|
Author
|
Topic: Answering: The advantages of theft over toil
|
Jerry D. Bauer
Member
Member # 756
|
posted 16. May 2004 20:11
NOTE TO ISCID BRAINIACS: I have been debating Dr. Elsberry in another forum (If one can call it debating) and both him and Dr. Wilkins have been notified that their paper has been refuted in this forum and they have been invited to come over and defend it.
Answering: The advantages of theft over toil: the design inference and arguing from ignorance
By Jerry Don Bauer
INTRODUCTION: In 2001, Wesley R. Elsberry of Texas A&M University at Galveston and John S. Wilkins, University of Melbourne, Australia published a paper entitled The advantages of theft over toil: the design inference and arguing from ignorance. This paper was published in Biology and Philosophy 16 (November):711-724. (1)
Abstract
Philosopher William Dembski has proposed an "explanatory filter" for distinguishing between events due to chance, lawful regularity or design. Elsberry and Wilkins proposed that if Dembski's filter were adopted as a scientific heuristic, some classical developments in science would not be rational. They further posit in the abstract that if background information changes even slightly, the filter's conclusion will vary wildly and that Dembski fails to overcome Hume's objections to arguments from design, neither posit substantiated in the paper. I will show this paper as not based on science or logic and blatantly erroneous, as peer reviewed herein, in its basal tenets.
KEYWORDS, Dembski, Elsberry, Wilkins, Intelligent Design, Complex Specified Information, Explanatory Filter, Darwin, Abiogenesis.
I’ll let the readers dissect the ‘Fingers’ analogy and get to the gist of the paper. The authors begin this paper with a false postulation: “Dembski's filter is a reworking of Paley's design inference (DI) in the forensic manner of identifying the ‘guilty parties.’”
This, of course, is not correct. Although Paley introduced his famous watch on the heath argument, the similarities stop here. Paley was a philosopher/theologian and intertwined his brand of design with much philosophy and a healthy dose of theology. Dembski communicates design as an antithesis to Paley‘s version.
Dembski is a philosopher, as well. However, he mathematically quantifies the design argument with this math being firmly rooted in science. (2) Dembski proposes an upper limit barrier in which events or reactions cannot occur spontaneously as when the odds reach 10^-150 (or 1:10^150). This value is also equivalent to 500 bits of information as defined by mathematician and the father of modern information theory Claude Shannon and the bits conclusion is sometimes used to describe CSI as well.
This figure is not pulled out of thin air. Dembski arrives at this figure by the following calculation: 10^80 particles in the universe x 10^45 state transitions (Planck time) x 10^25 seconds in the Universe = 10^150.
One can readily see this has nothing at all to do with Paley’s notion of design and the assertion that the two are related holds little merit.
The authors do describe the explanatory filter well and offer the following chart to illustrate the law, chance, design tenets of the explanatory filter: (3)
Then the authors introduce several analogies wrapped in other analogies that have very little to do with anything Dembski has postulated.
One analogy reads as follows:
“This naturalist - call him Charles - is on a voyage of discovery. He has read his Paley; indeed, he might almost have written out Paley's Evidences with perfect correctness by memory. Although he has not heard of Dembski's filter, he knows the logic: whatever cannot be accounted for by natural law or chance must be the result of design. Young Charles encounters some pattern of the distribution and form of a class of organisms - let us suppose they are tortoises - on an isolated archipelago and the nearest large continent. Each island has a unique tortoise most similar to the autochthon of the neighboring island and the island closest to the continent is most similar to that species. On the basis of the biological theories then current, he knows that there is no known process that can account for this pattern. It is so marked that one can draw a tree diagram from the continental form to the islands, and it will match a diagram showing the similarity of each form to the others. What should Charles rationally infer from this? Let us assume for comparative purposes that Charles is in possession of the filter; he will therefore reason like this:
“E: Species are distributed such that morphological distance closely matches geographic distance.
“HP? No, there is no regularity that makes this distribution highly probable.
“IP? No, the likelihood of such a distribution is extremely low.
“SP? Yes, it is a very small probability (made even smaller as more variables are taken into account).
“sp/SP? Yes, the problem is (more or less) specified.
“Conclusion: The tortoises have the biogeographic distribution and formal distribution they do by design.
“By Dembski's framework, Rational Charles should have ascribed the tortoises' situation to intelligent agency…..”
The first question asked by the EF is if the system under study can possibly be explained by natural law. Of course, tortoises that vary in sub-speciation with a different sub-species found on the main-land and different ones on neighboring islands can be explained by natural law. Its called evolution via random mutation and natural selection and this law was suspected before young Charles added his input to it. The EF, if employed by a knowledgeable Idist, would have detected this.
Thus, if the state of a system can be explained by natural law, the EF stops there and design cannot be detected with this tool. Yet, the very confused young Charles believes he detects design. Obviously, the young Charles is Charles Darwin and reading Origin of Species shows clearly he did not imply design, nor has anything ever posited by an Idist I’m aware of suggested he should have suspected design. This is little more than gratuitous musing by the authors.
The authors then bring up the old circuitous argument that the EF cannot be employed in the case of abiogenesis, as there is no evidence that anything “popped” into existence. What if the original organism from which all known organisms evolved, itself evolved into the first protista?
Naturally, they never offer any reasonable scenario, math, science or logic on how this could occur and the notion defies common sense when one examines it. Organisms must have genes to evolve. How could an organism breathe, eat, reproduce or do anything at all to interact with its environment before it exists as an organism?
What unknown force could suddenly cause thousands of proteins to come together, a cell wall to form--Pop into existence a cell Membrane: a flexible, semi-permeable barrier with lipid center that controls diffusion in and out of the cell. Magically bring in cytoplasm: the fluid-filled space inside the cell containing hundreds of different enzymes--and let’s not forget our ribosomes, DNA, RNA, and a “pool” of millions of small molecules and ions. Then more ribosomes: particles which are made of protein and RNA, for the sites of protein assembly.
How did proteins assemble BEFORE these protein assembly sites existed?
Of course, the authors believe all of this separately evolved, then came together. How? Who can even propose a scenario in which this could possibly occur?
The authors do not try to explain this and point out that Dawkins proposes this as well, but offers no math to support it.
