|
Author
|
Topic: Responses to Criticisms of Specified Complexity
|
Moderator
Administrator
Member # 1
|
posted 06. May 2003 09:09
For the time being, I am merely going to post a warning to Pim. I was tempted to go in and rip out a bunch of his post.
The problem here is that I will not be provoked. I won't tolerate stuff like:
"I'd hope that this board will not tolerate Nelson's comments."
"I would like to voice my discomfort and concern with the comment made by Nelson"
If you've got a concern, send me an email: moderator@iscid.org
Though Pim's concerns are partially legitimate (Nelson could have said what he said more tactfully) Pim's "concern" post rocks the boat more than is necessary.
Finally, I gave this warning to Frances several times before he was permanently banned from our board. If every single one of your posts contains the word "eliminative" in it, then I will judge that you've got an axe to grind here at Brainstorms, and will take away your posting privileges.
Nelson, you too should remember that this board is not your average internet discussion forum. One of our rules is that posters treat each other professionally. Please avoid accusing individuals or groups of people of having a lack of will.
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 06. May 2003 09:52
With all due respect to the moderator, I see nothing wrong with Pim's post (and I think I am fairly sensitive to the standards ISCID wants to maintain.)
I also don't understand the problem with the paragraphs containing the word "eliminative." Eliminating hypotheses of regularity and chance is fundamental to Dembski's ideas, and therefore many posts revolve around this issue: how to do it, can it be done, can it work reliably.
All of us have various issues that are most important to us: I am interested in probability calculations, common descent, ways of understanding how design might enter the world, etc. Mike Gene can be counted on to discuss the early development of biological features, Gedanken is interested in logical ideas concerning ID, Erik offers mathematical posts, etc. Pim is interested in the eliminative nature of the EF. When I read the paragraphs that contain this word in his last posts, it seemed to me he was making perfectly legitimate points.
The moderator sometimes says that old and worn-out topics are not welcome. The problem is that some topics appear old and worn-out because in fact they are central to the disagreements about ID and they are far from resolved - therefore they reappear. The eliminative argument will continue to reappear as long as all the various issues about how in fact regularity and chance can be eliminated in order to infer ID remain in such an undeveloped state.
[ 06. May 2003, 10:19: Message edited by: Evan ]
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 06. May 2003 11:51
I apologize to the moderator but Nelson's comments struck a wrong nerve.
Here is my updated posting that takes into consideration the moderator's suggestions. I would also like to thank Evan for his comments.
I would like to comment on Nelson's quote that:
quote:
C12: That no one has been applying the calculation is a testament, in fact, to the lack of will exhibited by Darwinists to test their pathways
Why are Nelson's comments erroneous imho:
First of all let me point out that it is the nature of the ID inference as "the set theoretic complement of the disjunction regularity and chance" (Quotations from Dembski) and the use of complexity as a measure that requires the ID proponent to calculate the probabilities for Darwinian and other chance or regularity pathways in order to be able to infer design. So it is obvious from the start that it is not the Darwinist's responsibility to provide for such probability estimates but rather the responsibility and task of the ID proponent.
Secondly let me point out that Darwinist are not unwilling to provide evidence for their proposed pathways but probability estimates are but one way of achieving such and in fact given the nature of the beast, probability estimates are intrinsically difficult if not impossible, as Dembski and others have found out.
Thirdly Darwinists are not unwilling to consider the evidence and falsification of their ideas and in fact in many of instances they use whatever available data to formulate ways to support or eliminate their hypotheses. Probability estimates for their pathways, due to their nature of being (almost) impossible to accurately calculate are not required nor even necessary.
IP: Logged
|
|
Moderator
Administrator
Member # 1
|
posted 06. May 2003 12:18
Evan,
Good points, but my concern still stands. The issue that concerns me is whether Pim is at Brainstorms to indoctrinate or to discuss. If his every post is harping about the eliminative nature of Dembski's arguments, then it raises my suspicions, especially on a forum where discussions are supposed to be moving forward.
