Another smackdown of Dembski & Marks

| 128 Comments

As most readers know, William Dembski and Robert Marks recently published a paper in an IEEE journal that purports to show that

In critiquing his [Dawkins’] example and arguing that information is not created by unguided evolutionary processes, we are indeed making an argument that supports ID.

Various science bloggers have critiqued it; see here, here, and here for examples.

Now the Metropolis Sampler has published a more technical analysis of the paper, concluding that

The fundamental lesson here is that the Dembski-Marks approach to evaluating model assumptions is both arbitrary and a poor reflection of scientific reasoning. Model assumptions are not accepted or rejected based on a numerical measure of how many logical possibilities that are ruled out or how far probability distributions deviate from uniform measures. Rather, model assumptions are accepted or rejected based on predictive and descriptive accuracy, domain of applicability, ability to unify existing models and empirical knowledge, and so on.

ID creationists persistently use models that misrepresent theories (or in the case of the WEASEL hoorah, misrepresent what the model is intended to represent), and then conclude (on the basis of syntactic manipulations of the model) that the theories are invalid. Dembski, of course, is a serial offender in this respect, and it’s a pity that he’s inveigled Marks into sharing his delusions.

128 Comments

In critiquing his [Dawkins’] example and arguing that information is not created by unguided evolutionary processes, we are indeed making an argument that supports ID.

Does Dembski ever give a rational for that? Or is it just a bald assertion. Lol. What a ridiculous statement. And there’s that weasel word “unguided” again. People are supposed to know what it means? How come laws of nature can’t “guide” things? (Actually they do know what it means. It means “without Jesus”.)

Um, evidence that may not convict Johnny of stealing doesn’t necessarily exonerate Jimmy. …

“Although the monkey/Shakespeare model is useful for explaining the distinction between single-step selection and cumulative selection, it is misleading in important ways.” –Richard Dawkins

Well, I guess I’m glad Mr. Dembski agrees with Mr. Dawkins about that. (I guess.) Lol.

Anyone who has done even the simplest of computer programs to solve for root of equations would recognize that simply defining the problem to be solved also hauls in the characteristics of the allowed sets of solutions and the domains over which the functions are defined.

Even the Newton-Raphson method for finding the zeroes of a function makes use of differentiability.

Things like continuity, differentiability, analyticity or whatever other characteristics apply to the possible solution sets are all used in the computer algorithms that solve problems. You can’t even talk about solving a problem without envisioning the allowed solution sets and their characteristics.

It is also a scientific fact that much of the world of physics and chemistry, and by implication biology, are very well represented by mathematical functions which have the properties of continuity, differentiability, analyticity as well as domains over which the functions have any meaning.

Even in cases involving discreet solutions, such as those in quantum mechanics, we can know from Schrödinger’s equation, for example, what the distribution of probabilities might be.

And in highly non-linear problems that model much of the physical world, we see emergent properties and increasing complexity that accurately mimic the real world.

So D&M’s objections are both childish and churlish. Our understanding of nature is reflected in the accuracy of our computer models in producing observed phenomena.

We don’t smuggle in the answer, as they keep saying; we put in the processes of Nature as we understand them. Those processes lead to the “answers” in our programs, just as they do in Nature.

This blog is listed on

www.creationism.com

Dueling news feeds!

Keep an eye on the other guy!

Mike Elzinga said:

We don’t smuggle in the answer, as they keep saying; we put in the processes of Nature as we understand them.

In much the same way that a game simulation conforms to reality: “Nope, the character can’t walk through a rock, gets hurt if he falls off a roof or gets run down by a car, has to swim across water, and so on.”

“But it’s arbitrary to sneak all that in!”

“No, it would be arbitrary if we DIDN’T. Elmer Fudd may not actually drop until he looks down after walking off a cliff, but a good gamesim RESPECTS THE LAWS OF PHYSICS.”

Incidentally, Dawkins discussed this particular issue in his chapter on spiderweb evolutionary sims in CLIMBING MOUNT IMPROBABLE.

And of course, it’s just another lazy attempt by creationists to “prove that evolution can’t be possible” without addressing biology at all.

