|
Author
|
Topic: The Evolution of the Bacterial Flagellum
|
Nel
Member
Member # 614
|
posted 09. December 2003 01:11
Ged,
Can you provide an example, where there was no prior knowledge of an act (i.e. the first time an archaeologist found a tool from the past), and this archaeologist, after using statistics to calculate a probability of the designer of this tool to act to make this tool, concludes design? (this doesn't have to be an archaeologist, any field with like parameters will do).
[ 09. December 2003, 01:13: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 09. December 2003 01:12
quote:
(I hope that those who collect police statistics are reading, and now know that their efforts are irrational!)
Nelson: This is after-the-fact information and has nothing to do with , as you call it, "probability of space ships".
It has everything to do with profiling, predicting crime behavior just like statistical polls provide probability measures. What about insurance companies who base their policy ratings on probability measures.
What do they know that makes these data so relevant....?
Or talk to Amazon.com whose datamining techniques are notorious.
As far as probabilities and archeology
link
quote:
In the field of psychiatry, the correct diagnosis of clusters of symptoms such as paranoia, schizophrenia, etc. is essential for successful therapy. In archeology, researchers have attempted to establish taxonomies of stone tools, funeral objects, etc. by applying cluster analytic techniques. In general, whenever one needs to classify a "mountain" of information into manageable meaningful piles, cluster analysis is of great utility.
Or this one which looks as Bayesian versus Fisherian approaches
quote:
Bayesian statistical methods have become powerful and efficient tools in diverse areas of statistics, particularly medicine, social studies, archeology, etc. The Bayesian approach to statistics provides a unified framework for optimally combining information from multiple sources and for incorporating previous experimental results and/or expert opinion into current experimentation and modeling. This results in simpler, highly efficient experimental designs and statistical analyses.
[ 09. December 2003, 01:26: Message edited by: Pim van Meurs ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 09. December 2003 01:23
Alonso said: quote:
Can you provide an example, where there was no prior knowledge of an act (i.e. the first time an archaeologist found a tool from the past), and this archaeologist, after using statistics to calculate a probability of the designer of this tool to act to make this tool, concludes design? (this doesn't have to be an archaeologist, any field with like parameters will do).
Quoting myself previously: quote:
...arguing this to the bitter end, rather than changing the subject.
I find Alonso's change of subject interesting itself, though rather surprising that Alonso cannot answer the questions posed by myself and others about the relationship of statistics and probability.
But this gets back to the analysis of wide ranges of probabilities given much earlier. ID procedure, whether based on eliminative or based on comparative methods are in fact fairly reliable when the joint probability of the "designer" causing the event in question are much higher than the joint probability of "missing distributions" or natural causes that have not been considered. I never said that ID inference procedures like the "explanatory filter" are never reliable--they are in fact reliable in certain circumstances.
In the case of archaeology, we know that people existed and were likely to produce tools. We do not have the conditions which Dr. Debmski identified in the urn example in which the two cases being compared are both unlikely. Rather one case being compared is the case of people creating the tools, and that is reasonably high. The other case to compare is natural forces, such as chiping of rocks. Is the object just a chipped rock, or a carefully chipped rock comparing to a particular pattern seen as useful as tools? These are reasonable questions and do not compare to the urn example.
Furthermore we have real theory here. The tool was created by humans for a particular purpose of cutting meat and working hides, etc.! This helps to reduce the effects of coincidence, because we are not just pulling the sharp edge of a stone out of thin air as a "design" cause, we are relating it to situational considerations involving many aspects and relationships.
But to actually answer Alonso's question, this is where the wide ranges of probabilities become relevant. If we have known theory of people chipping rocks into certain shapes (especially distributions of similar shapes across a larger region, for example), then we have a vaguely quantified value for that shape as joint probability of designer performing act of creating that shape. One does not have to perform the mathematics of the Bayesian inference for its results to be known. The bases of such calculatins can be done with intuitively judged values--this is perfectly reasonable and rational method. It ammounts to the same thing, and either an eliminative or comparative analysis is being made, and that analysis compares somewhat to the EF. No problem!
