Thursday, October 26, 2006

ID Optimism

I think that the optimism of IDers is related to how they think about the following diagram:


1. represents the total of all possible proteins that could be coded for in DNA.

2. represents all possible proteins with a selectable function.

3. represents all actually existing functional proteins.

4. represents proteins essential for the simplest living membrane bound organism.

IDers tend to think that the huge size of 1 and the comparatively tiny size of 2,3 and 4 points clearly to some other solution to random testing of possibles until we get the working examples that are needed.

IDers tend to think that even our present knowledge of genomes indicates the need for some other mechanism than random testing to find the necessary tiny needles in this vast haystack. Do the genomes we know look like efficient testing machines to find rare useful proteins?

IDers tend to think that the little orange circle is too big to allow chance to be a realistic explanation for the origin of life. Chance is just the wrong sort of explanation for what we see. It is a little like thinking that random selection of notes in sequence can produce a great symphony.

With objects like the rotary motor function of the flagellum the contraints upon so many proteins at once for selectable function tends to breed scepticism that this object can occur without a designer. This problem is compounded many fold whenever we then begin to think about the origin of life. If it is a real snag that we are struggling to deal with in the flagellum then it is infinitely worse for the origin of life. To reject other explanations outright says more about the rules we impose on our bank of possible explanations than about the real origins of the objects we are examining.

18 Comments:

Anonymous Hrafn said...

Andrew, this is exactly the sort of absurd, ignorant, woolly-thinking, garbled argument that makes me long for Dawkins.

1) The number of possible proteins is infinite (they are chains of amino acids, so, and these chains can be arbitrarily long) - this means that the number of proteins that have every been produced by any living creature that has ever lived (whether or not the protein was functional or a non-functional mutation), is by definition zero!

2) Whether a protein has a selectable function or not depends on its context - a given protein may have a crucial function in a muscle, but no function whatsoever if present in the brain, and vice versa. So to map the domain space of all potentially functional proteins, you would first have to map the domain of all potential physiological contexts.

"IDers tend to think that even our present knowledge of genomes indicates the need for some other mechanism than random testing to find the necessary tiny needles in this vast haystack."

This is because IDers are too STUPID AND IGNORANT to know that there are any other stochastic processes more complex than an independent uniform random variable! This stupid ignorance means that they are completely unaware that stochastic processes often result in outcomes that do not meet any layman's definition of "random." I would therefore suggest that they study a statistics textbook, so as not to sound so stupid and ignorant!

(Yes, I'm being abusive - I have a degree in Statistics, and I don't like seeing the subject being abused!)

What you have developed here, Andrew is a whopping big, garbled, Argument from Personal Incredulity, based on an ignorance of biochemistry and statistics. You don't understand how DNA and RNA constrain the number of potential proteins, or how a complex chain of stochastic events can give a deterministic outcome, so you simply and ignorantly assume that the whole thing is an impossibly improbable fluke.

PS, as I'm a Statistician not a Biochemist, I apologise if I've cocked up the details of the proteins argument. I'm fairly certain the over-arching argument is valid however.

PPS, while the technical definition of 'random' and 'stochastic' is basically identical, 'random' tends to have a layman's meaning of "equal probabilities and independence" (leading to fairly simple statistical systems, with fairly obvious results), so I tend to use 'stochastic' to emphasise the fact that the probabilities can be unequal and dependent (leading to potentially highly complex statistical systems, such as Markov chains, Queuing Theory, etc, with non-obvious results).

12:55 pm  
Anonymous Mike Godfre said...

Hi Andrew,
‘IDers tend to think that the huge size of 1 and the comparatively tiny size of 2,3 and 4 points clearly to some other solution to random testing of possibles until we get the working examples that are needed.’
You got it –although I hesitate to speak for most Iders,
Looking in Simon Conway Morris’s book’ Life Solution ‘ He tends to think that in this case a protein of a 100 amino acids would have ‘..the potential number of combinations in which this protein could be assembled is absolutely gigantic. Specifically it is 20^100, which is equivalent to 10^130.This is an uncomfortably large number…’ This number only increases with the number of amino acids need to construct the protein-the majority of which I believe are greater than 100 amino acids long.

