Thursday, February 28, 2013

Dembski Repeats the Same Tired Nonsense

Here's a video of Bill Dembski repeating the same old nonsense about intelligent design. Want to know why intelligent design is dead? Because even its leading proponents have nothing new to say.

- "specified complexity" and "complex specified information": incoherent, useless concepts that have been debunked long ago by me and many others

- the movie Expelled shows how intelligent design advocates get discriminated against (see here for the real story)

- same grotesque slurs against legitimate researchers, implying they've done something underhanded by "smuggling in" information

- same martrydom about how his career has been damaged by evil Darwinists.

- same mystical view of "information", without understanding how it can be created by any random process.

- same claim about prime numbers implying intelligence, even though there is evidence that natural processes that can generate them (e.g., cicada periodicity to avoid predation).

And he still doesn't understand that evolution doesn't have a goal and that evolutionary fitness landscapes come from the environment.

31 comments:

KeithB said...

Did you see this:
http://sandwalk.blogspot.com/2013/02/darwins-dead-idea-and-man-who-helped.html

The blurb is priceless.

mregnor said...

"same mystical view of "information", without understanding how it can be created by any random process."

So how about this: create a blog post, consisting of 10 or more grammatically correct English sentences, with a coherent meaning, using randomly generated letters and punctuation and no intelligently designed endpoints (no grammatical or semantic teleology).

You claim that non-intelligent non-teleological processes generated you. Show me that you can generate a meaningful blog post using such a process.

Have at it, Mr. Computer Scientist.

Jeffrey Shallit said...

That's what we love about Michael Egnor. He knows nothing about computer science or information theory, but still wants to lecture experts about it. It's that special brand of ignorance and arrogance that makes him so delightful, and a perfect spokesman for the Discovery Institute.

Egnor: go read my paper in Synthese and come back when you've understood it. If you do, you will understand why your challenge has nothing to do with the subject.

Hint: can you define "intelligently designed endpoints" in a way that computer scientists will agree it has a meaning? Which computer programs have "intelligently designed endpoints" and which don't? What test can we perform to decide if a program has them?

Can you define "grammatical and semantic teleology" in a way that computer scientists will agree it has a meaning? What tests can we perform to decide if a program has "grammatical and semantic teleology"?

Can you explain why "10 or more grammatically correct English sentences" could well contain very little information?

Anonymous said...

Isn't it interesting that the analogy proposed by mregnor goes back at least as far as Cicero? And was used in the 18th century to argue for preformation? (By Voltaire, for example.)

I'd note that the 18th century argument was a better use of the analogy because it was an argument for something, a decided contrast with the empty "Intelligent Design".

TomS

mregnor said...

@Shallit:

[Can you explain why "10 or more grammatically correct English sentences" could well contain very little information?]

Your blog posts are good examples.

All of your tap-dancing is quite funny. I've merely asked you to provide a computer model of your teleology-free theory of evolution.

You reply:

"Look-- a squirrel!"

Jeffrey Shallit said...

Egnor is too intellectually lazy to read my paper. Who is surprised?

I asked Egnor to define his terms. Which computer programs are "teleology-free" and which are not? How can we distinguish them?

I'll make it easy for Egnor. Consider the set of all 1-tape determinstic Turing machines with input alphabet {0,1} and tape alphabet {0, 1, B}. Out of (say) the set of all such 5-state machines, how many are "teleology-free" and how many are not?

Anonymous said...

mregnor:I've merely asked you to provide a computer model of your teleology-free theory of evolution.

How about providing a ... I'm not going to be nearly as demanding ... a partial description of an alternative to evolution?

"Should the believers in special creations consider it unfair thus to call upon them to describe how special creations take place, I reply, that this is far less than they demand from the supporters of the development hypothesis. They are merely asked to point out a conceivable mode; on the other hand, they ask, not simply for a conceivable mode, but for the actual mode. They do not say — Show us how this may take place; but they say — Show us how this does take place. So far from its being unreasonable to ask so much of them, it would be reasonable to ask not only for a possible mode of special creation, but for an ascertained mode; seeing that this is no greater a demand than they make upon their opponents."

The Development Hypothesis (1852) by Herbert Spencer

TomS

mregnor said...

@Shallit:

Odd. I didn't ask you a theoretical question. I issued a practical challenge. Generate a syntactically correct and semantically meaningful blog post, without teleology.