What about the very obvious measurable fact that when nucleic acids replicate in biological systems, the sequence of the copy strands is virtually never exactly the same as the sequence of the template strand (as would be predicted by everything we know about physics and chemistry)?

What about the undeniable fact that some nucleic acid changes will lead to phenotypic changes, and that some phenotypes will have a relative reproductive advantage, making it statistically likely that associated alleles will increase in frequency in the population?

What about the obvious fact that distribution of alleles can also be impacted by what we humans experience as “random chance”, and that such effects can be quite strong when population sizes are low?

All of the above can be observed and quantified with ease. If the above is happening, and it is, then life MUST evolve.

How does Dembski square his abstract claim that evolution must be impossible (and that is clearly what he is claiming) with the simple observed fact that it must be happening?

Also, of course, even if we throw up our hands and “agree” with him that evolution as we now understand it “must be impossible” due to some arcane characteristic of “information”, what is his explanation of the diversity of life on earth? If grizzly bears and polar bears aren’t more similar to each other than they are to raccoons because of evolution, how does HE explain life’s diversity? If he can “prove” that our major explanatory theory is wrong, what does he propose to replace it with?

wile coyote said:

“No, it would be arbitrary if we DIDN’T. Elmer Fudd may not actually drop until he looks down after walking off a cliff, but a good gamesim RESPECTS THE LAWS OF PHYSICS.”

In fact, this is exactly why the Road Runner cartoons were so funny. It was the ways in which the laws of physics were violated so unexpectedly that caught your attention and made you laugh.

Everything from Acme rockets to Acme fishing rods to Acme bird seed, and on to painted tunnels that Road Runner could enter and from which trains came screaming out; all these tricks indicated that the cartoonists knew their physics.

And they had a better grip on reality than do the idiots at DI.

Any student of physics knows that cartoons have their own Laws of Physics.

DNAJock said: Any student of physics knows that cartoons have their own Laws of Physics.

Also memorialized at http://en.wikipedia.org/wiki/Cartoo[…]s_of_physics

What is also troubling is how the author of the Metropolis Sampler demonstrates how Dembski and Marks skewed their data:

In order to reach the theoretical lower bound, which in this case is 93.28 nats (or 134.6 bits), the Weasel program must be modified so that fitness is evaluated on an absurdly biased scale with 10^40 possible unsatisfactory fitness values for every possible satisfactory value.

Simply put, MetSam’s analysis of their results shows D&M (as I suspected) re-wrote the “Weasel Program” until it gave them the results they wanted.

Not to mention the bird seed as bait, for a bird that would prefer a nice juicy lizard or snake… ;)

harold said:

How does Dembski square his abstract claim that evolution must be impossible (and that is clearly what he is claiming) with the simple observed fact that it must be happening?

The first time I ever read any of Dembski’s stuff, I was immediately struck by just how naive he was about the real world.

Even of more concern has been his continued resistance to criticism. His god complex kicked in pretty early, so there is nothing he can learn from “mere mortals”. Thus his errors are particularly egregious.

He can’t even acknowledge the uses of the properties of solution sets in mathematics. Even that acknowledgement should bring one up short to start reconsidering the fact that solutions sets have properties that allow one to converge on particular members of those sets.

Obtaining knowledge of how the physical world finds “solutions” would then be a logical extension of that realization.

Daffyd ap Morgen said:

What is also troubling is how the author of the Metropolis Sampler demonstrates how Dembski and Marks skewed their data:

In order to reach the theoretical lower bound, which in this case is 93.28 nats (or 134.6 bits), the Weasel program must be modified so that fitness is evaluated on an absurdly biased scale with 10^40 possible unsatisfactory fitness values for every possible satisfactory value.

Simply put, MetSam’s analysis of their results shows D&M (as I suspected) re-wrote the “Weasel Program” until it gave them the results they wanted.

Maybe I was a bit unclear. D&M don’t skew the data in this way. D&M make model assumptions that restrict the logical possibilities by 6 * 10^38 bits. The lower bound (134.6 bits) is mentioned to emphasize that it rarely, if ever, is reached.

Creationists never stop trying to find a way to disprove evolution. Attack the biological evidence, get refuted. Attack the paleontological evidence, get refuted. Attack the geological record to say the earth is young, get refuted. Try irreducible complexity, get BLOWN OUT OF THE WATER on the witness stand. So information theory is the last, and furthest removed, desperate area in which these simpering buffoons can pin their hopes.