2:29 pm  
Anonymous Hrafn said...

"He tends to think that in this case a protein of a 100 amino acids would have ‘..the potential number of combinations in which this protein could be assembled is absolutely gigantic. Specifically it is 20^100, which is equivalent to 10^130.’"

But only a tiny subset of this "absolutely gigantic" number can in fact be created by known RNA processes - rendering this argument "absolutely" inane. The vast bulk of these conceivable proteins would never have seen existence (unless artificially created in a Biochem lab).

"‘This is an uncomfortably large number…’"

Only if you're a small-minded Creationist. Any scientist worth their salt would know that this "gigantic" number would in reality be reduced down to a far smaller number by various constraints.

3:40 pm  
Anonymous Mike Godfre said...

Hi hrafn,
Seems like you might need a bit of anger management there.
YOU SAID:’ But only a tiny subset of this "absolutely gigantic" number can in fact be created by known RNA processes - rendering this argument "absolutely" inane.'
The numbers quoted by noted evolutionist Simon Conway Morris, were not proteins that have been made, but the possible number of permutations. You think someone has made 10^130 protein variations?
RNA along with everything else has evolved, so saying the current subset of rna's limits the possible proteins that could have been made throughout time is, although true, not applicable to the whole evolutionary story. The point I was making, is that the number of possible protein combinations is 'an uncomfortably large number' (another noted evolutionists said that, not IDers) there is no accounting in chance/dumb luck for the small number of actual samples we see in nature taken from this vast proteins terrain.
You say:’”‘ This is an uncomfortably large number…’"
Only if you're a small-minded Creationist. Any scientist worth their salt would know that this "gigantic" number would in reality be reduced down to a far smaller number by various constraints.'
Well these quotes in my previous post are by cutting edge scientist not IDers or creationist. Your caricatures of IDers are very simplistic and I guess meant to insult, this doesn't take the argument forward. The ‘argumentum ad hominem’ always seems to highlight the opponents lack of subtlety and how impoverished their argument is-I am sure that’s not the case with you.
Many people not just IDers find the current level of complexity and functional complexity being discovered in nature ample evidence to question the power of dumb luck. Francis Crick for one decided to opt for panspermia as get out because he felt chance was not a powerful enough tool for the job.

4:58 pm  
Blogger Andrew Rowell said...

hrafn,
to get an infinite number of possible proteins you would have to have an infinite amount of DNA so my very large number is constrained both by the available DNA and by what can be made in a cell without killing it.

I agree that selectable function depends on context and in multicellular organisms this makes matters worse not better for evolutionists. However in this post I am more interested in simple single celled organisms and the proteins that are selectable in this context.

"IDers are stupid and ignorant."
Some IDers are ignorant and stupid but some are very bright and well read.

I think I sense that I come into your catagory of stupid and ignorant... I do not understand your point about complex stochastic processes solving all the problems. Can you explain?

I have got used to you being abusive and focus on your arguments :-)
I do not mind being called stupid and ignorant as long as I can see that it is true.

My argument is not based upon total ignorance of biochemistry and it was an attempt to answer Matt's question about ID optimism.

It is not just personal incredulity the real issue is whether it is rational incredulity and how we decide whether it is or is not.

5:10 pm  
Anonymous Tempus Fugit said...

"This stupid ignorance means that they are completely unaware that stochastic processes often result in outcomes that do not meet any layman's definition of "random." I would therefore suggest that they study a statistics textbook, so as not to sound so stupid and ignorant."

Earth to hrafn calling: Dembski's Design Inference, whether you buy his argument or not, is precisely about trying to arrive at the relevant complexity in stochastic processes that is definitive of design. IDers do not presume that any stochastic process that is not random is therefore the product of intelligent design. In contrast, it is by no means clear what use Darwinists want to make of stochastic processes in their own thinking. Is there some probability-based proof out there that shows that something could NOT have been the product of design. I rather doubt it. All that Darwinists ever say is that in principle any apparently designed organ, etc. can be understood as the product of some sufficiently complex stochastic process that nevertheless, by some (what?) definition, is not the product of design. Of course, they can't tell us which possible process is the actual one because the historical record of evolution is necessarily gappy, etc.