Perhaps Pim is very interested in this topic. In fact, that seems to be the case. But repetition of criticism is a red flag to the moderators. One of the primary red flags. It should be fairly obvious to you why this is the case. We want to create an environment where actual thinking is done, not where people repeat the same mantra over and over. It is easy to memorize a criticism, and then repeat it...repeat it...repeat it...repeat it...etc. However, in doing so, people often miss the nuances of the discussion and merely force-fit the conversation into their "mechanical criticism." Pim is on the verge of realizing this tendency.
This is the end of the online discussion about moderator policy. For future offline discussion, please contact me at moderator@iscid.org
[ 06. May 2003, 12:20: Message edited by: Moderator ]
IP: Logged
|
|
Erik
Member
Member # 160
|
posted 06. May 2003 13:04
Nelson Alonso, because I think that you read the stuff you reply to very sloppily, I will not engage in any long discussion with you. I will reply to your comments on C10-14, but I won't proceed beyond two exchanges of posts with you.
C10: You claimed that Dembski indeed took into account the hypothesis that T3SS evolved to flagella and that the set (not "sample space") of testable evolutionary pathways seems empty. None of your points, whether correct or not, are relevant. To infer ID by Dembski's standards, we must identify a sample space (do not confuse the sample space, containing among other things the event under study and the specification, with the set of hypotheses to be eliminated), a set of hypotheses, ... and then we must do the calculation. My objection C10 does not concern the validity of Dembski's method nor is it an observation that details of his examples should have been different. My criticism C10 concerns the vagueness and lack of details (not correctness) in the few examples Dembski has provided. There may well be other ways to conclude ID, in C10 I simply note that Dembski fails to live up to the standards he set up for himself.
C11: You claimed that having eliminated non-ID hypotheses is not part of the definition of "specified complexity". In reply I note that p. 72-73 of the book "No Free Lunch" supports my statement. Indeed, on page 73 we can read "Specified complexity, the conclusion of the Generic Chance Elimination Argument, is a statistical notion", which together with the context suggests that "specified complexity" really is an event for which we have eliminated a set of non-ID hypotheses by Dembski's method. This is clearly not equivalent to Paul Davies's notion with the same name.
C12: It is Dembski, not evolutionary biologists, who set out to test hypotheses by "doing the calculation". That neither evolutionary biologists nor any other scientists have put their models in a form convenient for Dembski's method does not mean that they do not test their models. It means that they use other tests than Dembski's method of inference.
C13: You didn't significantly adress my criticism. The best way to adress it would be to:
(i) Explain why we should try to eliminate non-ID hypotheses and then automatically accept a general ID hypothesis (the EF method), rather than trying to eliminate ID hypotheses and then automatically accept a general non-ID hypothesis (the DEF method).
(ii) Explain why we hypotheses should be treated differently by our methods of inference depending on whether they feature intelligent agents.
(iii) Explain why intelligence being a fuzzy property is not a problem (or, alternatively, explain why we can assume that intelligence is a sharp either-or property).
C14: Then state the rigorous criterion for deciding if background knowledge "explicitly and univocally identifies the rejection function f". Some examples would nice too.
Erik
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 06. May 2003 13:45
Before I begin my responses in the various threads for today, I do apologize for using the phrase "lack of will". I could have phrased this a little better.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 06. May 2003 16:55
Erik,
C10: I don't think that Dembski was ambiguous with respect to the sample space. Here is Dembski:
quote:
In assessing whether the bacterial flagellum exemplifies specified complexity, the design theorist is tacitly following Earman's guidelines for making an eliminative induction work. Thus, the design theorist orders the space of hypotheses that naturalistically account for the bacterial flagellum into those that look to direct Darwinian pathways and those that look to indirect Darwinian pathways (cf. Earman's requirement for an ordering or topology of the space of possible hypotheses). The design theorist also limits the induction to a local induction, focusing on relevant hypotheses rather than all logically possible hypotheses.