Next, the authors attempt to assert gods into the math of the EF:
“What the filter lacks that real-world design inferences already have is a (sic) "Don't know" decision. If we can say of a problem that it is currently intractable or there is insufficient information to give a regularity or chance explanation now, then the Filter tells us we must ascribe it to design if it is specifiable. But it can be specifiable without the knowledge required to rule out regularity or chance explanations. This is clearly a god of the gaps stance, and it can have only one purpose: to block further investigation into these problems.”
This is tautology based on logical fallacy, if I can get the argument far enough along to apply logic to it. The filter certainly does have an ‘I don’t know’ built into it. Any conclusion other than design is an ‘I may or may not know what caused the existence, but we cannot show it was design via the filter.”
And how in the world did the god-of-the-gaps get in there? This is little more than the argument from ignorance used in the title of their paper: ‘He can’t show this was designed, therefore he surmises that God did it.’ What?
Next the authors demonstrate they have little understanding of how the EF works:
“As Dembski's probabilities are Bayesian assignments made on the basis of a set of prior knowledge and default hypotheses, this seems to be a perfectly reasonable move. However, it has one glaring problem - it blocks any inferences of design, and that is too much. There are well attested cases of design in the world: we humans do things by design all the time. So an explanatory filter had better not exclude design altogether. How can it be included here? When is a design inference legitimate?”
The EF never excludes design, it only shows design. Lack of proof for a positive does not necessitate a negative. This is argumentum ad ignorantiam all over again. Its purpose is to show beyond a shadow of a doubt that something is designed. Yet, there can be things designed of such simplicity that the EF doesn’t detect it.
I’m going to take a sheet of paper and graph it off into 100 squares which we will consider as microstates. Then I’m going to arrange 10 pebbles in a circle on that graph paper. This is, quite obviously, a designed system. So does this system meet the strict standards for complex specified information? Let’s do the math and see.
Each pebble has a 1:100, or 1:10^2 chance of existing in any given microstate. For 10 pebbles to exist together in those microstates, the odds become 1:(10^2)^10 or 1:10^20. This figure is much less than the 10^150 barrier of CSI, and so, I cannot show it as designed and therefore the EF has no place in this scenario to detect design.
Here, I have to say, I don’t know, from the perspective of the EF because the filter is not designed to be used in these types situations. The EF is designed to be used in extremely low probability scenarios such as tissues and the like. The authors seem not to understand this at all.
Next, we will examine the logic of skirnobs as proposed by Elsberry and Wilkins:
“The problem with a simple conclusion that something is designed, is its lack of informativeness. If you tell me that skirnobs are designed but nothing else about them, then how much do I actually know about skirnobs? Of a single skirnob, what can I say? Unless I already know a fair bit about the aims and intentions of skirnob designers, nothing is added to my knowledge of skirnobs by saying that it is designed. I do not know if a skirnob is a good skirnob, fulfilling the design criteria for skirnobs, or not. I do not know how typical that skirnob is of skirnobs in general, or what any of the properties of skirnobs are. I may as well say that skirnobs are "gzorply muffnordled" for all it tells me. But if I know the nature of the designer, or of the class of things the designer is a member of, then I know something about skirnobs, and I can make some inductive generalizations to the properties of other skirnobs.”
This logic tells us several things. Toasters cannot make toast unless we know exactly who the design engineer of that toaster was.
Throw out the vacuum cleaner, the hair dryer, the train you take to work, the airplane you take every Christmas to see Grandma, the shoes you are presently wearing and the big screen you are watching because if one does not have knowledge of the design engineer, the designed item is useless.
This is also another non-sequitur which we are seeing occurring with frightening regularity throughout this paper. The concept of design detection does not necessarily extrapolate to, therefore the designer is named Pete. It’s not necessary to know who the designer is in order to detect design any more than it is necessary to know who the surgeon was in order to deduce that surgery was performed on an organism. The irrelevancy is starkly obvious.
The rest of the paper needs little addressing.
(1) http://www.talkdesign.org/faqs/theftovertoil/theftovertoil.html#pgfId-197275
(2) William Dembski, No Free Lunch: Why Specified Complexity Cannot be Purchased without Intelligence, Rowman & Littlefield, 2002.
(3) http://www.arn.org/docs/dembski/wd_explfilter.htm
[ 20. May 2004, 17:33: Message edited by: Jerry D. Bauer ]
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 16. May 2004 21:09
Jerry, thank you for providing your ideas about why 'Theft over Toil' should be rejected as a valid argument against the design inference. I will be in more detail be exploring your statements but in the mean time I ran across a statement which I believe to be wrong.
Jerry objects to the authors stating that:“Dembski's filter is a reworking of Paley's design inference (DI) in the forensic manner of identifying the ‘guilty parties.’”
However Dembski himself seems to agree with the authors
quote:
Though intuitively appealing, Paley's argument had until recently fallen into disuse. This is now changing. In the last five years design has witnessed an explosive resurgence. Scientists are beginning to realize that design can be rigorously formulated as a scientific theory. What has kept design outside the scientific mainstream these last hundred and thirty years is the absence of precise methods for distinguishing intelligently caused objects from unintelligently caused ones. For design to be a fruitful scientific concept, scientists have to be sure they can reliably determine whether something is designed.
Link
From Answers In Genesis we find
quote:
Complexity of a different kind—‘specified complexity’—is the cornerstone of the intelligent-design arguments of William A. Dembski of Baylor University in his books The Design Inference and No Free Lunch. Essentially, his argument is that living things are complex in a way that undirected, random processes could never produce. The only logical conclusion, Dembski asserts, in an echo of Paley 200 years ago, is that some superhuman intelligence created and shaped life.
Perhaps Bill can comment on Jerry's conclusion?
Btw Jerry hits the head on the nail when he states that "The EF, if employed by a knowledgeable Idist, would have detected this."
The relevant words are "employed by a knowledgeable IDist". Lacking any knowledge about variation and selection, a seemingly "knowledgeable IDist" would likely conclude design, showing how the EF is not free from false positives. Thus the improvement proposed by Wilkins and Elsberry to include 'rarefied design', which would avoid false positives.
