Rob Sheldon kindly writes to say:
The bottom-up view has been heavily researched. We have learned a few things about it. I’ll make a quick list just to show that we can say a lot more about bottom-up than simply “it doesn’t work.”
1. Random searches
Dembski’s “no free lunch” algorithm shows we can’t get information by random searches. There just isn’t any better way to do random.
a) Gaussian distributions are the fingerprints of random.
b) Diffusive transport is the footprint of random search
c) Diffusion velocity goes as v = x/sqrt(t), so in 4 times the time, you have gone twice the distance. Not a speedy way to get somewhere!
d) Entropy is maximized (Info is minimized) at constant Energy (system adiabatic and closed) for a Gaussian. Think, sodas in the cooler.
2. Non-random searches
Actual observations of, say, grazing bacteria show they do not follow a random pattern.
a) “Fat tails” on the distribution are the fingerprints of non-random.
Entropy is maximized at constant Energy FLOW.
b) “Superdiffusive” transport is the footprint of non-random search
c) Levy-flight (infinite diffusion velocity) v = x / t^n where n > 1/2 and < 1. Not quite as good as knowing where you are going, v=x/t, but still better than diffusion. Especially when working in multiple dimensions.
d) Entropy growth is maximized at constant Energy FLOW. (system not adiabatic, but isothermal and open) The distribution has a fat tail. Think water boiling on the stove.
So Darwin, and most of the 20th century, was looking at closed systems with finite amount of energy. These systems are classic thermodynamics obeying the 2nd law, and never showing spontaneous information conversion.
The latter half of the 20th century, and 21st century have been looking at systems with energy flow, where the fastest growing mode is the one that makes the most entropy. This field is called MEPP “maximum entropy production principle”. Prigogine got a Nobel Prize for this, in part, because people thought it would lead to information creation. You can google it and see just how many people are betting the farm that this will solve the bottom-up problem.
It is true that a MEPP system–open, energy flowing through–does produce structure and structure does encode information. Benard convection cells form if you put a thin layer of fluid between two plates and heat the bottom one, and soon a hexagonal honey comb forms spontaneously. One might imagine another system that draws energy from the honeycomb cells to form its own structure within each cell, and hence a massively cascading complexity that erupts from heating a fluid between two plates. These fevered dreams go under several names–one was called hypercycles, and posited that order could emerge from chaos if the right starting materials were used. The greatest “hypercycle” of all, of course, was the Big Bang, where temperature and gravity gradients generated stars, galaxies and the complexity we see around us today. We even have “models” of various parts of this cosmic hypercycle, though we have to fine tune the dials of “gravity” “density” and “explosive energy” to get the hypercycle to start, and then we have to add in “anti-gravity” and “dark matter” to get it looking like our universe.
Now this is pure physics, of course, and you are right to object that this says nothing about “Origin of Life” or biology. It is admittedly a big stretch to go from cosmology to biology, but if you are a reductionist enough to love physics, then biology just reduces to physics at some microscopic level. Or to say it the other way, biology *emerges* from physics at some macroscopic level, the way galaxies emerge from the Big Bang.
“Emergence” is the mantra of the bottom-up school. The Santa Fe Institute was founded to study this, and has spent 30 years on it, Stuart Kauffman being the most famous name from there.
Did it work?
Not really. It stalled out on the same reasons the cosmology stalled out–all the information was encoded in the fine tuning of the starting conditions. The behavior you wanted to see emerge had to be encoded into the start-up sequence. This would be a version of Dembski’s “No Free Lunch” or “Being as Communion” thesis.
Well, can that fine tuning arise from a Monte-Carlo probability, a Quantum-Mechanical casino? This has been the recourse of the “multiverse” bunch, and they wouldn’t be going there if they had any other answer.
But even QM and even multiverse can’t give us those boundary conditions. The reason goes back to the physics of “fat-tail” distributions above. They are not simply the result of energy flows, they are the result of global coherence. Mathematician Mittag-Leffler back in the 19th century showed that
fat tails occur when diffusion is fractal, when the dimension isn’t integer (1, 2 or 3), but something like 2.375. For example, you are travelling along the coast of Maine to get to Boston, and the journey takes longer the shorter your boat, (because the coast twists more). The coastline is fractal, and transport is not 1-D, but something approaching 1.5D. (If that explanation didn’t help, you can google fractals to see some pretty pictures.) The net result is that the transport equation isn’t a simple derivative, its a *fractional* derivative, and the fractional derivative is both an integral and a derivative, its an integro-differential equation. You have to know what is going on everywhere, in order to know what is going on locally.
That’s a big deal. You can’t do a random search if you have to have a map before you take your first step. Because that map is teleology, it is purpose, it is global information. And that is the one thing random searches and Darwin exclude. In other words, all those cosmology models, all those Stuart Kauffman “emergent” physics models, all those sand piles (in my grad school days, we discussed piling sand into pyramids and how big avalanches would occur so as to keep the sand pile pyramidal shaped, calling it “self-organizing criticality”) had to have GLOBAL boundary conditions if any fine-tuned excitement is going to occur.
So sure, let there be a billion multiverses, but the ones that are exciting have finely tuned global boundary conditions so that they can never be modelled as a random system. You might argue that the Random God made them, but you cannot argue that they evolve randomly, because they don’t, they emerge globally.
Follow UD News at Twitter!