Earlier post:
quote:
This is after-the-fact information and has nothing to do with , as you call it, "probability of space ships".
I also have never said that the probability of "space ships" was low, for example. If one believes that the probability of aliens beaming objects into and out of one's work space is very high, then based on that assumption it might be quite rational to assume that my continued loosing of pencils is caused by alien space ships. If aliens were in fact beaming pencils out of drawers in the area at a high frequency, that might be a superior explanation to my randomly and unthinkingly taking them elsewhere.
I was quickly loosing candies out of an open dish but no one was eating them--and later found evidence of a mouse and the wrappers collected in a corner. I had eliminated natural processes causing the candies to evaporate. I had also eliminated all the human interventions. So this is very much like Dr. Dembski's urn problem or the eliminative method--this shows that a third alternative should be considered! But is the "mouse" an "intelligent agent"?
If the mouse is an intelligent agent and mouse action to steal food is "design", the are you saying that one cannot estimate the probabilty of mouse eating food out of one's house is much closer to 1.0 than it is to 10^-150? Is this irrational to make such an estimate of probability--because the "mouse" is an "intelligent agency"? But wait, if the mouse is not an intelligent agency, then the unexplained loss of candies was not ID! So the ID inference failed to be reliable! Which is it, Alonso, can I estimate probabilities of mouse ID, or is the ID inference unreliable?
[ 09. December 2003, 01:53: Message edited by: gedanken ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 09. December 2003 01:59
Gedanken: I think that it is rationality itself, combined with a knowledge of the logic of mathematics, that provides this result. Not something that is a special limitation of "science", appart from rational thought itself.
Yet once again, the truth of the flagellum’s design (for example) does NOT entail we’d have access to a base of knowledge about its designers. This is then another demonstration that science, rationality, the logic of mathematics, etc., does not provide access to all truths.
Perhaps it is then time to address the question which is at the base of ID, as put by Dembski. According to Dembski, ID begins with this question:
Can objects, even if nothing is known about how they arose, exhibit features that reliably signal the action of an intelligent cause?
Your answer seems to be “No!” That is, nothing about the object itself can signal design. Only the mechanism of its becoming can possibly signal design, right?
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 09. December 2003 02:14
Mike: Yet once again, the truth of the flagellum’s design (for example) does NOT entail we’d have access to a base of knowledge about its designers. This is then another demonstration that science, rationality, the logic of mathematics, etc., does not provide access to all truths.
Of course not, science may have to conclude based on the lack of data, that 'we don't know'. Again Mike seems to demonstrate that 'truths' are limited by the supporting evidence of lack thereof.
Mike: Can objects, even if nothing is known about how they arose, exhibit features that reliably signal the action of an intelligent cause? Your answer seems to be “No!” That is, nothing about the object itself can signal design. Only the mechanism of its becoming can possibly signal design, right?
How they arose is but one of the ways to support a design inference with much needed information about motives, means, abilities etc.And design is not really comparable to 'intelligent cause'. While intelligent cause may include design, design can incorporate more than just intelligent cause.
Perhaps a limitation of science but science does include the reality of admission of 'just not enough data (yet?)' as an explanation.
It's not that intelligent cause is not an impossibility, it's that intelligent cause needs to be constrained in order for it to have some scientific relevance.
I believe that Gedanken has made it clear what is needed for a reliable design inference.
[ 09. December 2003, 02:22: Message edited by: Pim van Meurs ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 09. December 2003 09:06
quote:
Yet once again, the truth of the flagellum’s design (for example) does NOT entail we’d have access to a base of knowledge about its designers. This is then another demonstration that science, rationality, the logic of mathematics, etc., does not provide access to all truths.
Let's forget science. But let's remember the basic rationality that has been learned in the last few hundreds of years, in specific areas like mathematics.
For example, are we to forget Bayes theorem? If we forget it, but agree to basic rationality upon which mathematics is based, I can remind us again by presenting a mathematical proof of Bayes theorem again (I don't have to name it "Bayes", we can call it the "posterior probability theorem" if we wish, or the "after the calculation theorem", whatever.)