KP said:

…Try irreducible complexity, get BLOWN OUT OF THE WATER on the witness stand.

Which was why Dembski ran away with his tail between his legs, and let Behe crash and burn?

So information theory is the last, and furthest removed, desperate area in which these simpering buffoons can pin their hopes.

And it looks like Dembski is crashing and burning at this route, too.

No-one seems to have commented on the nature of the journal that Dembski published in: IEEE is an acronym for “Institute of Electrical and Electronics Engineers” . I am not surprised that he chose (was obliged to?) publish in an Engineering Journal, given the number of creation-supporting “scientists” who are actually engineers. The paper would not have gotten past a biologially-trained referee, who would (likely) be familiar with the true nature of the Weasel program.

Search ‘IEEE’ and you will find a site listing that is followed by the blurb: “Technical objectives center on advancing the theory and practice of electrical, electronics, and computer engineering and science.”

How D’s paper meets these goals is not clear to me.

Marks is a fellow of IEEE. However, judging from the length of the list of fellows in just the southwest region that’s not a particularly rare phenomenon.

Henry J said:

Not to mention the bird seed as bait, for a bird that would prefer a nice juicy lizard or snake… ;)

However, a cartoon road runner’s diet is presumably dictated by the exigencies of the plot.

Come to think of it, doesn’t the same apply to D&M’s mathematics?

Henry J said:

Not to mention the bird seed as bait, for a bird that would prefer a nice juicy lizard or snake… ;)

That isn’t a cartoon law of physics. It is cartoon biology. Someone should stack up creationist biology against cartoon biology. My bet is that cartoons would win in terms of biological accuracy even though cartoons aren’t any more constrained by reality than creationism.

Would Barney be considered a cartoon? Is he vegetarian? Land Before Time and Little Foot or the Creation Museum?

djlactin said:

No-one seems to have commented on the nature of the journal that Dembski published in: IEEE is an acronym for “Institute of Electrical and Electronics Engineers” . I am not surprised that he chose (was obliged to?) publish in an Engineering Journal, given the number of creation-supporting “scientists” who are actually engineers. The paper would not have gotten past a biologially-trained referee, who would (likely) be familiar with the true nature of the Weasel program.

Search ‘IEEE’ and you will find a site listing that is followed by the blurb: “Technical objectives center on advancing the theory and practice of electrical, electronics, and computer engineering and science.”

How D’s paper meets these goals is not clear to me.

The paper is about information and search algorithms, and so is appropriate for this IEEE publication (which has a pretty low impact factor compared to other IEEE journals). I doubt the paper would have passed muster in a biology-oriented journal because it’s not about biology, but I’m not a biologist.

If I had been one of the reviewers, my principal comments would have focused on the unremarkable nature of their conclusions and their failure to apply their concept to a significant example of an evolutionary search, or even a significant in silico instantiation of random variation + mutation such as is found in Avida or Tierra. Without such a demonstration, their conclusion that “attempts to characterize evolutionary algorithms as creators of novel information are inappropriate” is not justified. My recommendation would have been “may be publishable after major revision.”

And by “random variation + mutation” I meant “random variation + selection” …

I’ve never been able to get an ID creationist to tell me where the ‘smuggled information’ came from in the case study lineage of The Evolution Origin of Complex Features. See here for a discussion here on PT )note that a platform migration totally wiped out the paragraphs in that post!).

SWT said:

djlactin said:

No-one seems to have commented on the nature of the journal that Dembski published in: IEEE is an acronym for “Institute of Electrical and Electronics Engineers” . I am not surprised that he chose (was obliged to?) publish in an Engineering Journal, given the number of creation-supporting “scientists” who are actually engineers. The paper would not have gotten past a biologially-trained referee, who would (likely) be familiar with the true nature of the Weasel program.

Search ‘IEEE’ and you will find a site listing that is followed by the blurb: “Technical objectives center on advancing the theory and practice of electrical, electronics, and computer engineering and science.”

How D’s paper meets these goals is not clear to me.