Concluding
quote:
If science is to be possible given a fallibilistic account of knowledge and if the knowledge it generates depends on empirical rather than innate rational information, then no rarefied design inference is needed, and all inferences are sensitive to the current state of knowledge. A priori assignments such as Dembski's filter requires make the human enterprise of discovery through trial and error impossible except where the metaphysical commitments of scientists and the broader society are unthreatened.
[ 16. May 2004, 22:50: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 17. May 2004 04:54
Jerry Bauer may wish to rework the section on abiogenesis. As it stands now, the characterization of Elsberry & Wilkins ("no evidence of popping into existence") doesn't match the refutation ("What unknown force could suddenly cause thousands of proteins to come together"). I suppose the goal was to argue that there is evidence of popping into existence, or that popping into existence is the only possibility, or that the poppingness of one's existence is not relevant to an application of ID, but none of these arguments seem to have materialized.
The only argument that directly addressed E&W's claim was a statement that the notion of gradual abiogenesis "defies common sense when one examines it", continuing with "Organisms must have genes to evolve." This is a weak argument on three counts. First, abiogenesis is essentially chemistry, and it's less than clear that "common sense" is a good guide when considering planetary-scale chemical reactions with largely unknown constituents. Second, genes aren't necessary for evolution (at least not genes composed of DNA with start and stop codons, etc.). Third, the whole point of abiogenesis is that you're starting with something that isn't an organism yet.
Also, I'm not sure the criticism of E&W's "god of the gaps" phrase hits the mark. The phrase "God of the gaps" is now used to describe a fallacious mode of argumentation, whether or not the conclusion is "God did it".
I think that the more pertinent criticism is that the filter actually doesn't tell us that we must continue through to a design inference if a problem is intractable or there is insufficient information to give a regularity or chance explanation. In fact, the filter says the opposite: the problem must be tractable and we have sufficient knowledge to show conclusively that an event is really low probability.
This error makes me think that E&W are paying attention to how IDists, including Dembski, attempt to use the filter. For instance, Dembski attempted to use the filter on the flagellum, a system which is ancient, complex, and only modestly well understood as a mechanical system, let alone as a potentially evolved system.
In this misapplication of the filter to a poorly-known situation, E&W's criticism finds its mark: by proceeding through the filter to generate an answer where one has insufficient knowledge, one is biased towards an incorrect finding of design. It is easy to mistake lack of knowledge for knowledge of impossibility. It is in this sense that the filter is prone to a "God of the gaps"-style problem. (Or rather, the filter isn't prone to it, but it's set up such that people who use it are likely to fall into that trap.) That's perhaps why E&W want to make an explicit "don't know" outcome--to really focus attention on whether or not chance processes and regularities are all known well enough to make the low-probability determination.
(The criticism of E&W's claim that the filter rejects too much seems on-target. The filter is a tool to detect design with confidence, not its absence. 95% confidence intervals reject too much also--that's the point!)
I think Jerry Bauer misses E&W's point about skirnobs, too. E&W complain:
quote:
Unless I already know a fair bit about the aims and intentions of skirnob designers, nothing is added to my knowledge of skirnobs by saying it is designed. . . .I may as well say that skirnobs are "gzorply muffnordled" for all it tells me.
If we extend this to toasters, E&W would be claiming that knowing that a toaster is designed (by Black & Decker, perhaps) doesn't tell you anything unless you know something about the designer. You can use a toaster to toast bread without knowing that it was designed to toast bread. Since we know a bit about Black & Decker and product liability laws in the U.S., we can have extra confidence that our toaster will, in fact, toast bread as promised, and not instead, say, electrocute us or catch the house on fire. But if we didn't know anything at all about the designer of the toaster, knowing that it was designed wouldn't be much help.
It's not an argument that toasters can't make toast. It's an argument that we can make toast just as well with a toaster made by a completely alien designer about whom we know nothing, and by a toaster that we don't even know was designed.
A more useful tack to take might be to argue that E&W underestimate the value of knowing that a designer exists. First, we can be interested whether something is designed or not even if the information isn't actually useful to us. Second, if A, B, and C are designed and have certain characteristics, and we find an object D of a similar class and it is also designed, we can guess that the designer of A,B,C was also the designer of D, and thereby gain knowledge of likely design features of D given our knowledge of the design features of A,B,C. (One could then counter that if we know enough about A, B, C, and D to assign them to the same class, we'd expect similarities in D regardless of whether all or none of them were designed, and so again, design isn't really helping us. And to counter that, one could postulate two classes of objects, designed by a certain designer and undesigned, which otherwise were kind of hard to distinguish from each other at first glance.)
[ 17. May 2004, 04:59: Message edited by: Rex Kerr ]
IP: Logged
|
|
Jerry D. Bauer
Member
Member # 756
|
posted 17. May 2004 16:38
Hey, Rex:
I may need to work on the abiogenesis section. Rather than “popping” into existence, I should have used the terms gradually morphing into existence.
But you are really claiming that genes are not necessary for evolution? I find this rather odd since the very definition of evolution is change in the gene pool of a population over time.
What other alternative methodologies can you offer to describe how that first organism appeared other than that all of this stuff just came together by chance. And whether, or whether not it ‘Popped” suddenly or over a period of time doesn’t change the probability math, does it?
But I disagree that chemistry is not common sense. For example, I’m not going to try to get hydrogen and oxygen to react hoping that gold is produced in the reaction.
I agree that the god-of-the-gaps argument is used in logic. However, I researched it briefly and this page seems rather typical:
"This argument has the form
1) There is a gap in scientific knowledge.
2)Therefore, the things in this gap are best explained as acts of God."
http://www.don-lindsay-archive.org/creation/god_of_gaps.html
I believe my assertion was correct in that the authors, whether intentionally so or not, introduced deity into the math of the EF.
As I pointed out somewhere, the EF is just a tool that needs to be used where it is called for. I would not attempt to use a screwdriver to drive a nail. My point is that the EF is not a universal, its just very effective on those things it applies to.
And, I’m afraid there is no bias in math. The filter either detects design mathematically or the ‘I don’t know’ factor comes into play. And if I understood you correctly, there is a big difference between detecting design and detecting non-design. Dismissal of a positive does not prove a negative.
