Joined: Aug. 2007
I've just watched a video (http://edinburghcreationgroup.org/intelligent-design.xml) by the Edinburgh Creation Group which I found slightly interesting. Not because it was correct (it's a load of crap, which I'll go into in a moment) but because after years of watching near zero ID output from Dembski, this chap called Kirk Dunston comes along and actually presents what he claims is a working ID hypothesis.
He begins with a long ramble on what ID is, how science really uses it even though evolutionists don't admit it, it's not about God, etc, but then he gets to his Design Hypothesis, which, to spare people having to listen to the rest, is at about 18:00 to 30:00, and it is: Entities which are intelligently designed should have significant amounts of functional information.
We've heard this from Dembski, of course, but he then goes on to define functional information. He claims that this is a new formula from a paper in 2007:
I(Ex) = - log2(M(Ex)/N)
I = the amount of functional information, in bits
x = the function you're considering
Ex = the degree of function x (basically how well it performs function x)
M = the number of ways there are to perrform function x this well
N = the total number of ways function x could be performed.
Beneath all the gloss it's very simple: it boils down to 'how many bits do you need to distinguish between the ways of doing the function well enough, versus the ways of not doing it properly?'
For example (his was crap) suppose there are a billion keys in the world, and you have a door with a lock, and you decide that function x is 'using a key to open the lock'. If it's such a crap lock that any key will open it, then the functional information is zero, because you don't need any kind of specification of how to do it. (in the formula, it's
- log2(1 billion/1 billion) = 0).
If half the keys work, the functional information is 1 bit - the amount you need to distinguish between two possibilities. And if only one key works, then it's the negative log of 1 billionth, which is probably around 20 bits.
It's basically the same as Dembski's negative log of how improbable stuff is, yet he doesn't mention Dembski in the video at all. That alone is suspicious; surely Dunston has heard of him?
Anyway, it gets worse. With this defined, he then goes on to use some bogus math to disprove evolution. Just like Dembski. Here it is:
He first defines any non-intelligent search as 'blind search'. Just like Dembski. So already he's ignoring the cumulative search ability of evolution.
He then claims that for a natural process (which uses 'blind search'), M(Ex) can only reasonably be 1, and N will be the number of trials in the blind search, which, by considering a hypothetical population of billions of organisms mutating, he works out to be 10^42 trials. Plugging this into the formula then gives: 140 bits of functional information which can possibly be produced by natural processes.
This, of course, is such a load of nonsense that I find it difficult to believe he isn't a fraud.
Even if you accept that evolutionary search is a blind search (it isn't), he has plugged these numbers into the formula for no reason whatsoever. The formula measures the amount of information required to distinguish a good solution from a bad one. It says nothing about the amount of information inherent TO a solution. It says nothing about performing a search, either. Dunston does not define a function x; he does not say how many possible ways there are of doing x, and therefore cannot say how well nature can do x. I can only assume that Dunston is trying to say 'there are 10^42 possible ways of doing this and nature can only do one', which is meaningless.
I mean, am I wrong here? Has Dunston just totally shoved in these numbers to sound sciency (just like Dembski)? I cannot believe he did this without deliberate intent to deceive, since he apparently does this in his job.
He then goes on to give examples of intelligent design and how it has more functional information, and goes into tornado-in-the-junkyard mode (just like Dembski) and says that just one protein has hundreds of bits of information, a minimal genome needs about 267,000 bits.
By this time I was fed up with this nonsense, and frankly I don't know how he can do his job and mess all this up so badly. He doesn't even give a 'conservation of information' law like Dembski does, he just assumes that there is one, that you can add up amounts of it, and also that functional information is a universal value and not totally dependent on how people define a function.
In fact, the actual mechanism of evolution only comes up once, and he dismisses it as 'vague'. Nice personal incredulity there.
Possibly the laugh of the video comes at the end, where he moans that scientists always unfairly accuse ID of being all about religion. Who's he complaining to? A group of people at an event organised by Edinburgh University's Christian Union.
Anyway, I just wanted to get that off my chest. It really annoys me to see these people and not know if they're totally incompetent (which is worrying) or deliberately lying (which is annoying). If he is merely stupid, then I'd like to offer him the following advice: scientists aren't shunning ID because it's religious. They're shunning it because, like you've just shown, it's a load of bullcrap.