Complexity: the paradox of order and disorder

May 23, 2018 6:14 AM

The Key to Everything (Freeman Dyson reviews Geoffrey West's Scale: The Universal Laws of Growth, Innovation, Sustainability, and the Pace of Life in Organisms, Cities, Economies, and Companies) - "West is now making a huge claim: that scaling laws similar to Kepler's law and the genetic drift law will lead us to a theoretical understanding of biology, sociology, economics, and commerce. To justify this claim he has to state the scaling laws, display the evidence that they are true, and show how they lead to understanding. He does well with the first and second tasks, not so well with the third. The greater part of the book is occupied with stating the laws and showing the evidence. Little space is left over for explaining. The Santa Fe observers know how to play the part of a modern-day Kepler, but they do not come close to being a modern-day Newton."
posted by kliuless (16 comments total)

Let's just get this out of the way: Freeman Dyson is still alive?
posted by Halloween Jack at 6:21 AM on May 23, 2018 [2 favorites]

Let's remember that it's easy to fit data to a power law but that doesn't mean a power law captures the phenomenon any better than some other function. See e.g. this paper by Cosma Shalizi

"The purpose of this note is to argue that, at least for the United States, while there is indeed a tendency for percapita economic output to rise with population, power-law scaling predicts the data no better than many other functional forms, and worse than some others. Furthermore, the impressive appearance of scaling displayed in Refs. [3, 4] is largely an aggregation artifact, arising from looking at extensive (city-wide) variables rather than intensive (per-capita) ones. The actual ability of city size to predict economic output, no matter what functional form is used, is quite modest."
posted by escabeche at 6:46 AM on May 23, 2018 [11 favorites]

fractals are faractals are fractalls and self similar is self simlar.
posted by danjo at 6:46 AM on May 23, 2018

See e.g. this paper by Cosma Shalizi

Seriously that thing needs to be passed out like candy.
posted by PMdixon at 6:54 AM on May 23, 2018 [5 favorites]

fractals are faractals are fractalls and self similar is self simlar.

The B in Benoit B. Mandelbrot stands for Benoit B. Mandelbrot.
posted by Pyrogenesis at 8:08 AM on May 23, 2018 [14 favorites]

A more generalized approach to power law BS "So you think you have a power law"
posted by lalochezia at 8:09 AM on May 23, 2018 [4 favorites]

Also: a pox on both their houses.

While Dyson is clearly one of the greatest minds of 20th century humanity (and I make that assignation in all seriousness), his glibertarian impulses and pronouncements - and the cover it provides for genuine glibertarian idiots/psychopaths - are galling.
posted by lalochezia at 8:11 AM on May 23, 2018

I say this with respect, as someone who worked at the Santa Fe Institute as a very young person decades ago. Dyson's critical review is a 94 year old man criticizing the book written by a 78 year old man. They are both brilliant people and still contribute to science in useful ways. From other reviews of West's book I think it is probably worth reading.

But it's time to look to younger researchers for insights. West and the SFI program is looking for something radical, for a brand new Theory of Everything that explains systems behavior. This goal isn't new; the cyberneticists were doing the same thing in the 50s. So far it's largely elusive. But from what I understand in the history of science, the kind of breakthrough West is grasping for tends to come from someone very young. And somewhat randomly, often from an adjacent field.

Sadly, the best minds of the new generation are thinking about how to make people click ads.
posted by Nelson at 8:42 AM on May 23, 2018 [3 favorites]

I found the diversion into futurism (here's what human diversity and space travel will look like in 200 years) really strange in this review. It made me think the reviewer simply has their own pet theories and would like the reader to know about them, and shoehorned them into a discussion of someone else's.

That makes the generally unfavorable impression of West's book less convincing, on the whole.
posted by dbx at 9:18 AM on May 23, 2018

The Economist has another review of West's book that is much less idiosyncratic than Dyson's.

This reminds me I've been meaning to read Melanie Mitchell's book on Complexity. She's an amazing computer scientist and clear thinker and has been just as central to the Santa Fe Institute complexity program as anyone. It came out nearly 10 years ago, so I've been remiss. She's writing a new book right now.
posted by Nelson at 9:28 AM on May 23, 2018 [3 favorites]

It made me think the reviewer simply has their own pet theories

He does footnote an earlier book review of his we he also pushes his warm-blooded plant agenda. That's permissible in the NYRB, and he's not the only author there to use a book review of as a pretext for a personal essay.
posted by paper chromatographologist at 9:31 AM on May 23, 2018

Does anybody have a good pro/con summary of the genetic drift thing?
posted by bukvich at 9:46 AM on May 23, 2018

Which part, bukvich? Some of it is just algebra, although if you're trying to explain actual populations with it you rapidly get to (interesting, well-founded) arguments over which neutral theory is most applicable.

... In the actual review, I wasn't convinced by the implication that Jerusalem, Venice, and Manchester in their glory days were genetically isolated. (They were ports! Of great trading concenrs!) at which point it seemed like crank vs crank. (And neither of these guys is a total crank, and I like it when researchers near the end of their lives try for a grand synthesis even if they fail, and all of this is better than the Kind-Of science of Wolfram.)
posted by clew at 12:32 PM on May 23, 2018 [1 favorite]

I took the point of the sudden jump into highly detailed highly speculative futurism as an implicit claim that West is basically doing the same thing in a different direction, but maybe I'm being overly generous to Dyson.
posted by PMdixon at 1:35 PM on May 23, 2018

Nelson, thanks for the "Complexity" link. Looks good, picking up shortly.
posted by aleph at 1:50 PM on May 23, 2018

Thank you clew. The Hubbell book looks close to what I was looking for. I thought the genetic drift and population size argument was Dyson's way of dancing (poorly) around the high voltage eugenics maypole and am predisposed to discount the rigor of all scientific reports and commentary. They mostly seem tainted by that old Wilson-Gould-et al brouhaha.

Also Shalizi's power law post is a classic.
posted by bukvich at 7:17 AM on May 24, 2018