Mark Chu-Carroll has a post up that does two admirable things: it deflates yet another creationist and his grandly fallacious claims, and it gives me a new toy to play with.
The first part is a debunking of Granville Sewell, a mathematician and darling of the Intelligent Design creationists. Sewell actually is a professor of mathematics, so it's somewhat embarrassing to see a fellow professional humiliate himself with such ancient, bogus creationist complaints, such as that evolution violates the second law of thermodynamics, or that stepwise change can't occur. I'll also recommend debunkings of Sewell's bad math by Jason Rosenhouse and Mark Perakh.
The second part is fun. Download the breveCreatures simulation software, and play. This is a program that runs blocky, crudely jointed creatures through cycles of evolution, and you can watch rectangles with hinged limbs evolve to walk under the pressure of selection for greater distances of movement. They start out pathetic and feeble, little blocks that flail about and might randomly flick themselves forward a few steps, and after a half hour or so and tens of generations, they're rather clumsily galloping across the screen.
I've only run a couple of the demos—this thing is programmable, too, and there's much more it can do—and the performance is remarkable even with the handicaps imposed on the design. One thing I noted right away in the simplest demo ("Walker.tz") is that there's no symmetry imposed on the systems, so the poor creatures are afflicted with four limbs that may each have completely different properties, making them particularly thrashworthy. There really ought to be something in the code to require the two upper forelimbs, for instance, to have identical controls, with some kind of central regulatory circuitry that could impose phase differences. Are there no structuralists and developmental biologists among the coders at breve?
- Log in to post comments
But you see, that's evidence of intelligent design. The simulation was designed, therefore there must be a program behind our own evolution, which must have been designed.
Goggley moogley, this stuff's easy to parody. And look, I've implied that a character from The Matrix is the intelligent designer. Wow.
Id like to bring your attention to a program I toyed around with about 5 years ago called Framsticks
http://www.frams.alife.pl/
Seems more "advanced" than breveCreatures
But you see, that's evidence of intelligent design. The simulation was designed, therefore there must be a program behind our own evolution, which must have been designed.
Why sure! And there must have been a program before that program, which must have been designed... all the way down to the bottom turtle.
Now you can expect the standard ID objection to software simulations of Evo. They always have an Evaluation Function (EF) that compares design changes against other designs, and so the EF is said to "secretly" drive the design toward a goal. True enough: when you're hungry, you look for food.
But that's a marvelous admission by the ID side of how Natural Selection can work without an a priori plan. The true driver is the goal (food, a mate, a winning chess position), and the EF only compares how far the goal is for alternate designs.
There really ought to be something in the code to require the two upper forelimbs, for instance, to have identical controls, with some kind of central regulatory circuitry that could impose phase differences.
What are you asking for? Wouldn't that be teleological?
"There really ought to be something in the code to require the two upper forelimbs, for instance, to have identical controls, with some kind of central regulatory circuitry that could impose phase differences. Are there no structuralists and developmental biologists among the coders at breve?"
What makes you think that symmetry must apply in all creatures? Just because Earth creatures are rather symmetric does not mean that all creatures in the universe have to be.
Yes, that's one of the things I truly love about the IDers. They're too FSM-damned stupid to understand that they've just given the game away when they make this "objection" to the simulations. It's almost as funny as Dembski being so proud of proving that selection averaged across all possible fitness landscapes goes nowhere.
That's not teleology, it's structural.
And sure, the point of the simulation is that even without a constraint common to all bilaterians on this planet, evolution still manages to crank out a functional behavior.
... that Michael Behe's arguments against neo-Darwinism from irreducible complexity are supported by mathematics and the quantitative sciences, especially when applied to the problem of the origin of new genetic information.
More arguments "against" something, but no arguments "for" something, because they can't have any positive arguments since somebody else could always say that their "poof" is wrong because of this here other "poof" thingy. Besides, why would they use mathematics and the quantitative sciences to argue against anything when they don't like things that don't have "poof" stuff in it? Ha ha, what hypocrites! Lol.