Not being a professional statistician myself, God knows what subtlety I've missed here. But from this amateur's standpoint, the evolutionists are winning merely because the players presume their terms of engagement. If you shift the burden of proof on to the evolutionists, their position looks at least as underargued as the IDers'.

10:13 pm  
Anonymous Hrafn said...

"Seems like you might need a bit of anger management there."

Mike: willful abuse of statistics gets me angry. Get over it!

The argument I was making was that the gigantic number of proteins is statistically completely irrelevant. The vast majority of these proteins will never occur in nature. Their probability is therefore zero. This means that the probability of any one of the proteins that do (or even might) occur is considerably greater than one divided by gigantic number. Thinking that this probability is "one divided by gigantic number" is a fallacy that I see repeated countless times in creationist literature. It bugs me. This is why I made the "small-minded Creationists" crack.

I will cease "caricaturing" IDers, when IDers cease caricaturing statistics.

Fair enough?

2:36 am  
Anonymous Hrafn said...

IDers are stupid and ignorant."

Andrew, that is not what I said. This is a paraphrase of a fragment of a sentence of mine, misrepresented as a direct quote of a complete sentence. And you wonder why I make a big issue about quote-mining.

What I in fact said was:
"his is because IDers are too STUPID AND IGNORANT to know that there are any other stochastic processes more complex than an independent uniform random variable!"

This is because I have yet to come across an IDer who demonstrates a nuanced view of statistics, rather than a caricature of it (this includes Dembski, who should know better).

"think I sense that I come into your catagory of stupid and ignorant... I do not understand your point about complex stochastic processes solving all the problems. Can you explain?"

Okay.

Simplest (and thus most caricatured) level: independence and equal probability. Think of an ordinary six-sided die. Assuming that it is fair (i.e. symmetric), the probability of it yielding any number is equal, and the probability of what the next number is unaffected by (is independent of) what the last number thrown was.

A slightly more complex example might be special dice that is used in wargaming, that replace the 1 & 6 with 3 & 4, which would have unequal probabilities (but still retain independence). A real-world example of this is that the probability of a randomly picked baby growing to be 7" tall is considerably less than that of growing to be 5'10".

Dependence is far more common than independence. Your height and weight will depend on your gender (itself a random variable) and on height/weight of your parents. Similarly, your hair and eye colour will depend on that of your parents. In the extreme, two blue-eyed blondes will never have a brown-eyed brunette child (as the genes are recessive). Thus dependence, over times, makes some outcomes more likely, and other less so (or even inevitable or impossible).

There are too many potential additional complexities to mention - the sum total of them amounts to several years study, but some idea might be gleaned from reading a first-year university statistics textbook.

Taken in the extreme, stochastic processes can result in complex, seemingly-planned outcomes such as snowflake symmetry, insect compound eyes and fractal art.

Stochastic patterns frequently lead to quite deterministic outcomes - which is why bookmakers and casinos can make a profit.

"It is not just personal incredulity the real issue is whether it is rational incredulity and how we decide whether it is or is not."

I cannot speak with any particular authority on the biochemical side of it, but from the statistical side, your incredulity is definitely not rational.

3:26 am  
Anonymous Hrafn said...

"Earth to hrafn calling: Dembski's Design Inference, whether you buy his argument or not, is precisely about trying to arrive at the relevant complexity in stochastic processes that is definitive of design."

Earth to Tempus Fugit: I have yet to see an argument of Dembski's that even attempts to model the constraints and complexities that Biochemistry imposes on possible outcomes in his statistical calculations.

"IDers do not presume that any stochastic process that is not random is therefore the product of intelligent design."

Stochastic processes are always random, in the strict technical sense of "random." In the colloquial sense of "random" your statement is trivially obvious, as the list of obviously-not-designed stochastic processes that exhibit neither independence nor equal probability is very long.

In contrast, it is by no means clear what use Darwinists want to make of stochastic processes in their own thinking.