After all, you assert that life evolved without teleology. Surely you (who evolved without teleology) can generate a short blog post without teleology.

"Oh", you reply, "I don't know what teleology means". "So", I reply, "Why do you deny it in evolution?"

Let me help you with your computer problem. A non-teleological blog post would be generated by a random letter/punctuation input to a non-deterministic Turing machine in which the state and symbol do not uniquely determine the transition function, which is itself governed by a random letter/punctuation generator.

If you really want to make sure it isn't intelligently designed, you might just use one of your previous blog posts for the transition function.

Better get to work! Time's a wastin'.

Jeffrey Shallit said...

I didn't ask you a theoretical question.

You didn't ask anything meaningful at all.

Why do you think asking for a criterion by which it can be judged whether a program is teleological or not is "theoretical"? It's entirely practical.

in which the state and symbol do not uniquely determine the transition function

I have no idea what this means. Do you mean that
delta(q,a) = delta(q',a') for some pair of states q, q' and input symbols a, a'?

We know that any non-deterministic TM can be simulated by a deterministic one, so why is it important that your TM be non-deterministic?

transition function, which is itself governed by a random letter/punctuation generator

Word salad. Why are you so incapable of answering a simple question?

How would a transition function be "governed" by a random letter/punctuation generator? You do realize that if you are talking about randomized Turing machines, then these are not TM's in the ordinary sense, right? Or are you referring to a probability distribution over Turing machines? In that case, are you claiming that "non-teleological" is a property of a TM or a distribution of TM's?

I'm teaching a course right now on theory of computation. Let me make it even simpler for you. Is the first TM on this page teleological or not?

mregnor said...

@Shallit:

It's teleological.

Only randomly generated endpoints are non-teleological.

Still waitin' for that Darwinian-generated blog post...

Jeffrey Shallit said...

Only randomly generated endpoints are non-teleological.


What are endpoints? I've taught theoretical computer science for more than 30 years and we somehow don't talk about "endpoints". Please, educate me.

OK, if that one is teleological, please give me the transition diagram for one that is not.

Why didn't you answer my other questions? Is "teleological" the property of a TM? Or a distribution of TM's?

mregnor said...

I'm still waiting for that undesigned/non-teleological blog post. You seem so confused about teleology-- so sad for a man who bases his metaphysics on the denial of teleology.

Why don't you generate your un-designed blog post, takin' your best 'computer scientist' guess at what's teleological and what's not, and then publish your algorithm for public discussion.

Show us what you can generate without design. I'd love to see what you come up with.

Jeffrey Shallit said...

I'm still waiting for that undesigned/non-teleological blog post.

Sure, I'm happy to comply, once I understand what it is you are asking for. Now answer my questions:

1. Is "teleological" a property of TM's or a property of a distribution of TM's?

2. What are "endpoints"?

3. What is a single example of a "non-teleological" TM?

4. Why do you insist on a nondeterministic TM, when it is known that a deterministic TM can simulate it?

5. Was my guess about "delta(q,a) = delta(q',a') for some pair of states q, q' and input symbols a, a'" your intended meaning? What is the point of this condition?

6. How would a transition function be "governed" by a random letter/punctuation generator? Give a more precise definition, like the ones found in any introductory theory of computing textbook.

It is hard to argue with such a disordered mind like Egnor's, because he thinks he is saying something profound when all he has is word salad.

mregnor said...

[It is hard to argue with such a disordered mind like Egnor's]

But there's no argument. I challenged you to do something. No argument involved.

Generate a proper 10-sentence blog post without using intelligent design or teleology.

Show your work when you're done.

If you can't do it, or you can't even define "intelligent design" or "teleology", then I'll accept that you have nothing meaningful to say on the subject.

mregnor said...

I kinda' know that you're gonna keep stallin'. So let me rephrase my challenge, so you have a little less room to wiggle:

Generate a proper 10 sentence blog post using a computer process as close to a Darwinian process as you can make it-- i.e. model RM + NS to generate the blog post.

Richard Dawkins famously tried to do it with his "Me thinks it is like a weasel" program, but he made a little faux-pas and used the phrase as a target. It was a beautiful computer simulation of teleological evolution. Oops.

See if you can avoid that. Generate a blog post using a computer simulation of a Darwinian process-- do the best you can.

I know it's hard, but stop running away. Don't chicken out on this, like you chickened out on abortion.