But you see, that's evidence of intelligent design. The simulation was designed, therefore there must be a program behind our own evolution, which must have been designed.
The program mimics a process that happens naturally. It shows that natural selection will improve a design without divine intervention. It doesn't address abiogenesis, it just shows you what would happen afterwards.
http://www.uncommondescent.com/index.php/archives/1348#more-1348
William Dembski is bizarre...
What stage is evoltion as a theory in comparison to the Black Knight?
Ummm... how bout GREAT if you realize that ID is the black knight and the Monty Python made fun of all things religious.
Framsticks looks pretty interesting.
A minor limitation - it's mostly Windows software, with limited Linux and Java support. So most Mac users can't play.
Breve not only has Mac, Windows and Linux versions, the Mac version has OpenGL support and also offers a screensaver module as well as the simulator. All are very CPU-intensive, natch.
Hmm. Perhaps the "real" problem is not the lack of forcing symmetry, but that the cost of inventing new controls is "equal" in all cases to the modification of existing ones? In other words, its fairly easy to produce symmetrical solutions naturally in circumstances where its easier to copy an "existing" system, than the manufacture a completely different one out of no where. The failed constraint isn't in requiring it, its in making it too easy to just make up some completely new class of behaviours, without any prior basis for them.
On symmetry: couldn't Breve be understood as a taxon starting with a kind of radial symmetry like starfish? (...or *cephalapods*?) Not all animals are bilateral, of course. But I don't offhand recall the actual evolutionary history and/or developmental constraints of radial critters.
Bilateralism seems to correlate with a lot of movement, radialism with being fairly sedentary or at least slow. If you move a lot, the basic worm design looks pretty functional: senses, brain, and mouth in the front, anus in the back. A Usenet poster once told me that larval starfish are bilateral, and turn radial as they grow up, suggesting an evolutionary story like that of flatfish, but more so.
If you started with a really radial organism, and it then turned mobile and bilateral, you might have holdovers like quintuplicate organs. Then again, maybe the spares would atrophy.
QUOTE: "But you see, that's evidence of intelligent design. The simulation was designed, therefore there must be a program behind our own evolution, which must have been designed.
Goggley moogley, this stuff's easy to parody. And look, I've implied that a character from The Matrix is the intelligent designer. Wow."
How is it ID if this is an elaborated picture of an observed event? Who's really doing the design here? A natural process, or someone's informed analysis and reconstruction of that process.
This isn't even close to intelligent design.
QUOTE: "But you see, that's evidence of intelligent design. The simulation was designed, therefore there must be a program behind our own evolution, which must have been designed.
Goggley moogley, this stuff's easy to parody. And look, I've implied that a character from The Matrix is the intelligent designer. Wow."
How is it ID if this is an elaborated picture of an observed event? Who's really doing the design here? A natural process, or someone's informed analysis and reconstruction of that process.
This isn't even close to intelligent design.
These simulated critters have no development at all. And because of the simplicity of their genotype-phenotype mapping they don't need any symmetry ( a symmetrical creature would not evolve faster locomotion than an asymetrical one, I've tried many times in framsticks - a similar program ) , so, atrificially imposing it would be teleological.
steve c, that UD post by Demsbki is out there. First off its a ridiculous analogy regardless what the commentors think. Secondly it's another in the long line of posts that shows Dembski and the DI are thick in the fog of ignorance. DaveScot's departure must have sparked something in Billy to put up something wildly insane and off target to fill in.
The train wreck in slow motion over at the Allen Loves Hannah weblog continues
http://evolutionanddesign.blogsome.com/2006/07/19/analogy-induction-and…
Be careful about posting there. If your argument is too cogent or if you actually insist on getting a direct answer to a direct question, you'll be banned.
Which party benefits from such rules, do you think?
GWW, enlighten me - what is the context of your remarks? Clearly you are making a remark to something that has been debated earlier, but I can't recall that blog being mentioned before.