Now consider whether the apparent match of the concept from human history of the motor, and the mechanical-appearing aspect of the flagellum, were matched by coincidence rather than because there was some causal relationship.
Let's not forget Dr. Dembski's "ProbRes" factors, and why they are present in his Generic Chance Elimination procedure. I advise study of No Free Lunch, the sections on "ProbRes" factors.
Namely, study both the "SpecRes" factor N, and "ReplRes" factor M, and why they are given as important in the GCE procedure (a version of the "explanatory filter" given in NFL chapter two). These factors each deal with "inflation"-like problems which must be accounted for to avoid finding a match by coincidence rather than because there is any actual basis in "design" for the match of the aspect picked from human experience, and the aspect in nature (as in the rotating element of flagellum). These factors are present to avoid the problem like the search for Bible codes in which one establishes a match of some algorithmic transformation of the text of the bible and some language statement--but could do so by chance, that there was a coincidence of a match rather than a rational basis.
In a simplified description in NFL page 79, consider a number of targets for arrows N, and that you have M arrows in your quiver. If the chance of a single arrow hitting the target is p, then there is a significance to the mathematical expression M*N*p. That significance is that the probability of hitting the target by "chance" (p), is amplified by the "ProbRes" factors M*N. (Dr. Dembski chooses a base reliability for the GCE of 50%, in other words the base reliability of the explanatory filter is 50% of the time it will not generate false positives due to coincidence matches--and thereupon derives the cutoff alpha by the equation 0.5 > M*N*alpha. Another way to view it is that the probability of "design", controlling for coincidence, had just tipped toward "design" over natural causes when alpha is set at least as low as 0.5/(M*N))
What are the "ReplRes" and "SpecRes" factors worth considering in examining the pattern match of the motor from human experience, and the rotating element of the flagellum? Remember that the rotating element of the flagellum does not have steel or brass bearings, and that it is not powered by a magnetic field from electrical coil windings. So the match is more distant--how many ways could such a match be chosen due to allowing for vague similarities rather than exact match to the properties of the electric motor? And over how many different biological forms could such a search be made for a match, which must be controlled for as possible coincidental matches?
Remember that one cannot reasonably just assume a truth to "design" of the flagellum--at least not within reason without saying that one has simply assumed this. The question is whether we have simply picked a chance match at random, like seeing a face in the clouds. What I am getting at is why we even need to consider rationality here.
If we don't need to consider rationality and its limitations, then we based the concept of design of the flagellum upon some other aspect of humanity, a religious belief possibly, not a rational evaluation?
[ 09. December 2003, 09:24: Message edited by: gedanken ]
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 09. December 2003 12:16
What is exceedingly important about gedanken's last posting is that it makes it clear that choices about auxiliary information govern the probabilities that are calculated in Dembski's procedure. Those choices, as gedanken has shown, inform the assignment of numbers to M and N, and thereby control the probability ("complexity" in Dembski's usage) that one calculates for the object in question. But there is no principled way to make those choices: they're made in the light of the current knowledge (and whim) of the investigator. (I'm beginning to read gedanken's analysis to imply that in making those choices, Dembski smuggles in assumptions about the design process and the designer.)
It follows that the answer to Dembski's question, quoted by Mike Gene, is no, one cannot determine from properties of an object itself whether given features reliably signal the action of an intelligent cause. Nothing about the object itself, by itself, reliably signals design.
'Designed by an intelligent agency' is not a property of an object. It is a claim about the history of production of an object. Interpreting the putative traces of an intelligent designer's activity left on an object by that history, whether the interpretation is formally probabilistic or purely subjective, requires invoking auxiliary knowledge and making choices about matters outside the object itself, and there is no formalism for determining how those choices are to be made. Hence there is no mechanical procedure for inferring design.