Jeffrey Shallit said...

I challenged you to do something.

But you steadfastly refuse to answer even the simplest questions about what it is you want me to do.

Shorter Egnor: "I know nothing about computer science and have no training in the subject. But my religion demands that something be impossible to do with a computer program. I won't say what it is, exactly, and I won't answer questions about it. Instead, I'll spend two minutes browsing some Wikipedia page so I can throw in some nonsense about nondeterminism and transition functions. But if you don't create a program to do it, then you lose."

Did I get it right?

Jeffrey Shallit said...

he made a little faux-pas and used the phrase as a target.

Sure. Define "target", and tell me how I can tell whether a program has one. After all, I wouldn't want to spend all this time writing a program, only to learn that I misunderstood what you meant by "target", would I?

Let's go back to the TM here. Does it have a target or not? Show your work.

mregnor said...

Shorter Shallit: "I know a lot about computer science and have a lot of training in the subject. But my irreligion demands that I use my training and knowledge to obfuscate rather than elucidate, because a guy who sees through my bluff is asking me to do things that would demonstrate that my irreligious beliefs and my creation myth are idiotic.
So I'll throw a lot of chaff and hope and pray (I don't actually pray) that this guy goes away. I'll do anything I can to avoid actually engaging the issue he challenges me with"

BrianG said...

Here's one Egnor - http://crd-legacy.lbl.gov/~dhbailey/dhbpapers/dhb-english-text.pdf. The function has no pre-specified target.

As for strawmanning Dawkins, his example was only intended for showing how cumulative selection differed from pure chance.

Paul C. Anagnostopoulos said...

I'm not sure what an "endpoint" is, either, but Egnor appears to be asking you to employ a random mutation and natural selection process to generate a 10-sentence blog entry that utilizes grammatically correct and semantically meaningful English.

The endpoint is apparently a previously established natural language, so the teleology has already had its say. If the program is allowed to use the rules of English as the selection pressures, then I daresay you would not have too much trouble producing the program he requests.

If Egnor wants a teleology-free experiment, then he cannot dictate the results a priori.

Sometimes IDers seem to have difficulty understanding that a simulation of evolution must involve interesting selection pressures if it is to end up with interesting results, and that those selection pressures are not necessarily "rigging the deck." That is, unless they also want to call natural selection pressures "rigging the deck."

~~ Paul

Jeffrey Shallit said...

Egnor: If asking you for clarifications is not engaging, what is? Why not just admit you have no answers and are just bluffing? Of course, that would take honesty.

I count 7 different questions you have no answers to. Pathetic.

Diogenes said...

@Smegnor:

Generate a proper 10 sentence blog post using a computer process as close to a Darwinian process as you can make it

Explain to us, Smegnor, why ANY inability to create human language sentences is in any way relevant to observed processes of gains in BIOLOGICAL complexity.

Suppose Shallit cannot produce English sentences by a Darwinian algorithm. So what?

There are no English sentences in any biological lifeform. There are no English sentences (or sentences in any other human language) in the human genome or in the genome of any other species.

There are no English sentences (or sentences in any other human language) in RNA, protein molecules, lipids, ribosomes, mitochondria, etc. etc.

There is nothing like human grammar or human written languages in DNA, RNA, protein molecules, lipids, ribosomes, mitochondria, etc. etc.

Do you think the human genome was written in English, like your Bible?

Why do you believe English language sentences are in any way analogous to, or present in, biological structures?

Give your answer in the form of an EQUATION. That is, write down an equation which takes as input a structure, and returns a property which:

1. Returns nonzero value when applied to an English sentence

2. Returns nonzero value when applied to a biological structure, e.g. a genetic sequence

and

3. Cannot be increased by already-observed evolutionary processes

Write this equation down NOW, an equation with properties 1, 2, and 3 above.

Write this equation down NOW. I mean in your very next comment on this thread.

Jeffrey Shallit said...

asking me to do things that would demonstrate that my irreligious beliefs and my creation myth are idiotic

You can't define your vague concept rigorously; you can't answer the most basic questions about your concept; you insult and mock a genuine attempt to understand what you are saying.

That's the difference between mathematics & science on the one hand and the pseudoscience of intelligent design creationism on the other. When Turing formalized the previously-vague notions of "algorithm", "decision problem", what it means for an algorithm to "solve" a problem, and so forth, he did so in a tightly-argued paper published in a journal. Other mathematicians and scientists quickly realized his definitions were useful & fruitful and the intellectual revolution started by Turing began.