Generally, they don't. They know that the level of assumptions needed to make such calculations (in terms of input probabilities, distributional assumptions, etc) are so heroic as to render the results essentially meaningless. Garbage in, garbage out.

One quote on this I have come across is:
"We cannot calculate the probability that an eye came about. We don't have the information to make the calculation"
- Martin Nowak, a Harvard professor of mathematics and evolutionary biology.
http://www.time.com/time/magazine/article/0,9171,1090909-6,00.html

3:51 am  
Anonymous Hrafn said...

"I agree that selectable function depends on context and in multicellular organisms this makes matters worse not better for evolutionists."

I would disagree with you on this point. The wide variety of contexts yield a wider range of proteins that would be functional in some context, and so a wider pool of potential antecedent-proteins that could be changed through a mutation (or other change) to a protein that might have a selectable function in a related context - thus far increasing the chances of evolution by cooption.

3:58 am  
Blogger Andrew Rowell said...

Hrafn,

"The vast majority of these proteins will never occur in nature."

I am not sure I understand your argument here hrafn. Can you explain?

2:40 pm  
Blogger Andrew Rowell said...

Hrafn,
You said:
"But only a tiny subset of this "absolutely gigantic" number can in fact be created by known RNA processes - rendering this argument "absolutely" inane. The vast bulk of these conceivable proteins would never have seen existence (unless artificially created in a Biochem lab)."

I do not understand your argument here. Can you explain? Can you tell me some more about how you get to your "tiny subset"?

2:44 pm  
Blogger Andrew Rowell said...

In multicellular organisms you have the problem of control and a cell knowing where it is in the body in order to trigger the correct set of proteins. A new protein may work well in one context but not in another and this will tend to mask selective advantage.

3:18 pm  
Anonymous Hrafn said...

"I am not sure I understand your argument here hrafn. Can you explain?"

A finite (very large, but a tiny fraction of 10^130) number of different species that have ever lived, a large (but small in terms of orders of ten) number of proteins per species (with considerable overlap between species).

To pluck numbers out of thin air, let us assume a thousand new species per year, over a billion years - that's 10^9 species. Let us assume a million different proteins per species, and ignore the overlap. This is still only 10^15 proteins. Let's assume that I underestimated somewhere by an order of magnitude of 5, so bump it up to 10^20. This is still 1/(10^110) of the original number of proteins.

It doesn't matter if a protein is functional or not, if no living creature can be found that produces it (in fact, 'functional' becomes meaningless, as there is no context for it to function in). It may as well not exist, from a statistical point of view. In fact if neither creature nor lab process produces it, then it literally does not exist, except in a Biochemist's imagination. It then has no more effect on the real world than the unicorn or any other imaginary beast.

3:36 pm  
Anonymous Hrafn said...

Eep - obvious arithmetic error in the above. It should read:

To pluck numbers out of thin air, let us assume a thousand new species per year, over a billion years - that's 10^12 species. Let us assume a million different proteins per species, and ignore the overlap. This is still only 10^18 proteins. Let's assume that I underestimated somewhere by an order of magnitude of 5, so bump it up to 10^23. This is still 1/(10^107) of the original number of proteins.

(But then again, I could be out by a trillion trillion and the argument would still hold.)

3:43 pm  
Blogger Andrew Rowell said...

Hrafn,

Your argument is looking at all the proteins that have ever existed.

It is not a question of which proteins have ever existed in the universe but a question of how the proteins that have existed came to exist. How did that small subset of actual proteins come to be the ones inside living cells when there are so many possible ones that have to be rejected?

6:01 pm  
Anonymous Hrafn said...

"How did that small subset of actual proteins come to be the ones inside living cells when there are so many possible ones that have to be rejected?"

In the normal manner of natural selection. Proteins that are lethal will be rejected immediately. Proteins that are inefficient will result in the organism starving or being eaten, so will also tend to be rejected. Trial and error, with the errors getting eaten.

After all "selection" and "rejection" are just two faces of the same coin.

6:55 pm  
Blogger Andrew Rowell said...

Hrafn,

You have to go back to the initial building of proteins to make a system that can reproduce. How do you get all these proteins in the first place?

12:27 pm  

Post a Comment

<< Home