RBH
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 09. December 2003 12:39
But I see no reason in principle why such a formalism could not be developed. The Explanatory Filter would seem to be an attempt, albeit an imperfect one.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 09. December 2003 19:30
Ged:
quote:
In the case of archaeology, we know that people existed and were likely to produce tools.
In fact, we do not know that people existed and were likely to produce tools. We simply assume this or infer this from evidence.
Nonetheless, this illustrates the very point I wanted to make when I asked my archaeology question. With polls, police statistics, and even these examples of archaeology we have prior experiences of the intelligent designers that we want to make the cause for such actions. These statistics and/or assumptions are based on after-the-fact observations of the designer (their methods and motivation) that can explain the event in question.
Consider for example giving the probability for the event for a presidential election. According to most staticians, reasons for voter behavior fall under, for example, whether economic growth is good. How do we know this? Well we look at past observations. For example, in 1932 Hoover was the incumbent, but he only got 40% of the vote, the economic growth rate for that year was horrible. On the other hand, in 1936 Franklin Roosevelt (the incumbent) got 62% of the vote, that year the economic growth was very good. Thus economic growth rate is a good factor in predicting who will win an election.
The question is, do we have experience designing life forms and the complex machines that we find? No. Does that mean that the EF is unusable or that we can never hope to detect design? Not in my opinion. First, the EF is an eliminative filter. It eliminates chance and regularity explanations. Does it eliminate the need to know why the designer did it or how? No, it simply leaves that as a seperate question. As to how we would know such a thing? Here we go back to what Mike said earlier:
quote:
The best way to understand design is to do it. We’ve learned much about aerodynamics as a consequence of designing planes. I recall watching a documentary about some engineer who sold the military on plans for a type of flying saucer. Stuff worked great on the blackboard. But when they actually designed it, they couldn’t get it to fly.
Attempts to engineer homeostatic, self-perpetuating cellular systems would help us get a much more rigorous handle on the design principles involved. And that we would not simply borrow from LAWKI (life as we know it) would give us greater confidence we were looking at general principles defined by the design problem.
Here not only have a possible motive, but how it was possibly done. I can envision arguments among design theorists to the effect of (the designer probably didn't do it that way because of X Y Z). As to what are the probabilities that someone was around for the job? There are no laws or processes that would tell us there could not possibly be anyone around for the design.
Does the design community not take motive seriously? I think it does (and here I am only expressing my own opinion). Julie Thomas, who participated on T.O. stated:
quote:
Since intelligent design entails intent, each of the four
systems can be viewed as serving the intended existence of
these bacterial cells (this is especially true of DNA and its
replication, along with the bacterial cell wall). But what intent
could possibly exist behind the design of the cell? The working
hypothesis that I have proposed is that of terraforming. Namely,
bacteria-like cells were deposited on the abiotic earth with the
intent of transforming the environment of the early earth into
one that would become bio-friendly.
In summary, I think that Ged's concerns are at best, backwards.
Ged:
quote:
Furthermore we have real theory here. The tool was created by humans for a particular purpose of cutting meat and working hides, etc.! This helps to reduce the effects of coincidence, because we are not just pulling the sharp edge of a stone out of thin air as a "design" cause, we are relating it to situational considerations involving many aspects and relationships.
But this goes back to something that Sober stated:
quote:
To infer watchmaker from watch, you needn't know exactly what the watchmakers had in mind; indeed, you don't even have to know that the watch is a device for measuring time. Archaeologists sometimes unearth tools of unknown function, but still reasonably draw the inference that these things are in fact, tools.
Ged:
quote:
If the mouse is an intelligent agent and mouse action to steal food is "design", the are you saying that one cannot estimate the probabilty of mouse eating food out of one's house is much closer to 1.0 than it is to 10^-150? Is this irrational to make such an estimate of probability--because the "mouse" is an "intelligent agency"? But wait, if the mouse is not an intelligent agency, then the unexplained loss of candies was not ID! So the ID inference failed to be reliable! Which is it, Alonso, can I estimate probabilities of mouse ID, or is the ID inference unreliable?
No I think that the mouse is an intelligent agent.