It is not up to me to try to formalize your vague, incoherent notions. That's the typical strategy of intelligent design creationists -- they want credit for things done by other people, without doing the hard work involved.

I would find it amazing and fascinating if you could define "teleological TM" and "non-teleological TM" in a way that is rigorous, and then prove that teleological TM's can do things that non-teleological TM's can't. It would be a breakthrough in theoretical computer science and artificial intelligence equal to Turing's. You would probably get the Turing award.

It's far easier, though, for you to bluff and insult and claim victory. That's what you're doing.

By the way, I am travelling for the next 48 hours so it is likely that comments will be held up in moderation until I get internet access.

Jeffrey Shallit said...

That is, unless they also want to call natural selection pressures "rigging the deck."

Well, of course. The whole thing amounts to a massive misunderstanding of algorithms on the one hand and mutation & natural selection thought of an algorithm on the other. It's like their eyes are incapable of viewing anything computational except through the distorting lens of "teleology" -- a term they cannot even define coherently.

Their thinking is still in the middle ages. Meanwhile the intellectual revolutions of Turing and evolutionary computation and natural computing have happened, and they are still stuck in 1200.

Curt Cameron said...

Jeffrey, did you ever comment on Frank Tipler? Recently the DI has had him on a series of their podcast ID The Future. Apparently he's a mathematical physicist who did post-doc work at UT under John Wheeler, so has apparently very good credentials.

I listened to the podcasts, and he took an unusual tack: he defended Einstein's view that the laws of physics cannot be fundamentally by chance, that "God does not play dice." I've always considered this to be a good demonstration that even great scientists can be plainly wrong, because we know that subatomic events really are random.

But Tipler described how the math of QM demands there to be multiple universes, and what looks at first like random events is really just us being in one of many universes where all possibilities are played out. I could just imagine the DI flaks and regular audience being pretty uncomfortable with this talk.

So how did he tie this back to biological evolution, which is what the podcast is about? Well, you see, since there are no random events and everything is deterministic, that rules out Darwinian evolution. How? Because the Theory of Evolution says that change happens by random mutation combined with natural selection, and Tipler says that there is no randomness in the universe, so there.

It's like he doesn't understand that "random mutation" doesn't actually mean non-determinism, it just means that it's the product of processes that are so outwardly complex that biologists can treat them as random.

I've read more about Tipler's ideas, and he's quite the crackpot.

Telly O' Logical said...

"And he (Dembski) still doesn't understand that evolution doesn't have a goal..."

Huh? Isn't that precisely what he tells everyone?

SLC said...

Re Curt Cameron

On has to be careful here not to conflate the concept of parallel universes introduced into quantum mechanics and multiple universes as introduced by the strings hypothesis.

Parallel universes are introduced into quantum mechanics to attempt to explain conundrums like the 2 slit problem or quantum entanglement. The deal is that all the parallel universes have the same set of values of fundamental constants (e.g. Planck's constant)

The multiple universe concept is introduced to attempt to explain why the fundamental constants have the values they have, which Tipler claims imply a designer (that's why the IDiots like him, it's called fine tuning). Each of the universes may have the same values for the fundamental constants or different values. The point is that we just happen to live in the universe that appears fine tuned for life. Like the appearance of design in life forms, it's an illusion. Of course, the set of parallel universes is, presumably, a subset of the set of all universes.

BrianG said...

Did Egnor crawl back into his cave or where did he go?

Telly O'Logical said...

"And he (Dembski) still doesn't understand that evolution doesn't have a goal..."

I can't help but conclude that you're misrepresenting him.

Dembski writes, on his website (in the Orzack-Sober_review): "According to Darwinism, biological evolution proceeds without discernible plan or purpose."

It's as clear as day what he thinks, and what he rejects.

Jeffrey Shallit said...

I can't help but conclude that you're misrepresenting him.

I guess you didn't watch the video then. His probabilistic analysis assumes a specific outcome (the one we see) and works backwards from it.

It's as clear as day what he thinks, and what he rejects.

The statement is correct, but not in the way you think.

BrianG said...

Hi Shallit,
Has Dembski's definiton of 'specified' information changed since your critique in your 2003 paper?