I note after running a few dozen generations that the mechanism for testing fitness (drop the new entity with random orientation onto the floor, then let it twitch) encourages a trait beyond just forward motility: while sometimes you get a form that works well regardless of orientation, more often you get a form that can right itself before proceeding.
That's cool. That's a remarkably interesting behaviour to develop from such a simple test environment and criterion for selection.
T_U_T, thanks for highlighting that fundamental flaw in these models.
There aren't many people who've included any developmental dynamics in their artificial lives. The only one I knew of was F. Gruau for his artificial neural network models. But he did his dissertation and a few publications and talks in the early-mid '90's, including participation in John Koza's Stanford working group on Genetic Programming - and then he seemingly vanished.
Sample citation: Gruau, F. (1995). Automatic definition of modular neural networks. Adaptive Behaviour, 3:151-183.
Oh? Fred Gruau, a name from some distant past (he got his phD
in the lab I was in Lyon). Still alive at Paris XI Orsay, LRI
or the like. Maybe not much kicking research-wise, recently,
but he published at least until 2004.
http://www.informatik.uni-trier.de/~ley/db/indices/a-tree/g/Gruau:Fr=ea…
Yup, that's him. Thanks GuLi!
Any other modelers include a developmental process between genetype and phenotype?
Just had dinner on a boat in Puget Sound with Lee Spector, the advisor of the author of breve.
Note that breveCreatures 1.1 (the screensaver) uses arbitrary morphological rules, so the nature of "leg" and even "walking" need to be discovered by the search as well. Some of our machines here in the house have developed V-shaped walkers, some sidewinder snakes, some pole-vaulters, some vibrating wigglers; each run of the search tends to result in novel solutions to "get as far as possible as quickly as possible". It's cool stuff... but I'd say that. It's what I do.
That said, one issue that needs to be addressed. Genetic Programming is analogous to artificial selection, not natural selection. That's crucial to keep in mind, on both sides of the argument. We do specify a fitness function in evolutionary optimization applications --- that's why it's not a good analogy for natural selection, except insofar as artificial selection is.
For a better analogy, you should examine Tierra, or Avida, or derived systems. Those are open-ended evolutionary Artificial Life systems, and display much more realistic aspects of community formation, coevolutionary dynamics, niche discovery, and such diverse adaptation that one has to do basic research to try to discern just how the creatures have managed to function.
What stage is evolution as a theory in comparison to the Black Knight? Evolution is to BillD as the uber ultimo knight is to the Black Knight. Just as the Black Knight keeps taunting the rest, BillD is trying to raise some laughs from his factotums. From Baylor U all the way down into obscurity is not bad at all. BillD is doing a pretty good Black Knight!
"How is it ID if this is an elaborated picture of an observed event? Who's really doing the design here? A natural process, or someone's informed analysis and reconstruction of that process.
This isn't even close to intelligent design."
Oh, yeah, it's logically invalid alright. It's an argument I've seen from ID advocates before, I couldn't really overlook the chance to lampoon it.
... all the way down to the bottom turtle.
Are you saying that god programs in logo?
If you liked that one, PZ, yer gonna love this one.
Jordan Pollack is an aquaintance and a cool guy. we don't all get to have one of his evolved robots but on the other hand, the fact that they have been built and manage [with increasing effectiveness, of course] to exhibit fitness [eg moving in deep soft sand] makes them a bit more interesting that mere animations.
I'm considering running one of these evolution simulators on my semi-monstrous desktop, since I haven't been using it lately. One thing I'd like for the selection criteria is essentially the search for "food." Anyone have suggestions for which program I should run/mods I should download?
Breve reminds me of a great piece of work by Karl Sims. Here's a brief movie called Evolving Virtual Creatures. The thing I love about it is that it demonstrates a number of very creative locomotive strategies. With design thinking, we have pretty rigid forms of locomotion: "have rotating axle, will travel." But with evolutionary thinking, you get really creative forms of locomotion, and some of them are so close to natural forms of locomotion it's almost eerie.