Now it is quite obvious, in my understanding of the calculation that Dembski was eliminating both chance and a combination of chance and natural selection, and doing so by eliminating the relevant chance hypothesis, natural selection. Dembski did not take the type III secretory system into account in No Free Lunch, but he did so in a reply to Miller. Selection theory assumes functional intermediates leading all the way to the flagellum. Dembski treated the DCO with this hypothesis in the sample space, and did the calculation, where functional intermediates would lead all the way up to the bacterial flagellum. Instead what the calculation shows is a sea of nonfunctionality, where natural selection becomes sterile (and therefore it is nothing but random assembly). This was the testable hypothesis in the sample space that Dembski was using. I, for one, don't find this ambiguous at all, although there is a lot more work to be done.
C11: With respect to eliminating chance hypothesis as part of the definition of specified complexity, I disagree that pages 72-73 show this. What we see with the Argument is the consequence of specified complexity. Remember that a specification refers to an external pattern. An external pattern cannot be explained by any chance hypothesis, this is simply a consequence of the fact that the pattern refers to function (in the case of biology) that is not reducible to any one of the components. To see this consider Dembski's statement:
quote:
it is not fair to say that [specified complexity] is established via a purely eliminative argument. If the argument were purely eliminative, one might be justified in saying that the move from specified complexity to a designing intelligence is an argument from ignorance...it takes considerable background knowledge to come up with the right pattern (i.e. specification) for eliminating all those chance hypothesis and thus for inferring design.
p. 111
Notice how he equates finding the right pattern with eliminating all relevant chance hypothesis. In other words it is the "right pattern", not just eliminating the hypothesis that points us to a specification. And you can see this "right pattern" in Paul Davies, which is exactly equivalent to Dembski's notion of specified complexity:
quote:
To bring out this point clearly, consider the way in which the four bases A, C, G and T are arranged in DNA. As explained, these sequences are like letters in an alphabet, and the letters may spell out, in code, the instructions for making proteins. A different sequence of letters would almost certainly be biologically useless.
Note how equivalent these two notions of specificity are. Both Dembski and Davies points to biological function and how the specification is beyond the individual components.
C12: When I see a statistical test for the ability of natural selection building any complex system I will recant my response to C12, even if it does not use Dembski's. It is quite disenguous, in my opinion, to say that no Darwinist uses Dembski's calculation because it is useless when in fact, no Darwinist uses it because they think that the entire probability of these pathways are irrelevant (what always matters is it's mere possibility, Dembski has always encountered the logical possibility argument). The latter is what I think is at play here.
C13:
There are many reasons why we go from eliminating chance and concluding design instead of the other way around. I'll name a couple here in the interest of time, and maybe I can get into it a little more when I have more time. One reason is that we can see that the designer can perform what chance and regularity can accomplish.
What we actually see intelligent agents accomplishing is what we cannot see chance accomplish. It is in that very argument that when we see what we actually do not see chance accomplish, we infer what we actually have observed accomplish such a feat, intelligent design. This is not only commonly intuitive, but it is an inference that is used in various sciences:
quote:
If the probability of rain is 5% or less, the statistician says that the chance of rain is so improbable that someone seeded the clouds to make it rain on that day.
source
This also addresses ii and iii which was already addressed in my previous post.
C14: The answer to this criticism, and a rigorous calculation for it, is quite clear in Dembski's writings.
quote:
if the significance level alpha is small enough, there will not be enough items in our background knowledge that are conditionally independant of E and that enable us -- in case E is due to chance --- to recover E within a detachable rejection region R of probability less than alpha...Indeed, the presumption is that someone else used that very same item of background knowledge -- the one used by S to eliminate H -- to bring about E in the first place.
Thus it is the most talked about aspect of Dembski's writings that C12 claims does not exist. It is an objective constraint on a supposedly subjective background knowledge, the Universal Probability Bound, indeed, as Dembski says, it is the smallest significance level we will ever need.
[ 06. May 2003, 17:02: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 06. May 2003 22:20
Nelson, I can't even use Dembski's method to calculate whether a set of coin flips are specified or not. Case in point:
TTTTTHTTTHTHTHTH
(I just pulled these out of my pocket.) There are 16 coins, so the probability of that exact sequence is 2^-16 ~= 1.5*10^-5. Calculating the probability is easy.
This sequence of head-tail flips is a binary number, of course, and the standard conversion is H=1, T=0. Curiously, the number 0000010001010101 it is 1109--day and month of the World Trade Center terrorist attack, arguably the most notable date since I've been pulling coins out of my pocket. Did the World Trade Center explicitly and univocally identify this pattern?
