Over at [Dispatches][dispatches], Ed Brayton has been shredding my old friend Sal Cordova.
Ed does a great job arguing that intelligent design is a PR campaign, and not
a field of scientific research. Ed does a fine job with the argument; you should definitely click on over to take a look. But Sal showed up in the comments to defend himself, and made
some statements that I just can’t resist mocking for their shallow stupidity and utter foolishness.
Let’s start with a mangled metaphor from [here][comment-turing]:
>The theory is that a fundamental component of life, the self-replicating Turing Machine, will
>not arise from undesigned primordial elements. Easy enough to falsify.
Life is not a turing machine. Sal, as usual, is trying to throw around terminology to
make it look like he’s saying something deep, when in fact he isn’t. Throughout the
comment thread, he continually invokes computer science, computation, and Turing machines as
part of his argument.
DNA can be used to *construct* a Turing-equivalent computing device. But is it correct to say
that *life* is a turing machine? Clearly not. Is it correct to say that life is a
Turing-equivalent computing device? No. Is it even correct to say that living cells *contain* a Turing-equivalent computing device? Not clear. Cells do a lot of amazing things, but we have not yet developed a sufficiently complete understanding of the internal biochemical processes of a living cell to be able to determine if, modeled as a computing device, living things actually are actually doing anything that requires the equivalent of Turing-complete computation.
A metaphor will help clear up what I’m saying. [Conway's game of life][life], the cellular automata, [is a Turing complete computing system][life-turing]. But *most* actual Life automata grids *do not* perform any Turing complete computations. You can’t grab a life grid that implements a glider gun (like the pattern to the right), and say that the glider gun is
a turing machine, or that it’s Turing complete just because the underlying automata is; it’s not doing anything TC.
But Sal doesn’t get that.
Further down the comments, [Sal comes up again with another boner; this time, a cleverly circular argument to weasel out of admitting the successes of genetic algorithms][ga]:
>And as I’ll point out genetic algorithms in the computer industry are products of design
>running in designed environments. Man-made GA’s hardly constitute a counter-example to the
>questions I pose, and if anything tend to strenghthen the arguments.
>Appealing to intelligently designed as proof that mindless forces can design is disingenous
>Furthermore, the evoltuionary biolgists must give reasonable accounts as to why the known
>physical world of the present and past should be modelled like an intelligently designed
It’s a common criticism from ID folks that genetic algorithms don’t count, because they “run in designed environments”, and therefore strengthen the argument for design. Nonsense, but
convincing to a layman. The basic ID argument comes down to the idea that when we run a GA, we’re “smuggling” design information into the system, and so it’s not the same as biological evolution, where we claim that there is no agent intervening in the process in a way that inserts information. The thing is, there are some pretty damned brilliant GAs out there, which *don’t* always do what we want. They satisfy the selection criteria, but not always the way we
would want them to; not even always in ways that we *understand*. And there are plenty of examples of GA systems that are trying to emulate life, and end up evolving things like symbiotic relationships, parasites, etc – things which the builders of the GA system certainly
didn’t hard-wire into the system.
But what’s more interesting is taking a look at what he’s saying on a deeper level.
First, he argues that GAs *strengthen* the ID claim, because they smuggle information in by way of the design of the environment. That is, he’s making the argument that while evolution appears to be doing things, it only works because of an outside agent’s intervention.
Then he turns around, and says we need to explain why the natural environment in which biological evolution occurs looks so much like a GA system.
There’s a circle in there, if you think about it. GA systems *are modeled to emulate
biological evolution*. Sal wants us to explain *why biological evolution looks like man-made GA systems*. Biological systems look like man-made GA systems, because man-made GA systems are made to look like biological systems.
That is, because we designed systems to emulate what we observe in nature, we need to explain
*why* the things in nature look like the things we designed.
That circle is a key part of the paragraph before: the reason that we must be smuggling in information is because the information can’t be created randomly. The information can’t be created randomly, because the process by which it’s created looks just like the system we designed, which only works because we’re smuggling in information through the design.
There’s really nothing there but the circle: when we imitate nature, we can’t do it without cheating; nature looks like what we did; therefore nature must be cheating, because what happens in nature looks like what we did, and since we know that we must have been cheating, then nature must be cheating because it’s doing the same thing that we did when we were cheating.
One final stupidity from Sal – a classic case of how IDists like to abuse [information
theory][it], and [deliberately mix Kolmogorov-Chaitin and Shannon theories.][shannon]:
>And that’s just for starters. In information science we have the concepts of information
>storage capacity and channel capacity. I’m afraid, even when these rudimentary questions are
>posed to evolutionary biologists we get indications they have hardly looked into the issue,
>or when they do, they find it incompatible with the prevailing paradigm. Rather than
>admitting the paradigm could be totally wrong, they sweep the problem under the rug and
>obfuscate it into oblivion….
Information storage capacity and channel capacity are concepts form Shannon theory. But
when IDists talk about the generation of information in biological systems, they’re using
the Kolmogorov-Chaitin formulation of information theory. When Sal says that biologists brush aside questions about information storage and channel capacities because they
“find it incompatible with the prevailing paradigm”, what he’s really saying is “When we ask biologists a question about the channel capacity of biological information, they respond by saying ”Sorry, channel capacity is a Shannon theory question, but we aren’t studying
information in terms of Shannon theory, we’re using Kolmogorov theory, because Shannon isn’t applicable to what we’re studying!””
When a scientist answers a question by explaining that the question makes no sense, Sal claims victory.