There is no such thing as a ‘95%’ confidence interval in the EF. If something is determined to be complex and specified, it was designed because the odds are so staggering against it having formed on its own than it cannot happen in reality. I mean, Dembski’s 500 bits is three times the 10^-50 of Borel’s law where with odds at that level, Borel claimed nothing would happen in the real universe.
But if there were, it would still be a great tool for ID. Go wih Ocaam’s Razor and you will be right 80% of the time. The razor is not universal, but its used somewhere in science every day.
As to the toaster, I may need to clarify that this sucker is going to make toast or it will not. You having knowledge of the design engineer is not going to affect anything in reality. But if you care to have knowledge of that design engineer, go for it. But a design engineer is simply irrelevant when the question is was a system designed or is it the result of natural processes. The latter is all that Idists are concerned with.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 19. May 2004 02:53
I am uncertain whether Jerry and I are both authorized for posting, so I'll keep this brief.
A few points: popping "suddenly" vs. slowly does change the math if anything in the system is selectable. Oxygen + hydrogen producing gold probably didn't seem as far-fetched to ancient alchemists--it's common sense now that this doesn't happen because we've studied the relevant chemistry in great detail. But we have not studied prebiotic chemistry in great detail (in part because we aren't really sure what it was). Genes, meaning selectable heritable elements, are needed for evolution. Genes, meaning protein-coding DNA sequences, are not needed for evolution.
I agree with the parts about E&W being too critical of the filter's stringent criteria, so I won't comment on that further.
Finally, if you are agreeing that knowledge of a system being designed tells us nothing about the function or usage of a system, that's fine; but if so, you should modify your criticism of E&W to basically say, "Knowing it's designed isn't useful, but we are nonetheless curious, and the EF provides a mechanism to try to satisfy that curiosity."
IP: Logged
|
|
Moderator
Administrator
Member # 1
|
posted 20. May 2004 08:03
Jerry,
Comments like the following are not appropriate for our forums. Such statemetns are largely hostile and dismissive.
"The rest of the paper is largely nonsensical"
Please remember that posting at our forums is a privilege that can be taken away if we don't feel as if the poster is contributing to *our goal* of fostering a civil community of critical discussion.
IP: Logged
|
|
Jerry D. Bauer
Member
Member # 756
|
posted 20. May 2004 18:06
I fail to see how evolution by increments would change probability math. If I were to calculate the formation of a protein forming from 250 chirally ‘left-handed’ amino acids those odds come out to 1:10^74 against this happening without intelligence guiding the process. And if a simple organism had 100 of these proteins, I would multiply them all together to get the final calculation: 1:(10^74)^100. And this would be the final calculation no matter if the proteins came into existence all at once or one at a time over a period of 1000 years.
I agree that we have not adequately studied pre-biotic history, but for good reason. No one, it seems, can describe a pre-biotic atmosphere or an assembly scenario for that first protist. We can surmise and suppose, but that’s about as far as we can take it.
Rex states:
*****Genes, meaning selectable heritable elements, are needed for evolution. Genes, meaning protein-coding DNA sequences, are not needed for evolution.*****
I’m afraid you lost me here with this comment. The EF can be used to show (beyond a shadow of a doubt to those who will keep an open mind) that the first organism was designed. However, detractors of this point out that the first organism probably occurred as a series of steps rather than via one ‘poof.’ What I fail to understand is how this could happen. And if it did, how would this affect the EF?
A major tenet of evolution is that single organisms don’t evolve, populations do. And I fail to understand how a single organism could evolve before it can eat, breathe, reproduce or interact at all with its environment. One glaringly obvious point is that proteins don’t have genes with which to evolve.
Finally, I feel that ID has no place for a designer, only design. If that designer were deity, one would have to leave science and enter the realm of meta-physics, or theology. The scientific method disallows us of doing this because although ID is based on teleology, its governed by methodological Naturalism.
So, as for me, I’m not curious at all about the designer. And if I were, I still wouldn’t use the lab to satisfy this curiosity.
[ 20. May 2004, 18:47: Message edited by: Jerry D. Bauer ]
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 20. May 2004 22:58
Jerry D Bauer offers a common type of calculation for the formation of a biological entity - one which uses the basic multiplicative rule for multiple independent events. Using this rule assumes that all the component parts came together by chance (let us call this the “pure chance” hypothesis,) which is quite unrealistic and not relevant to genuine biological hypotheses.
In particular the pure chance hypothesis neglects the role of any laws by which components parts might combine by other than pure chance, and the pure chance hypothesis neglects the effect a series of steps might have by which the entity in question is formed over time.
Jerry D Bauer illustrates this common argument when he writes,
quote:
I fail to see how evolution by increments would change probability math. If I were to calculate the formation of a protein forming from 250 chirally ‘left-handed’ amino acids those odds come out to 1:10^74 against this happening without intelligence guiding the process. And if a simple organism had 100 of these proteins, I would multiply them all together to get the final calculation: 1:(10^74)^100. And this would be the final calculation no matter if the proteins came into existence all at once or one at a time over a period of 1000 years.
I have been thinking about a simple way to explain the flaws in this reasoning. I am not a biochemist, and even if I were the actual laws and combinations would be extremely complex. So let me simplify things (in the time-honored tradition of those who have offered mousetraps, combination locks, scrabble tiles, and stacks of coins) by talking about the game of Yahtze.
The Rules of Yahtze
For those of you who don’t know the rules, the game of Yahtze is played by throwing five dice and trying to get all five dice the same number. You get up to three throws, and after each throw you can leave as many dice lying on the table as you wish. For instance, if your first throw had 3 sixes, you would leave them there and try to get 2 more sixes with two more throws of the other two dice. If you get all 5 dice the same, you have a “yahtze.” (There are more rules about other, lesser combinations you can get, but we will ignore all that.)
So let’s think about the probability of getting a yahtze.
A simple strategy
The chances of getting, for instance, 5 sixes on one throw would be 6^5 = 1 out of 7776. Since you could also get 5 fives or 5 fours, etc., then odds of throwing a yahtze on one throw would be 6 * 1/7776 = 1/1296. These are simple applications of the multiplicative rule for independent events, as each dice is independent of each of the other dice. So far so good.