I have to keep comming back to the fact that designers don't act like mechanistic processes (yes they do things that are sometimes predictable, but this is not always uniform, a designer can select steps that natural selection would not, we would never expect natural selection to act so unpredictably so as to select something that is functionless, in fact, this is what irreducible complexity is all about ). Even if the probability you speak of is something akin to criminal behavior, and this is what you want applied to the hypothesized designer, the fact that ID does not involve this is not because of any irrationality, it is because of a current limitation of science (the fact that we don't have that background knowledge!).
As Dembski himself states, the EF does not rule out such questions, it simply makes it a seperate one. But the question is, one can only hypothesize whether a mouse would go for the food, or if it would not, based on past events. Surely, the probability that a mouse was the one taking the cheese out of the mousetrap rather than a bird from Africa can possibly be calculated. But what does that say of the actual event? If we could not possibly calculate the probability of a certain designer and it's motives, does that take away at all from the fact that it is simply too improbable to be a natural occurance? That is the question. One can base their inferences of design action on what is likely to be around at the moment. But what if one has no prior experiences of mice eating cheese from mousetraps? What if every mouse was dead after having attempted to take food from the trap? We would have expected the mouse to be dead, if it was in fact a mouse. Especially if the mouse trap is an elaborate one. What if it was in fact, a clever (and hungry) human? Now we see the approach to all this constantly making false inferences as to who or what was the motivation behind the theft, as a mouse becomes an unlikely candidate (based on statistics). But because intelligent agents are unpredictable, dynamic entities (and not uniform laws/processes) there is no reason to rule out a clever mouse.
Ged:
quote:
Remember that the rotating element of the flagellum does not have steel or brass bearings, and that it is not powered by a magnetic field from electrical coil windings. So the match is more distant--how many ways could such a match be chosen due to allowing for vague similarities rather than exact match to the properties of the electric motor? And over how many different biological forms could such a search be made for a match, which must be controlled for as possible coincidental matches?
I think that what the parts are made up of is less relevant than the structure and function of the system (speaking in terms of stators, rotors,things are driven by a certain revolution per second, at a certain horsepower,etc). After all, we make molecular brakes, ratchets, antennas, etc, and these things have undeniable human characteristics and functions.
Coincidental matches go right at the heart of the complexity issue. A face in the clouds is far less complex than the face on Mount Rushmore. Both are specified, but we usually infer design from the latter.
Note, since this has gone way off topic from Nic's article, any further replies will be done at another, more relevant thread (i.e. the motive thread or a new one).
[ 09. December 2003, 20:27: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 09. December 2003 20:37
Pim: Of course not, science may have to conclude based on the lack of data, that 'we don't know'.
Sure. In this case, we could say that we don’t know if the flagellum was designed or if it arose from natural causes.
Again Mike seems to demonstrate that 'truths' are limited by the supporting evidence of lack thereof.
What happened is not limited to supporting evidence. Only our beliefs about what happened are limited in this way.
How they arose is but one of the ways to support a design inference with much needed information about motives, means, abilities etc.
But one of the ways? What other ways did you have in mind?
-----------
Gedanken: Remember that one cannot reasonably just assume a truth to "design" of the flagellum--at least not within reason without saying that one has simply assumed this.
I am not saying that it is true that the flagellum is designed. I’m pointing out that if it is true the flagellum was designed, this does not entail that we’d have access to an independent base of knowledge about the designers, the very thing that is necessary (according to some) to infer design. If science/rationality needs such a base of knowledge to infer the flagellum’s design, then we clearly see its limitations for accessing truths about reality. Put simply, they are incapable of detecting design, thus passing judgment on design. Yet I commonly encounter critics of design who go much further than this and make a case against design.
The question is whether we have simply picked a chance match at random, like seeing a face in the clouds.