Now what do I do? I had background knowledge, and this knowledge seems to have identified those coin flips pretty directly. There were a few arbitrary choices (British date order--but I was born in Australia; concatencation of digits rather than using the day of year--but that's the only way to make it readable; using it as a date rather than some other common set of 4 digits (time of day?); and so on), but not many. So how do I do this rigorous calculation?
Also, with respect to C12, I would love to do statistical tests for the ability of natural selection to build a complex system but unfortunately I have no idea how to do it in a meaningful way. I can do the calculations Dembski did--they're not particularly difficult. I can do calculations assuming selection and get results that we should expect to have millions of independently evolved flagella, if I'm generous enough with the selection. There is absolutely no argument that selection happens; the critical questions are: how simplified can the system become before it fails to work, and how many other systems can be co-opted to use in the one under consideration?
The answers are that I have no idea. A flagellum structurally would *have* to be at least a few hundred amino acids long, just to get through the membrane and have enough sticking out to wiggle. But maybe almost all of it could be generic unstructured or alpha-helical? On the other hand, maybe the flagellum can exist in only a handful of variants from what it is now, and even single amino-acid changes aren't tolerated.
While Dembski is willing to stick his neck out and make a calculation without all of the relevant facts (actually, I would argue that it is very nearly without any of the relevant facts), I am not willing to do so. It would simply be a waste of my time. By very slightly altering assumptions, I know I can make the calculations come out any way I want, and that's a pretty stupid way to do science.
Anyway, I'd rather not sidetrack you just yet; there are 8 or so criticisms remaining. But I think you have overestimated the rigor and clarity of Dembski's instructions, and underestimated the difficulty of performing calculations that are at all meaningful. Doing calculations is easy; demonstrating that they are meaningful is much more difficult, but of course is absolutely essential.
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 07. May 2003 01:29
Nelson: Selection theory assumes functional intermediates leading all the way to the flagellum. Dembski treated the DCO with this hypothesis in the sample space, and did the calculation, where functional intermediates would lead all the way up to the bacterial flagellum. Instead what the calculation shows is a sea of nonfunctionality, where natural selection becomes sterile (and therefore it is nothing but random assembly).
Actually Dembski did none of the kind, he did not look at the viability of the intermediates, their fitness or anything relevant to the theory of evolution via natural selection. In fact Dembski's argument seems to fall apart now that the data for both RNA and proteins shows that the world of structures is dominated by a few common structures and many uncommon ones and that the common structures can be found throughout sequence space and that these common structures are closely interconnected. That means that for the common structures at least evolutionary pathways are plentiful.
Additionally the work of several researchers has shown how natural selection as well as neutral mutations can both play essential roles in evolution. In fact the simulations show that during evolutionary epochs, fitness improves while during neutral epochs fitness remains the same, a punctuated equilibrium arises when neutral mutations have 'found' a link to another domain and selection takes over again.
Quite fascinating findings which go straight against any argument of protein 'islands' separated by vaste oceans of unhospitable oceans.
If you want to continue your claim that Dembski looked at intermediates and found a sea of nonfunctionality then I encourage you to show us in a few simple steps how Dembski achieved this. Please explain to us how Dembski took into account the effects of neutral mutations as well as selection.
I have read Dembski's works quite carefully and I have seen none of the calculations to which you allude.
As Rex so well states
"But I think you have overestimated the rigor and clarity of Dembski's instructions, and underestimated the difficulty of performing calculations that are at all meaningful. "
as an example of simulations which show how information in the genome can increase through simple processes of mutation and selection see the following figure
quote:
b, The information content at binding sites ( Rsequence) of the organism making the fewest mistakes is plotted against generation number. Selection for organisms making the fewest mistakes was applied from generation 0 to 2000 (top curve, green). The simulation was then reset to the state at 1000 generations and rerun without selection (bottom curve, red). The dashed line shows the information predicted, Rfrequency = 4 bits, given the size of the genome and the number of binding sites.