Given that you don’t throw a yahtze on the first throw, you need to have a strategy for the second throw. (Yes, I know we are bringing intelligence in here - we will deal with that later.) A simple strategy (but not the best) would be to just throw all 5 dice again, and if that doesn’t get a yahtze, try again. A naive application of the addition rule would say that probability of getting a yahtze after three throws would be 1/1296 + 1/1296 + 1/1296 = 1/432
The situation is a little more complicated than that though, because you wouldn’t try for a yahtze on the second throw if you got one on the first throw. So the actual probability of this simple strategy would involve a probability tree like this:
Throw 1: yes 1/1296 no 1295/1296
Throw 2 (if throw 1 was no): yes 1/1296 no 1295/1296
Throw 3 (if throw 1 was no): yes 1/1296 no 1295/1296
So the probability of a yahtze would be 1/1296 + 1295/1296*1/1296 + 1295/1296*1295/1296* 1/1296 = 1/432.33
A slight difference, but the principle is important - once you start adding steps which can be affected by previous steps, simple combinatorial probability or simple exponential probability no longer applies
A better strategy
Now real people don’t use this simple strategy. Instead, the most common (and perhaps most effective strategy) is to always keep any groups of the same number (like 2 sixes,) and always keep groups with more members if you have a choice (that is, if you had 3 sixes and 2 fives, keep the 3 sixes.)
As you might guess, calculating the odds for this strategy now becomes a quite complicated probability tree. For example, on throw 1:
if a yahtze, end
if four of a kind, keep them
if three of a kind, keep them
if two of a kind or two pair, keep a pair
if no numbers match, just throw again (for you purists out these, real Yahtze gives points for straights, but we are ignoring that.)
Then, based on the probabilities of each of those (which can be calculated using combinatorial probability) move on to throw 2:
if four of a kind on throw 1 and a match on throw 2, end
if four of a kind on throw 1 and no match on throw 2, go to throw 3,
and so on.
You can see that this is not an impossibly difficult problem, but certainly more than I’d want to go through here. However I did work all this out a very long time ago (before I knew much probability - I did it the long way.) Suffice it to say that your odds of getting a yahtze are considerably greater than with the simple strategy. (I will dig out my work and post the actuals odds at a later time.)
Eliminating intelligence
Now let’s try to make this more like the real world by substituting laws for the intelligent player. Here are the rules.
1) Every minute (a throw) the dice randomly jiggle so as to produce a random number (random in both the sense that the numbers 1 through 6 are equiprobable and random in respect to the goal of the game or the state of the other dice.)
2) However, if two dice have the same number, those dice “bond” together and do not jiggle any more - they now stay the same except for the application of rule 3.
3) If there are two bonded groups on the table, the strength of the higher group (either having more matching dice or the higher number on the dice if they have the same number of dice,) dissolves the bonds of the weaker group (and therefore they jiggle again on the next throw.)
These are the same rules followed by the player above, but now the rules are embedded in the interactional nature of the dice themselves - they are analogous to the laws of nature.
Therefore, the odds of getting a yahtze are the same as above.
One more point
Suppose instead of three throws we get five throws, or ten. The pure chance hypothesis represented by the simple strategy is approximately additive - for ten throws, the probability of a yahtze is approximately 10 * 1/1296 or 1/130.
However, for ten throws the mechanical rules outlined above will almost certainly produce a yahtze - the probability approaches 100%.
Summary
Let me return to my original claim: “In particular the pure chance hypothesis neglects the role of any laws by which components parts might combine by other than pure chance, and the pure chance hypothesis neglects the role that a series of steps by which the entity in question is formed over time.”
My example shows two things that are relevant to the real world:
1) There are laws which cause things to interact together so that once they have combined they stay that way. The pure chance hypothesis does not take this into account.
2) Given that things bond together as they do, the effect of repeated steps becomes important. In the example above, there is “selection” going on, not in the biological sense related to replicating entities but in the simple sense that once something is build it isn’t as sensitive to (or is even immune to) the random and chance interactions it might have had before it bonded.
I believe these considerations strongly support my original claim that the “pure chance” hypothesis “is quite unrealistic and not relevant to genuine biological hypotheses.” Furthermore, Jerry says he doesn’t see why multiple steps would change the odds, but I believe my example shows clearly that with natural laws operating at each step, multiple steps can dramatically change the probabilities.
I am also aware that one response to this post is that we don’t know what laws and steps might have led to the first organism (which is often but not always the subject of the pure chance hypothesis.) I grant that, but we do know that there are natural laws which work to combine elements of the chemical world and we do know that things happen in steps whereby the result of one step affects the next step.
My analogy is meant to highlight the fact that this is the way the world works. The pure chance hypothesis is meaningless irrespective of how much we do or don’t know about any particular biological entity, because we do know that things don’t just come together by pure chance. We don’t see that happening now - and there is no reason to think that the world was any different in this respect 3 billion years ago.
(P.S. A little side point: if in fact natural laws did not operate to bind things together, then even if Jerry’s 2500 amino acids happened to come together in just the right way to make his organism, there would be no reason for them to stay together - they would just move away the next moment and go back to randomly bouncing around until some other improbable event happened. The pure chance hypothesis carries with it the implicit assumption that only that one improbable combination has the ability to create a permanent biological or biochemical entity, but that all other lesser and more probable combinations will not combine to form simple entities by which more complicated entities might be built. This also is an unrealistic aspect of the pure chance hypothesis.)
Thanks for reading. I am interested if others find this analogy useful in explaining at least some of the flaws in the “pure chance” hypothesis when related to biology.
[ 21. May 2004, 08:07: Message edited by: Evan ]
IP: Logged
|
|
Jerry D. Bauer
Member
Member # 756
|
posted 21. May 2004 16:43
Hmmm….What a thought provoking paper, Eric. Very well thought out, carefully crafted and well communicated.
If I might sum up your argument (let me know if I’m missing something) you are addressing a concept we might term ‘stickability.’ IOW, the odds tend to reduce as events occur and become “stuck” or stabilized into the system.
If I might relate this to a system in a mind experiment (unfortunately about all we can do over the Internet) In an organism consisting of 100 proteins, if these proteins form all at once, we have one odds calculation. But if 50 proteins form in single, individual steps and then the other 50 form together, we come out with different figures.
Let’s regress to an uncomplicated system and then take this thread directly into a simple organism to apply it. And I will begin with…yes….the dreaded coin toss.