Indeed. But go back to your flagellum example. What if the flagellum had steel bearings and was powered by a magnetic field from electrical coil windings? I see a rabbit hole. ![[Wink]](wink.gif)
A further problem is this eager willingness to abandon a design inference because it could be a “chance match at random.” There is no doubt in my mind that detecting design without the luxury of an independent base of knowledge about the designers is a difficult task indeed. I have found that many are itching to abandon the effort at the first sign of difficulty. Given my high tolerance for ambiguity, I am not bothered by the things that bother others. I just keep on probing, pondering, testing.... And I am not discouraged.
What I am getting at is why we even need to consider rationality here.
If we don't need to consider rationality and its limitations, then we based the concept of design of the flagellum upon some other aspect of humanity, a religious belief possibly, not a rational evaluation?
Sure. I never said anything about a design inference needing to be rooted in rationality/science. I just have a sincere hunch that remains even when I understand the non-teleological perspective and have no psychological need/motive to declare design. A splinter in the brain, as some movie character once said. And I find it quite intellectually stimulating to look for ways to follow up on this hunch.
Anyway, as Nelson points out, this discussion as far off the topic of the OP.
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 09. December 2003 21:28
Alonso said:
quote:
Coincidental matches go right at the heart of the complexity issue. A face in the clouds is far less complex than the face on Mount Rushmore.
First, I had not intended to argue about the results that come from my “motive” thread here, rather to discuss their implication to the flagellum. (Further discussion of the essence of this would be fairly relegated to the misnamed “motive” thread.) This is the primary issue here, Dr. Dembski brings up Nic’s article and its implication specifically for the EF and flagellum are at issue.
That is why I am reminding participants about Dr. Dembski’s use of the “ProbRes” factors. (Re-read my last post for issues, I don’t want to repeat them in detail.)
quote:
I think that what the parts are made up of is less relevant than the structure and function of the system (speaking in terms of stators, rotors,things are driven by a certain revolution per second, at a certain horsepower,etc). After all, we make molecular brakes, ratchets, antennas, etc, and these things have undeniable human characteristics and functions.
This gets precisely to the issue of “SpecRes” and “ReplRes” factors N and M. How loosely are we taking those features from humanly described objects? The more loosely we take those features, the greater the “SpecRes” factor used in the pattern match to organism’s features. And how many different organisms do we search over to find a match to any of those N different descriptions of features, combined with how many different subsets of the features within the organisms? This maps to the “ReplRes” factor M.
You see the issue is precisely whether the flagellum was chosen out of all the M * N potential matching combinations simply because the stretch of loose association (scaled as part of M) and choice among all possible organism sub-part descriptions (scaled as part of N) happened thus to find a match by coincidence—just like the coincidence of all the possible algorithms for sub-selecting codes from books times all the passages that could be chosen for such processing. We see precisely why the EF must use such a low cutoff factor alpha.
So what is the significance of Nic’s article? The significance is that is provides connection to research for potential paths by which the flagellum developed by natural causes, it provides potential pathways that increase our knowledge of higher probability of chance construction within the environment. The significance is in essence to reduce the “complexity” of the flagellum using that odd meaning of the term.
This is my response to Mike’s point as well. What I am suggesting is that the amplification of probability pathways by “ProbRes” factors M*N occurs with all chances for production of the effect, whether well and completely described by theory or only partially or even unknown. These make the EF unreliable when we do not have a much higher joint prior probability of the designer cause of the effect than the joint prior probability of the effect by natural non-intelligent processes.
The reliability of the EF in a case like this is precisely dependent upon the reliability of the information used to calculate the probability of the natural non-intelligent cause used. In other words, without acceptance of a high joint prior probability of designer action to cause the effect in the target class used, the EF reduces to the “argument from ignorance”.
If one believes in high prior joint probability of designer action, then one has brought in an element for comparative rather than eliminative approach. The acceptance of high joint prior probability of designer action is one leg of the comparison.
But the key element from the “motive” thread is the recognition that the issue is not one of whether one should use eliminative, or should use comparative methods. Both are unreliable of the two choices being compared are of low probability. And thus the reliability of such methods as the EF, even if properly used, depends highly on what is taken from belief about the joint probability of designer cause of the effect class in question. Since this belief is highly variable, it is the element that makes the EF reliability something that cannot be analyzed here. (Alonso has a point in part, but he and Mike miss the aspect that lack of existence of a designer with access to the events confers low joint probability of designer cause of the event in the class, whether or not one has chosen to state motive aspects as possible among the speculative designers.)
