Thursday, June 14, 2012

A Cambrian Explosion of Complexity

We've been discussing Bourgeois Dignity: Why Economics Can't Explain the Modern Worldthrough a series of recent posts on this blog, starting here.

By her own admission, McCloskey is eliminating other potential explanations why economic life was suddenly revolutionized in the industrial revolution. So the true explanation (if there is one) must lie in the residual.

For her that explanation is mostly a shift in ideas: an increase in respect and dignity for bourgeois strivers and tinkerers.

But she does not make much of a positive case for that argument in this book. That seems to be left for future books. From interviews and other writing on her site, it seems those will attempt to give more evidence that such a rhetorical shift did in fact take place, using a wider, qualitative range of evidence which economists usually ignore.

Still, despite her compelling and persuasive stress on the importance of encouraging innovation rather than just efficiency, she does not pay much attention to the work on how innovation happens (or does not happen) in other spheres. And not just innovation, but the evolution of complexity.


Evolution and Innovation

One consistent theme on this blog has been to see the economy as an evolutionary system. This originally stemmed from reading Eric Beinhocker's superb book, Origin of Wealth: Evolution, Complexity, and the Radical Remaking of Economics, which I mentioned in the first few days of the blog. But, as we shall see, it leads to a great deal of other work and thinking.

My own suspicion is that the shift in values that McCloskey describes was highly significant, because it facilitated a huge surge in the evolution of complexity of the economy. And that is more than just "talk, talk" acting alone.

There's various ways we can see this.

First, evolution is capable of sudden and stunning increases in complexity. Stuart Kauffman of the Santa Fe Institute develops this in his famous book At Home in the Universe: The Search for the Laws of Self-Organization and Complexity.

He works on simulated computational algorithms of evolution, and had been pondering how life first started. Chance alone was unlikely to generate those first steps in the lifetime of several universes, he argues. But fortunately there is some spontaneous order inherent in nature as well, which can produce sudden phase-shifts in complexity.

In the case of the beginning of lfe, he demonstrates how auto-catalysing sets could in effect suddenly spring into existence, in a S-curve of complexity, and start reproducing themselves. Once complexity hits a critical threshold it grows exponentially. A system goes "supracritical" in a short space of time once a network passes a key threshold.

So that is a first possibility here. There can be thresholds of complexity which, once crossed, can produce an enormous sudden expansion of possibilities.

An Economic Cambrian Explosion


In fact, that is precisely what seems to have happened in the later evolution of life on earth. Basic unicellular life evolved relatively early in Earth's history, not long after the surface of the earth cooled and meteroite bombardment tailed off, perhaps 3.9 billion years ago.

Then little happened for the next three billion years. Life was unicellular until a billion years ago, and evolutionary change was very slow.

That changed within the remarkably short timeframe of just 5-15 million years, the blink of an eye in historical perspective. Most of the current astonishing diversity of life suddenly appeared in that short interval around 530 million years ago, rather than gradual development of features over eons of time.

The Cambrian explosion completely transformed life. Before, there were three major phyla - main groups - of life on earth. After, there were the full thirty-eight or so that we still have today. The famous Burgess Shale in British Columbia is a snapshot of the transition, as celebrated by Stephen Jay Gould and others.

The reasons for the sudden change have been a major scientific mystery, with suggested explanations ranging from increasing oxygen in the atmosphere to the bottleneck effect of "Snowball Earth" freezing of the climate millions of years before.

The most recent work argues that there may have been slow internal changes in internal body form in the Ediacarian age which lay before it. Andrew Parker argues in In The Blink Of An Eye: How Vision Sparked The Big Bang Of Evolutionthat normal rates of mutation suggest that new phyla had in fact developed slowly before the Cambrian, but creatures still looked like similar worms on the outside. Then the development of eyesight triggered an enormous change in creatures' environment that led to a sudden huge radiation of the external forms of animals, as they now had to defend against predators or chase prey.

What has this to do with the industrial revolution? McClosely argues the central fact of modern economic life is the sixteen-fold increase in income, more than can be explained by efficient allocation or capital accummulation.

The number of basic divisions of living things increased more than tenfold in the Cambrian explosion.

Does this prove any similarity? No. But it does show that the evolution of complexity can follow a similar pattern. The more we look at evolution, the more intricate, finely balanced and sometimes fragile the process is.

The key point here is to think of evolution not as just the history of life on earth, and certainly not as the social darwinism (or even evolutionary psychology) that have provoked justified resistance in the twentieth century. This is not the old idea of nature red in tooth and claw.

Evolution as an Innovation Algorithm

Instead, think of evolution as a search algorithm for some kind of fitness relative to an environment, which, given enough time, generates dazzling complexity. It is an astonishingly effective generator of innovation. And in the right conditions there can be a sudden leap, a huge acceleration of evolution.

It is just the kind of small-scale, decentralized tinkering which McCloskey celebrates which can suddenly move that algorithm into high gear.

Evolution does not work in all circumstances. Some kinds of "fitness landscapes" , as Kaufmann explains, can be too random to be searched by evolution.

And you need some mechanism of variation, selection and adaptation. That is obviously a given in the natural world. You have genetic mutation, natural selection and the reproduction of selected individuals over time so that "fitter" genes spreaf through a population.

Similar processes may or may not exist in economic and political life. In fact, they likely do - but for thosuands of years they worked quite feebly. If you have a particular set of institutions and balance, the force of that algorithm creates evolutionary momentum.

But stasis is just as likely, especially as "fitness" is much more ambiguous in economic history. It is not just that variation was more difficult in traditional societies where change was perceived as a threat. But it was largely military or agressive success that was "selected" for, not wealth or innovation.

If you alter that, however, and tune the algorithm, you can suddenly get a surge of complexity with the full force of life itself.

In the next post we'll explore another one of Kaufmann's ideas, that evolution has a natural tendency to move a system to "the edge of chaos" which is where innovative and necessary change is most likely to happen.

No doubt McCloskey would see this as an attempt to reduce the independent significance of ideas, in favor of some other "as-if" process. But it's not. it is an explanation of why ideas can have such a startling effect. Not least because ideas evolve too, as Richard Dawkins' idea of the "meme" makes clear. And sometimes if you unleash the algorithm in a looser environment, a remarkable leap can result.



No comments:

Post a Comment