The paper is about information and search algorithms, and so is appropriate for this IEEE publication (which has a pretty low impact factor compared to other IEEE journals). I doubt the paper would have passed muster in a biology-oriented journal because it’s not about biology, but I’m not a biologist.

If I had been one of the reviewers, my principal comments would have focused on the unremarkable nature of their conclusions and their failure to apply their concept to a significant example of an evolutionary search, or even a significant in silico instantiation of random variation + mutation such as is found in Avida or Tierra. Without such a demonstration, their conclusion that “attempts to characterize evolutionary algorithms as creators of novel information are inappropriate” is not justified. My recommendation would have been “may be publishable after major revision.”

It also appears that Dembski as well as Marks belong to the IEEE. It is not difficult to join, get a sponsor, apply, pay your membership fee. I think many of the publications are quite abstract and as noted are not screened by those in other fields prior to publication. I did write the editor (I used to be a member as well).

The issue reminds me of an incident ~40 years ago when I was teaching at Georgia State. The library actually subscribed to creationist publications! An article by Harold Slusher “proved” mathematically that the universe was only 10,000 years old. What he’d done was take a basic astronomical calculation, reversed the variables, and lo and behold, when you stuck in the assumed density of the universe, you came out with an age 10,000 years. Unfortunately, you had to keep sticking in density values until the equation finally righted itself and you were swimming in protons. It was a clever trick, ala Dembski. The next article in the magazine had to do with the shell of water around the earth that gave rise to the great flood …

William Dembski:

The implications of intelligent design are radical in the true sense of this much overused word. The question posed by intelligent design is not how we should do science and theology in light of the triumph of Enlightenment rationalism and scientific naturalism. The question is rather how we should do science and theology in light of the impending collapse of Enlightenment rationalism and scientific naturalism. These ideologies are on the way out…because they are bankrupt.

Dembski’s goal is to destroy Western civilization. The Enlightenment and science are the basis of the 21st century West, including the leader, the USA. To set up another unworkable hell on earth theocracy.

He says stuff like this all the time, this is one quoe of countless.

Let’s play psychiatrist over the internet. 1. IIRC Dembski has an autistic child. 2. Autism has a strong genetic component. 3. Dembski is a kook, an a somewhat hostile one at that. Look what he did to Eric Pianka. He certainly hasn’t displayed much of human affiliative behavior and his social skills got him kicked out of Baylor twice. It is quite possible that Dembski has an autism spectrum disorder.

Just an opinion with some reasoning behind it and worth what you paid for it.

raven said:

Let’s play psychiatrist over the internet. 1. IIRC Dembski has an autistic child. 2. Autism has a strong genetic component. 3. Dembski is a kook, an a somewhat hostile one at that. Look what he did to Eric Pianka. He certainly hasn’t displayed much of human affiliative behavior and his social skills got him kicked out of Baylor twice. It is quite possible that Dembski has an autism spectrum disorder.

Just an opinion with some reasoning behind it and worth what you paid for it.

This is truly one of the most offensive things I have ever seen posted, and I’m ashamed that it appeared in this forum. Dembski and his family have my sympathy in dealing with autism, and you are the one who needs counselling.

Dembski and his family have my sympathy in dealing with autism, and you are the one who needs counselling.

Why? I gave the facts as are known to me. And derived a conclusion. It is what it is.

Autism spectrum disorders aren’t necessarily handicaps, just differences. Many of my colleagues, some of whom are also friends show it to varying degrees.

If the moderators don’t like it, feel free to erase it.

you are the one who needs counselling.

Well, that is offensive and insulting to me, but whatever. I suppose playing psychiatrist over the internet can be hazardous if someone plays it back to you.

As a senior member of the IEEE, I am ashamed.

However, I must concur with the previous statement that this is easily the most obscure journal of the IEEE; I was unaware of it until now.

Dembski and Marks (2009a) mention “a simple self-replicating molecule, an autocatalytic set, or a lipid membrane” as examples of biological targets. However, these chemical concepts presuppose model assumptions from physics, chemistry and biochemistry and they could hardly be independent of the chemical processes by which they are thought to arise.

Nice to see that thought put together so succinctly. The way creationists talk and manipulate the numbers has - as far as I’ve ever seen - completely failed to address the elephant in the room: chemistry.