If I flip a coin what are the odds of me getting heads or tails? 1:2. If I flip 50 coins and I get 25 heads and 25 tails, what are the odds when I flip that 51st coin that I will receive head or tails? 1:2. If I have flipped 99 coins and 47 have come up heads and 52 have come up tails, what are the odds for heads or tails in that 100th coin? 1:2.
Well what are the odds if I flip 100 coins they all will come up heads? 1:(.5^100). But what if I have already flipped 50 of the coins and 25 of them are tails and 25 of them are heads. Now what are the odds that all 100 coins will come up heads? They’re still the same 1:(.5^100). I’m not getting all heads, but with odds against me of getting them, I’m not surprised at the result.
So let’s place all 100 coins in a bag, shake them up all at once and see how many heads I get. What are these odds? 1:(.5^100). So it doesn’t really matter if I flip the coins all at once (a ‘poof’ as in spontaneous generation) or I flip them one at a time (individual, incremental steps), the odds in the big picture do not change.
Of course, chemical reactions are not coins and this happens a bit different in the real world.
For two atoms to “bond” (join together into a molecule) they must be within an “interacting neighborhood.” In fact, in order for two atoms to react together, they must be in the area of about 100 picometers (10 to the -10 power meters) in distance from one another.
The universe is big. And atoms must be moving in order to come into the “neighborhood” of another atom. The faster they are moving, the more opportunities they have to form a bond.
But this gets a little hairy because if they are moving too fast, the momentum will shoot them past each other before they can bond.
And, the temperature can‘t be too cold as reactions will not effectively occur and if it is too hot more bonds will be broken than are formed, and even when the temperatures are perfect, “bonds” of a long molecular chain may be broken simply because a random high energy atom or molecule knocks it loose. The point is, there is a certain finite number of opportunities available, even in 50 billion years for a reaction to occur in reality
For these reasons, Brewster and Morris concluded, based upon the size of the universe, the temperatures under which bonding occurs, the surmised age of the universe, the nature of bonds and how they form and break-- that 10 to the 67th power is the ultimate upper threshold for any chemical event to happen--anytime, anywhere in the universe, even in 50 billion years.
Dembski defines a universal probability bound of 10^-150, based on an estimate of the total number of processes that could have occurred in the universe since its beginning. Estimating the total number of particles in the universe at 10^80, the number of physical state transitions a particle can make at 10^45 per second (Planck time, the smallest physically meaningful unit of time) and the age of the universe at 10^25 seconds, thus the total number of processes involving at least one elementary particle is at most 1:10^150. Anything with a probability of less than 10^150 is unlikely to have occurred by chance. Previous to Dembski, statisticians concluded through Borel’s Law that 1:10^50 was the upper limit odds in which anything could actually happen.
The smallest known bacteria I’m aware of consists of around 500 proteins but I don’t think anyone would disagree with me that I am safe in using a 100 protein scenario in order to form an organism that could remotely be called life.
Proteins from which all of life is based are formed from amino acids. And these proteins are usually chains of from 50 to 50,000 amino acids.
Chemist, Stanley Miller showed long ago that under the correct conditions we can create amino acids in a beaker. A chirality problem is they come out completely “racemized.” The amino acids produced by Miller consisted of equal amounts of “right-handed” and “left-handed” molecules. The atoms that react to form amino acids bond together into cork-screw shapes--these cork-screws can curve to the right (right-handed) or to the left (left-handed). But a useable protein for life has to be composed entirely of left-handed molecules.
So, when an amino acid adds itself to a protein chain, the odds are one in two that it will be left-handed. That’s not a big deal if the protein chain is extremely short--say three amino acids long. Our probability would be one chance in 2 to the 3rd power or 1:8. That’s not bad odds for this type of thing.
So, let’s look at this primeval ooze from which that first protist popped and we are going to surmise that this ooze was racemized amino acids that had occurred naturally.
The odds against assembling a protein chain consisting of only left-handed amino acids by chance is 2 to the “n” th power. And “n” is the number of attached amino acids in the protein. So its not difficult to calculate that the odds against assembling a useable protein of only 250 left-handed amino acids from a racemized mixture is one chance in 2 to the 250th power. This is about 1 chance in 10 to the 74th power.
Well shoot, we are already past the Borel’s Law barrier with one tiny protein and we are nowhere near our organism. It would only take one more to catch up with Dembski’s UPB.
And some of the proteins found in nature are 50,000 chained amino acids. The odds of assembling a protein that long are 1:10^15,000
These were designed.
To calculate the organism, we have to multiply together the odds of each one of our amino acids. When we do we come out with a 1:10^7400 chance that this tiny, highly unrealistic and overly simplistic organism could ever form. These are staggering odds that could not occur in reality.
Now we can see why some Idists calculate that the odds against a fully functioning, much more complex human cell occurring by chance is one chance in 10 to the 100 billionth power. That’s one hundred billion zeroes. Us computer geeks can think of it as a 100 gigabyte hard drive full of nothing but zeroes.
And whether or not this cell forms one step at a time, or all at once, these odds don’t change.
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 21. May 2004 19:22
quote:
If I flip a coin what are the odds of me getting heads or tails? 1:2. If I flip 50 coins and I get 25 heads and 25 tails, what are the odds when I flip that 51st coin that I will receive head or tails? 1:2. If I have flipped 99 coins and 47 have come up heads and 52 have come up tails, what are the odds for heads or tails in that 100th coin? 1:2.
Well what are the odds if I flip 100 coins they all will come up heads? 1:(.5^100). But what if I have already flipped 50 of the coins and 25 of them are tails and 25 of them are heads. Now what are the odds that all 100 coins will come up heads? They’re still the same 1:(.5^100). I’m not getting all heads, but with odds against me of getting them, I’m not surprised at the result.
So let’s place all 100 coins in a bag, shake them up all at once and see how many heads I get. What are these odds? 1:(.5^100). So it doesn’t really matter if I flip the coins all at once (a ‘poof’ as in spontaneous generation) or I flip them one at a time (individual, incremental steps), the odds in the big picture do not change.
Jerry, what if you throw your coins onto a surface to which only "head" faces stick, but "tails" bounce off, flipping in the air and falling back again? Once all coins have settled, does that alter the probability outcome of the coin toss experiment?