[ 07. May 2003, 01:37: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 07. May 2003 12:28
I have re-read Dembski's claims and found the following of interest
quote:
What this means is that if a forward chaining search succeeds, it does so as a highly specific and isolated path through genomic space. In that case the step-by-step probabilities moving forward from A_i to A_(i+1) could still be large enough not to overturn my universal probability bound. But absent a successful forward chaining search, there is no reason to think that success is even possible. Successful forward chaining assumes that a sequence like A_1 through A_n and can be made explicit. There is no evidence of this.
The claim that there is no evidence of this first of all ignores the structure of protein space but more importantly shows that Dembski did not do the calculation but merely rejected the hypothesis as "there is no evidence of this".
In fact even the NFL 'calculation' of the flagellum seems to only consider original function:
quote:
In No Free Lunch, I offer a way to try to get a handle on such progressions through what I call perturbation identity and tolerance factors (see section 5.10). The idea is to take a functional system, perturb it, and determine how perturbation affects the probability of retaining function.
But perturbation of a system may not be a very good way of estimating evolutionary probabilities since it can be shown trivially that while forward probabilities may be large, reverse probabilities may be small. This lack of symmetry in connections and pathways between proteins/RNA complicates Dembski's (strawman) calculations for the flagellum.
So when Dembski states that "Miller's task, to vindicate Darwinism in regard to the flagellum, is to exhibit a forward chaining search through genomic space that issues in a genome coding for the flagellum."
But that is incorrect, there are other ways to support a Darwinian hypothesis. In fact as I have argued it is Dembski's task to show that all such pathways are NOT likely. Scientists in no way are constrained by Dembski's design inference and to therefor impose on them the limited approach to scientific investigation via the design inference which in most instances has to deal in probabilities that are impossible to calculate seems illogical.
In case of Schneider and Adami they both have shown forward pathways which increase the information content of the genome, so I would say that in these cases Darwinist have shown what Dembski required. Dembski's answer? That the fitness function somehow front-loads the information but that is exactly what the environment is doing to the genome through selection. Mutations increase the search space and selection retians those mutations which improve fitness. In fact information/complexity is thus transfered from the environment to the genome showing that CSI in an open system can increase via such simple processes as natural selection and mutation
[ 07. May 2003, 12:32: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Iain Strachan
Member
Member # 96
|
posted 07. May 2003 14:57
Rex Kerr:
quote:
Nelson, I can't even use Dembski's method to calculate whether a set of coin flips are specified or not. Case in point:
TTTTTHTTTHTHTHTH
(I just pulled these out of my pocket.) There are 16 coins, so the probability of that exact sequence is 2^-16 ~= 1.5*10^-5. Calculating the probability is easy.
This sequence of head-tail flips is a binary number, of course, and the standard conversion is H=1, T=0. Curiously, the number 0000010001010101 it is 1109--day and month of the World Trade Center terrorist attack, arguably the most notable date since I've been pulling coins out of my pocket. Did the World Trade Center explicitly and univocally identify this pattern?
Sure it specifies it, but it has to be specified and complex in order to be able to say anything about design. Your 1109 specified sequence simply isn't long enough. You have only 16 bits of information & the probability is only around 1e-5. I guess the chances are pretty good that if you'd got any other number you could have found something interesting to say about it. For example any of 911 119 1109 would be equally impressive and that already multiplies up the probability by 3. And what if you had got 12345? or 11111? Or any power of two looks pretty impressive, with one head and the rest tails. The probability of an "interesting" number that you can think up a specifier string for is not that low. But the longer the sequence, the less likely it is to conform to some specifying pattern. Now, if you pulled out a lot more coins and the sequence yielded up not only the dates but the times of the two attacks, and the flight numbers of all four flights, then there is a good case for saying that it's specified and complex. In fact if that lot had turned up, I wouldn't have believed that you'd genuinely pulled the coins out of your pocket.
But with only 16 bits, it's specified but not complex. Dembski uses the universal probability bound of 10^-150, which is around 500 coin tosses, but I guess most of us would raise an eyebrow at 10^-10 (around 33 coin tosses).
