slime mold are typified by:

  • Many nearly-identical very simple units (neurons, ants, etc.)
  • Ability to deal with new and unfamiliar environments
  • Robustness to partial damage (e.g. Phineas Gage)
  • Lack of a central program or directing-agent (e.g. no "leader" per se, it's distributed and implicit)

Whereas the systems that people design - such as computers, vehicles, and business organizational structures are typified by:

  • Several modules of very different types with strongly-specified inputs and outputs (e.g. your video card uses the AGP convention to get signals from your PC, and DVI to communicate out to your monitor)
  • Strong sensitivity to minor damage
  • Complete lack of sense in new environments
  • Very strong notion of a "central commander"

So, anyway, I like reading about this emergence stuff but have found very few books about it. Unfortunately I did not enjoy Johnson's Emergence because it seemed very topical, failed to reference many of the field's pioneers, and completely lacked mathematical or algorithmic meat.

A particularly scathing review is at Amazon:

Steven Johnson's "Emergence: The Connected Lives of Ants, Brains, Cities, and Software" (Scribner, New York, 2001), is a very bad book, shallow, careless, and disappointing. I was lured by its nominal subject, which interests me greatly, and now I'm sorry I bought it. Mr. Johnson is a young - very young - video gamer who has managed to parlay a superficial aquaintance with the vocabulary of modern science into a series of trendy popular books, incomprehensibly praised by such authorities as Steven Pinker and Esther Dyson.

The book opens with a fraudulent pictorial simile, juxtaposing a side view of the human brain and a map of Hamburg ca. 1850. Indeed they do resemble each other, and the reader is supposed to infer (with no help from Johnson) that the resemblance arises from the operation of similar governing principles. Quite apart from the validity of this conclusion, it apparently does not trouble Johnson that the brain is three-dimensional and the city map is essentially two-dimensional, or that the comparison would fail if a frontal view of the brain had been chosen, or if
Paris or El Paso or Denver had been chosen instead of Hamburg.

It gets worse. At the most fundamental level, after reading the book I find it impossible to say what the author means by "emergence", his nominal title. When he discusses ant colonies it appears to mean swarm intelligence; when he discusses video games it appears to mean interactive software; at still other places it appears to mean whatever recent developments in the realm of computers or biophysics or city planning that he approves of.

Moreover, he appears to be totally ignorant of all science and mathematics that preceded his own adolescence. Although he has a great deal to say about self-organizing systems, you will search the index in vain for the names of John Conway, Oskar Morgenstern, John von Neumann, Stanislaw Ulam, Stephen Wolfram, or most of the other pioneers of the field. When he does recognize a figure from antiquity (i.e., pre-1970), it is with worshipful adulation. He italicizes the name of Marvin Minsky as if he were a demigod, and finds a book by Norbert Wiener "curiously brilliant". What exactly is the curiosity?- that a brilliant mathematician should write a brilliant book?
Likewise, you will find no entry in the index under "Boolean networks" or "cellular automata" or "crystallization" or "ferromagnetism." Under "entropy" you will find only the ludicrous assertion that in nonequilibrium thermodynamics "the laws of entropy are temporarily overcome." In short, Mr. Johnson gives new meaning to the phrase "born yesterday," a degree of ignorance and juvenile solipsism that borders on arrogance. I note that other reader-reviewers assert that the book will provide lay persons with an introduction to a new science. No, it won't. The only thing it will provide is an introduction to bad science.


Here are some related sites and books:

2 comments:

Anonymous said...

This blog is awesome! If you get a chance you may want to visit this Free PDF Creator site, it's pretty awesome too!

Professor B said...

Dude, did your blog get spammed? That is so intensely lame...