IP: Logged
|
|
Jerry D. Bauer
Member
Member # 756
|
posted 21. May 2004 19:43
*****Jerry, what if you throw your coins onto a surface to which only "head" faces stick, but "tails" bounce off, flipping in the air and falling back again? Once all coins have settled, does that alter the probability outcome of the coin toss experiment?*****
It may, I'm not really sure but I would be interested in furthering the concept for clarification.
I'm trying to think of a real-life system in which this occurs. What you have done is to create a thermodynamical system in which entropy would steadily decrease over time. Interesting.
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 21. May 2004 19:58
quote:
It may, I'm not really sure but I would be interested in furthering the concept for clarification.
I'm trying to think of a real-life system in which this occurs. What you have done is to create a thermodynamical system in which entropy would steadily decrease over time. Interesting.
Of course, there are plenty of open systems in real life in which entropy decreases. But, more specifically, you might be interested perhaps in something like this: quote:
Orig Life Evol Biosph. 2004 Feb;34(1-2):93-110.
Spontaneous onset of homochirality in oligopeptide chains generated in the polymerization of N-carboxyanhydride amino acids in water.
Hitz TH, Luisi PL.
This article is concerned with the spontaneous onset of homochiral oligopeptide sequences. We will show that the polymerization of hydrophobic NCA (N-carboxyanhydride = cyclic anhydride)-amino acid racemates (i.e. tryptophane, leucine and isoleucine) in aqueous solution yields oligopeptides that are characterized by a high degree of homochiral sequences. Furthermore we will show that quartz enhances efficiently the mole fraction of oligopeptides with homochiral sequence by selectively adsorbing the more stereoregular oligopeptides from an aqueous solution of oligo-D,L-leucine. We find in particular that the mole fraction of the adsorbed homochiral 7mers is 17 times larger than the mole fraction calculated for a theoretical, random process. Experimentally the stereoisomer distribution for each oligomer length can be determined by the use of enantio-labeling and LC-MS (Liquid Chromatography-Mass Spectrometry). Furthermore, if we start the polymerization with an enantiomeric excess (e.e.) of 20% of L-leucine (L-amino acid: D-amino acid = 6:4, molar ratio) we observe a chiral amplification in the enantiomeric homochiral oligopeptides. We think that such processes are relevant to the chemical evolution of single handedness.
or this quote:
Selective adsorption of L- and D-amino acids on calcite: Implications for biochemical homochirality.
Hazen RM, Filley TR, Goodfriend GA.
Proc Natl Acad Sci U S A. 2001 May 8;98(10):5487-90.
The emergence of biochemical homochirality was a key step in the origin of life, yet prebiotic mechanisms for chiral separation are not well constrained. Here we demonstrate a geochemically plausible scenario for chiral separation of amino acids by adsorption on mineral surfaces. Crystals of the common rock-forming mineral calcite (CaCO(3)), when immersed in a racemic aspartic acid solution, display significant adsorption and chiral selectivity of d- and l-enantiomers on pairs of mirror-related crystal-growth surfaces. This selective adsorption is greater on crystals with terraced surface textures, which indicates that d- and l-aspartic acid concentrate along step-like linear growth features. Thus, selective adsorption of linear arrays of d- and l-amino acids on calcite, with subsequent condensation polymerization, represents a plausible geochemical mechanism for the production of homochiral polypeptides on the prebiotic Earth.
or this
quote:
Asymmetric autocatalysis and its implications for the origin of homochirality
Donna G. Blackmond
PNAS 101, 5732-5736, 2004
An autocatalytic reaction in which the reaction product serves as a catalyst to produce more of itself and to suppress production of its enantiomer serves as a mechanistic model for the evolution of homochirality. The Soai reaction provided experimental confirmation of this concept, nearly 50 years after it was first proposed. This Perspective offers a rationalization of the Soai autocatalytic reaction; accounting for enantiomeric excess and rate observations, that is both simple as well as gratifying in its implications for the chemical origin of life.
Truth is, we know of several natural conditions and processes that can generate homochiral products from reactive racemic mixtures.
[ 21. May 2004, 20:19: Message edited by: charlie d. ]
IP: Logged
|
|
Jerry D. Bauer
Member
Member # 756
|
posted 21. May 2004 20:54
.
If you have that paper you might email it to me as they want 37 bucks to view it on-line and I ain’t going that direction.
The abstract does not give the methodologies by which he came to the conclusions in the abstract. So it really isn’t telling us much, I‘m afraid.
If this is the same study we were discussing at ARN then it would seem not to say that much, practically, about abiogenesis or evolution in that the methodologies used in the lab were not the same as those found in natural processes. But, I reserve judgment until I can read the entire paper.
So back to your hypothetical thought experiment. There are entropic decreases found in systems every day, but I feel those systems are not comparable to your analogy.
And I think I now know why. I’ve come to the conclusion that you have designed a system governed by intelligence. Intelligence would be necessary for a governing mechanism to view a coin and decide whether it is heads or tails, then reject or accept it.
In the real world, I could just do this for you. I could throw 100 pennies on the coffee table, pick up all the tails and throw them again until all are heads. This is not even probability because the odds are 100% that all of the coins will eventually be heads. Instead of probabilities acting on natural processes, this is purposeful intelligent design.
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 21. May 2004 21:07
Jerry D Bauer starts by writing, “Hmmm….What a thought provoking paper, Eric. Very well thought out, carefully crafted and well communicated.”
Thank you for the complementary words, as I strive to be clear, concise, and stick to whatever point I am trying to make (although I am Evan, not Erik.)
I’d like to start my response by saying that I don’t believe any of the discussions about “upper probability bounds” (UPB) are relevant - not Dembski’s or Borel’s or Brewster and Morris’s. The reason is that until we actually figure out how to realistically compute biological probabilities, speculations on what the UPB is that divides naturally caused events and designed events are just that - speculations. We need to get some real numbers before we have an idea about how probable or improbable a biological event is.
And, of course, my central argument is that the pure chance hypothesis is fundamentally flawed and of no value in helping us produce realistic probability calculations about biological events (and therefore of no use in calculating a UPB.)