One key component is if you can specify the sequence with less data than writing it out in full (via a suitable specification language). For example suppose you had a palindromic sequence of 1000 coin tosses. Then you can specify it by 500 bits and just a little more to say "same in reverse".
Iain
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 07. May 2003 17:33
Note I reply here to Pim as well as some comments on Rex's post and Ian's response.
Exactly Ian. Now, consider such a specified event, coupled with complexity, such as that as the bacterial flagellum. The calculation would not only be a great test of the ability of natural selection, but a great way to infer design, and yet Rex still claims he doesn't know how to do a statistical test with regard to the ability of natural selection. Specification is the perfect test for the ability of natural selection, because natural selection is the king of reductionism, it manages small probabilities by breaking them down to larger probabilities with each step conferring a selective advantage, with the help of random chance. But when random chance plays too big a part, even Darwinists reject such pathways. How could you not use Dembski's calculation for this? I can't find in the peer reviewed literture where a "wiggle" is enough for motility. In fact, it would most likely be useless even if we have a simpler flagellum with a simpler motor made up of a simpler stator/rotor, because the bacteria wouldn't even be able to move. It can hardly do a good job right now due to the "brownian storm", again the issue of IC coupled with minimal function comes into play.
Pim,
Natural selection requires functional intermediates, and without functional intermediates, it's nothing but random chance. Dembski's calculation showed what you can obtain without perturbing function and this is highly relevant with respect to the calculation.
You comment:
quote:
In fact Dembski's argument seems to fall apart now that the data for both RNA and proteins shows that the world of structures is dominated by a few common structures and many uncommon ones and that the common structures can be found throughout sequence space and that these common structures are closely interconnected. That means that for the common structures at least evolutionary pathways are plentiful.
This has absolutely nothing to do with whether a multi-part protein machine can have selectable steps all the way up to the flagellum, which is why Dembski' points to a forward-chaining of Darwinian steps that the calculation is searching for, and couldn't find. Reversing this chaining is exactly how one would conclude irreducible complexity (by removing parts and retaining function). Now you assert that the number of evolutionary pathways leading up to the flagellum is "plentiful" but vague imaginary pathways are very different from testable ones, which is what is required to apply the calculation. Thus, can you show me 10 testable evolutionary pathways that lead up to the flagellum that does not require large amounts of unselectable steps? If the pathways are plentiful this should be easy.
Note, I plan to address the the binding site program at a future date when I am more familiar with both sides of the argument, right now I'm concerned with addressing the last 8 which I may not get to today (since I forgot NFL again when I came to work).
[ 07. May 2003, 18:11: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 07. May 2003 18:42
Ian makes some very good points about finding interesting-looking patterns after the fact. This is related to Erik's point on C15: "if intelligent agents can magically create instances of "specified complexity", then what prevents intelligent agents from magically inventing specifications for any events they take interest in?"
Luckily, we have an answer. Ian wrote:
quote:
One key component is if you can specify the sequence with less data than writing it out in full (via a suitable specification language).
Excellent! Now, where is the suitable specification language (and how do we know it is suitable)?
And if we don't already have one, what do we do?
Nelson seems to continue to overestimate our predictive powers in biology. He asks:
quote:
Can you show me 10 testable evolutionary pathways that lead up to the flagellum that does not require large amounts of unselectable steps?
And the answer for me is no. Then again, I can't even predict that Nae-I could function as a topoisomerase with a single point mutation.
Let's turn the tables: Nelson, can you demonstrate that it is possible to directly test any hypothesized evolutionary pathway? Under what conditions will the testing work, and under which will it fail?
Everything I can think of will fail, quite rapidly, either because our predictive powers aren't good enough, or because our labs can only deal with 10^10-10^12 bacteria, while evolution had 10^24-10^30 to work with to accomplish the same thing.
So we seem to be left with looking at less immediate correlates, such as relatedness of gene families and such.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 07. May 2003 19:20
I managed to respond to the last 8, however, I did so based on memory of what I read in NFL, without the book, so bring your grains of salt.
C15: I couldn't find anything in C15 that I didn't already address.