Egregious in particular when they want to make fun of the molecules part of “molecules-to-man”.

raven says:

Dembski’s goal is to destroy Western civilization.

Sadly, Raven, Dembski’s goal is probably nothing more than to stroke Dembski’s ego, and, given his reception by the fraternity of real mathematicians, the simplest way to do that is to play instead to the fawning creobots.

The collateral damage is incidental.

Sadly, this is a pattern we see a lot of these days, ego-stroking demagoguery in place of rational debate, innocent victims be damned. The only saving grace is that Dembski doesn’t have a television show on a news network.

Mike Elzinga said:

What cleared it up for me was when I got into the technique of adiabatic demagnetization in my low temperature superconductivity work while trying to characterize delicate networks of percolating superconductivity taking place throughout a background of nearly constant resistivity.

Were you one of Reddy’s students?

Dan said:

Were you one of Reddy’s students?

When I was doing low temperature and superconductivity work, I studied under Ctirad Uher at Michigan. He, in turn, had studied under Guy White.

I’m one of the lucky ones who, throughout my career, had the good fortune to have gained a fair amount of expertise in several areas of pure and applied physics, before, during, and after my PhD studies. I even got to do some theoretical work. Some the later stuff is, unfortunately, still classified.

I might add, with regard to the adiabatic demagnetization, I found a better way to get the data rather than having to push below 12 mK off the end of my dilution refrigerator.

The superconducting networks were so fragile to probing that I ended up developing a very highly sensitive SQUID magnetometer that enabled me to pull out both magnetic susceptibility and resistivity without having to attach anything to the samples. The rest involved very careful electromagnetic shielding to reduce interference inside the sample chamber of the dilution refrigerator.

The adiabatic demagnetization technique actually interfered with measurements and destroyed what I was trying to see. Measurements had to be made slowly and with only the slightest possible fields. That doesn’t work with one-shot dips to lower temperatures followed by warm-ups that are too fast to collect data.

I probably learned more about the difficulty of making an isolated system in that set of measurements than I ever learned from the literature or from textbooks.

I think I had mentioned earlier (probably on another thread; I can’t seem to find it) that on a recent flying trip I was outlining a “Weasel” algorithm for the HP 48/49/50 series graphing calculator.

Well, I finished a version that allows different probabilities for going from an incorrect character to correct than for going vice-versa. Latching can be done by simply setting to zero the probability for a correct character to change.

I also did a complete analysis of the program (similar to what Wesley Elsberry did) and plotted up these results in MathCad (I could also do that on the calculator as well). The results compare extremely well with everything we have been saying before on those threads discussing the algorithm. The graph is a straight line on a semi-log (log vs. linear) plot for many values of probabilities, population size, number of characters and string length.

String manipulation on the HP calculators is mostly non-existent, but I used lists of integers instead. Manipulation of lists and vectors is quite efficient on these calculators.

Also, the target is generated randomly at the beginning, as is the original parent. The parent can be made to have no matches with the target. The randomly generated target emphasizes the fact that there is no “information” whatsoever in the target. Just the rules of nature are involved; and that’s not cheating.

As I suspected, the algorithm is a battery eater for this calculator. The number of integers, length of list, population size, probabilities, can all be set at the beginning, so the algorithm can run with smaller sizes to conserve batteries.

The program works well, as expected. But the kicker is that it can be done on a graphing calculator while those characters over at UD still can’t figure out how to do it on a computer. Not only another language (this time RPL), but an entirely different platform. :-)

Now you’re showing off, Mike. :)

RBH said:

Now you’re showing off, Mike. :)

Yeah; but I figured the graphing calculator version would rub it in a bit. :-)

Now you’re showing off, Mike. :)

And he was getting graphic, too.

Henry J said:

Now you’re showing off, Mike. :)

And he was getting graphic, too.

And it was all very calculated. :-)

About this Entry

This page contains a single entry by Richard B. Hoppe published on September 9, 2009 10:52 AM.

Bloggingheads’ business plan: Borrow credibility and then blow it. was the previous entry in this blog.

‘Creation:’ A drama about the life of Charles Darwin is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Categories

Archives

Author Archives

Powered by Movable Type 4.361

Site Meter