My Yahtze example is meant to get us thinking about this by incorporating two aspects of the real world into the model - natural laws which cause certain things to happen with high probability and a sequence of steps (a history) by things can happen with a higher probability than if they happened in just one step.
Jerry accurately summarizes my argument by writing “you are addressing a concept we might term ‘stickability.’ IOW, the odds tend to reduce as events occur and become “stuck” or stabilized into the system.”
However, the examples of coin tosses he then gives does not take this concept into account. They are just again a set of independent events in which nothing “sticks” from step to step - they are just the pure chance hypothesis again.
Jerry starts out by correctly pointing out that the odds of getting a head on any one throw are 1/2 irrespective of past events - that is that the chances of a head are random in respect to the current state of affairs as produced by the previous tosses.
But the issue is not what the odds of a coin toss are - the issue is what happens after the coin toss? Are there rules which select for some situations (in the Yahtze example, the “bonding” of dice with the same number.) Jerry’s example doesn’t take this into account.
quote:
So let’s place all 100 coins in a bag, shake them up all at once and see how many heads I get. What are these odds? 1:(.5^100). So it doesn’t really matter if I flip the coins all at once (a ‘poof’ as in spontaneous generation) or I flip them one at a time (individual, incremental steps), the odds in the big picture do not change.
In this case, Jerry is correct that it makes no difference whether we flip them one at a time or all at once, because all the flips are independent events and there is no “connection” (no laws) by which the current state influences what happens next.
So this coin toss example is just another pure chance hypothesis that does not address the two issues brought up by my Yahtze example.
Jerry then states that “chemical reactions are not coins and this happens a bit different in the real world,” which is true, and then writes,
quote:
For two atoms to “bond” (join together into a molecule) they must be within an “interacting neighborhood.” In fact, in order for two atoms to react together, they must be in the area of about 100 picometers (10 to the -10 power meters) in distance from one another.
Then Jerry goes on to discuss various factors which affect whether molecules can bond (speed, temperature, etc.), and then goes on to discuss the UPB issue.
However, all of these considerations are made moot by the simple observation that chemical interactions and bonding go on all over the world all the time - billions of them every second (many times a billion, actually) for billions of years. We live in a universe where things want to bond (pardon the anthropomorphism,) and sometime it’s darn hard to prevent it.
As I explained in my P.S. yesterday, the odd thing about Jerry’s pure chance hypothesis is that not only does it leave out the possibility of things bonding (“sticking”) so as to produce steps by which the eventual entity is formed, but it (the chance hypothesis) also includes the implicit assumption that when all those parts suddenly come together then they do stick, as if that particular combination was the only possible combination that would cause everything to stick together, but that simpler events did not have this possibility.
It seems to me that the chance hypothesis includes the idea that, really, nothing is possible, because if things come together by chance without any influence of natural laws, then they will also fall apart by chance almost immediately, as if the whole universe were nothing but billiard balls that only occasionally made a fleetingly ephemeral pattern that dissipated as quickly and as randomly as it arose.
And of course we know the world doesn’t work like that.
As a last example, Jerry considers the example of a bacteria, and again offers a pure chance hypothesis. He writes,
quote:
So, when an amino acid adds itself to a protein chain, the odds are one in two that it will be left-handed. That’s not a big deal if the protein chain is extremely short--say three amino acids long. Our probability would be one chance in 2 to the 3rd power or 1:8. That’s not bad odds for this type of thing.
and he goes on to offer pure chance calculations to show that the odds of all the amino acids coming together by chance are, once again, astronomically small:
quote:
To calculate the organism, we have to multiply together the odds of each one of our amino acids. When we do we come out with a 1:10^7400 chance that this tiny, highly unrealistic and overly simplistic organism could ever form. These are staggering odds that could not occur in reality.
And whether or not this cell forms one step at a time, or all at once, these odds don’t change.
Notice there is no hint here that there might be laws of chemistry at work, and no hint that Jerry has taken into consideration the “stickability” issue that I raised with the Yahtze example. For instance, there is all sorts of research going on in the world of both organic and inorganic chemistry about chirality: about small differences in left and right-handed molecules, about how once a particular chirality got started in life it might have tilted the balance towards that chirality until one side took over, about the effect of polarized light on chirality,and so on.
The point is not that we know exactly how the chirality of amino acids in life got established. The point is that those scientists studying the matter are taking the “Yahtze” approach, not the “coin tossing” approach - they are looking for chemical and physical interactions that would cause, through a long series of steps, things to come together in probable ways.
The “pure chance” hypothesis doesn’t take chemistry and physics into account, much less biology, and that is why it is not at all realistic, and thus not useful. Probabilistic calculations, if they are to establish anything at all, must be based on methods that have been developed to take the real “rules of the game” into account - chemistry and physics, and the pure chance hypothesis doesn’t do that.
[ 21. May 2004, 21:41: Message edited by: Evan ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 21. May 2004 22:08
quote:
If this is the same study we were discussing at ARN then it would seem not to say that much, practically, about abiogenesis or evolution in that the methodologies used in the lab were not the same as those found in natural processes. But, I reserve judgment until I can read the entire paper.
Forget abiogenesis one minute, and the experimental details. All those papers show is that natural law can in the appropriate conditions generate homochiral polymers from racemic mixtures. What seemed on its face to be impossible, based on random chance alone, and to be the certain product of intelligence, in the light of this information might well be the inescapable product of natural law. Much like neat layers of gravel, sand and clay do not happen by chance, but can either be the product of design, or plain natural sedimentation. quote:
And I think I now know why. I’ve come to the conclusion that you have designed a system governed by intelligence. Intelligence would be necessary for a governing mechanism to view a coin and decide whether it is heads or tails, then reject or accept it.
But of course quartz and calcite are not intelligent. They just happen to be able to sort racemic mixtures based on chirality because of their inherent chemical/physical properties. Just like sedimentation is not intelligent, but efficiently distinguishes between gravel, sand and clay.
So, if you are now postulating that natural laws, which give rise to calcite and quartz (and abiogenic aminoacids, and sedimentation) can themselves have been intelligently designed, sure, why not? It's a metaphysical position, as science can only reach and explain natural laws, and not past them, but it's as fair a metaphysical position as any other which does not fly in the face of empirical evidence.
IP: Logged
|
|
|
Printer-friendly view of this topic