C16
Specificational resources actually vary in complexity and is relative to what Dembski calls "subjects" and their background knowledge and their computational resources. A specificational resource, by definition, snips away at the probability of the event. It is actually when a specification "matches up" with the event E that we infer intelligent design.
C17
This criticism relates to the one above, and shows how C16 is based on an error. Words here can define precisely what a specification describes, and even it's specificational resources, or what Dembski calls concept levels, without doing an injustice to the complexity of the object being described. For example, the words "bi-directional" , "motor-driven", "propeller", "rotary" all describe bacterial flagella, and thus can be taken as a level 4 concept, as discussed in Logical Underpinnings. If 10^20 specifications of level 4 or less can constitute specificational resources, then we can get a reliable probability for the origin of hte bi-directional rotary motor-driven propeller.
C18
Note, the calculation used for the flagellum, as a discrete combinatorial object, was at first illustrated by doing the calculation for a cake , then the Gettysburg address for Pperturb. I think that these constitute more than "sketches" that can be worked out as much as Dembski's calculation on the bacterial flagellum.
C19
C19 is based on an old argument indicating that the complexity and mechanism of a designer (i.e. an everlasting God) is more improbable then a "missed" chance hypothesis that the EF does not know about. However, this goes back to specified complexity not being purely eliminative, in that it is a consequence of findnig an external pattern, and the common observation that specified complexity stems only from intelligent acts, that I think defeats C19. If advanced technology or any increase in knowledge and/or ability is all that is needed to postulate a better explanation for an event then "we don't know" (especially since science is tentative) then I think that Dembski's methodology is useful as an indicator of intelligent design.
C20: I think C20 is based on a misunderstanding of what displacement referrs to. Displacement refers to the fact that, to use one of Dembski's illustrations, that in order for an EA to successfully get specified complexity, it has to pack levels upon levels of searches (or orders as Dembski calls them), all of which can lead back to an intelligent origin. For example, if you want to get away from a time-dependant fitness function, then you're gonna have to constrain that etc, and such focusing will lead an ID theorist to follow that information trail back to an intelligent origin. The examples that C20 gives are all based on false negatives, whic is ok in the EF, (i.e. not being able to decode a message because it was encoded using the audible range of vampire bats or something outside our background knowledge), nevertheless if the set of specificational resources included the audible range of vampire bats, and the message is successfully decoded, then we may conclude design.
C21: Many false negatives could be given that don't seem to effect the utility of the EF. For example, not having the proper decryption key for an encoded method, thus the supposed random noise is taken at face value. Indeed, there are many false negatives in Biology that provide fruitful research prospects for the design inference (i.e. sub-optimal design). A true negative can also be given, as for example, the hemoglobin case, where the specification is quite small and natural selection can indeed select functional intermediates.
C22 I've already addressed this criticism, in that an objective set of specificational resources can be used in the form of the UBP.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 07. May 2003 19:31
Rex,
The UBP and making specifications after the fact show that specifications are no problem to make, since as Dembski notes, the problem is that we don't know enough in order to invent specifications. Making specifications after the fact is no problem, as for example, finding message from outer space, and finding that they are actually a set of prime numbers.
Rex writes:
quote:
And the answer for me is no. Then again, I can't even predict that Nae-I could function as a topoisomerase with a single point mutation.
Why not? A nuclease is just a reverse topoisomerase. With the NaeI it depends precisely on what is at position 43.
I have shown with my proposed experiment with actual bacteria and with Dembski's calculation how one can either not find functional intermediates between each subsystem or how each subsystem requires unselectable steps. So here we have two-step process of fully testing a Darwinian pathway.
Now you do realize that there was only 100 million years between when life was habitable and when the first life forms appeared in the fossil record. How many bacteria was evolution left to work with really? Also, if the hypothesis that bacteria evolved flagella through natural selection and random mutation depends on a certain number of bacteria, then we are no longer speaking of a sochastic process, especially given all the hoopla about simpler and alternative functions.
Also, statistical tests don't need to gather large amounts of bacteria. Even if your objection was valid, we can look at the complexity of each machine and their subsystems, and find out if natural selection smuggles in too much random chance.
[ 07. May 2003, 19:39: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
|
Printer-friendly view of this topic