The mathematics still shows the EF to be unreliable in this kind of case, highly dependent upon philosophical/religious/belief considerations of joint prior probability for designer action for any sort of resolution to the reliability issue. If designers in fact act thusly in extremely low frequency (e.g. due to their absence or lack of access), then the EF is highly unreliable in terms of false positives and reduces completely to the form of an “argument from ignorance”.
And Nic’s article shows increasing probability of construction of the flagellum, much higher than say an alpha cutoff of 10^-150 for example. Dr. Dembski’s correct attempts (in part) to control for coincidence demand a low cutoff factor alpha to even achieve a 50% reliability of the EF. But those factors M*N also amplify any unknown distributions, not just the known distributions! Nic’s article helps move the probability of the flagellum more from the unknown to the known in descriptive terms.
--
quote:
But this goes back to something that Sober stated:
quote:
To infer watchmaker from watch, you needn't know exactly what the watchmakers had in mind; indeed, you don't even have to know that the watch is a device for measuring time. Archaeologists sometimes unearth tools of unknown function, but still reasonably draw the inference that these things are in fact, tools.
Ah, but “theory” here is that humans existed, constructing tools of the general class of the object identified. It is indeed not necessary to know the “function”, rather theory is of a different aspect of development. And note that theory only reduces M*N factors by the degree that the theory focuses the possible elements being considered, it is highly “theory” dependent. That is why theory must be more specific to be useful. One can’t as theory simply say my theory is that there were space aliens making changes, or the like. One has no way to pin that down and check it by observation—it suffers from the same coincidence factors.
Mike said:
quote:
Indeed. But go back to your flagellum example. What if the flagellum had steel bearings and was powered by a magnetic field from electrical coil windings? I see a rabbit hole.
But it doesn’t.
I’m not saying the EF is always unreliable. I’m saying one can describe when it is reliable and when it is not. But when it is reliable, then comparative methods are also reliable. Either only become reliable when one has theory to reduce effects of coincidence. And specific circumstances would be important—would this be a find that only occurred after the industrial age? Even if evidence pointed against this, one would suspect planting of evidence, and “theory” could include how such evidence could have been planted.
The real problem with the eliminative approach is not that it eliminates—rather the problem is that after eliminating it does not give the positive answer to another claim. Are there more than two possibilities?
One needs to do the calculation for the flagellum! (And in light of Nic’s article.) But then all that is eliminated is the calculation (if of very low probability).
[ 09. December 2003, 21:47: Message edited by: gedanken ]
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 10. December 2003 13:20
Pim: Again Mike seems to demonstrate that 'truths' are limited by the supporting evidence of lack thereof.
Mike: What happened is not limited to supporting evidence. Only our beliefs about what happened are limited in this way.
Exactly what I said. But an eliminative approach or front loading approach does not seem to help resolve this issue of limitations in our understandings.
Mike: Put simply, they are incapable of detecting design, thus passing judgment on design. Yet I commonly encounter critics of design who go much further than this and make a case against design.
Perhaps these critics are not arguing against design as much as against the design inference approach? After all the design inference ala Dembski would indeed infer design where it should have infered at most, we don't know.
Would Mike agree with those critics who point out that a design inference in case of such 'design' may be unwarranted? Wilkins and Elsberry have made a good argument to that extent?
Pim: How they arose is but one of the ways to support a design inference with much needed information about motives, means, abilities etc.
Mike: But one of the ways? What other ways did you have in mind?
motives means abilities, anything to limit the extent of the 'intelligent design' idea to a testable scenario
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 03. January 2004 18:03
I consider Nic’s claim that the F0F1 ATP synthetase is homologous to the type III secretory machinery here
IP: Logged
|
|
|
Printer-friendly view of this topic