|
Author
|
Topic: Design Inference Game II
|
Cre8ionist
Member
Member # 140
|
posted 31. January 2004 09:35
As with most of the other examples, this new picture doesn't seem to possess the characteristics which would make it a candidate for the filter.
quote:
Given the extreme "no false positives" claim I find it telling that there is such resistance to applying the filter to non-designed phenomena.
brauer, I understand your point, but is it resistance, or ignorance or a combination of these + other reasons (mentioned below)? I think in most cases it's the latter. Dembski's not a physicist, nor a biologist, otherwise I expect he'd be running more through the filter than he is. However, there are enough people out there, IDers and NDers alike, who could pick up the slack a little. I don't fault Dembski, he's trying. The point he's working on, that of SC being the designsignature, could potentially revolutionize science, so I think it's worth the effort, by everyone, and the fact that we're still discussing it vigorously means to me that it isn't at all settled. When either the vindication of Dembski, or his refutation comes, I think we'll all recognize it. What I won't do is count him out because of someone's sleight of hand.
RBH "Nor can I see an accurate method. But those kinds of stimuli are precisely the stimuli of
interest! All these examples of letters floating in soup and dots and dashes on a stone are
contrived, and do not veridically model the phenomena about which we really want to
know, the biological phenomena that ID repeatedly claims are designed. Sure, if I saw
"Manufactured by Og Nanobiotics, Inc., on Beta Centaurus. Patent applied for October 22,
4004 BCE" encoded in a cryptogram embedded in a string of mitochondrial DNA, then I'd
say "Hey! There it is! Wowie Zowie!!" But one hasn't yet seen that, and it is the only kind
of example of reliable design detection we are repeatedly offered."
Italics mine
Both those examples are candidates for the filter, and both wouldn't return positive as Pim rightly points out.
As for the biological phenomena. They are well suited for testing with the filter and I expect this will happen with some examples, sooner rather than later, but we must be patient. I would guess that part of the hesitation by the leaders in the ID movement (not just Dembski) is due to the potential for mistakes being made in the assessment of the variables. Namely S and C. Extra care has to be taken. As was mentioned by brauer, there is a no false positives statement necessarily connected to the filter. Therefore, 1 strike and your likely out, even if it's due to some operator error or more likely, as with the flagellum, a distant homologue is found. This more likely response to a positive doesn't necessarily count as a false positive, but to evolutionists it carries a lot of weight.
Let me say something about this second response as it is likely to occur. One of the only areas I can speak on with any real authority is programming. Fortunately, there is a lot of similarity between programming languages, and many things that hold true for one hold true for others. Homologous functions, or methods or sections of code are a good example.
If I were to put up two similar sections of java code:
buildConstraints(constraints, 2,0,1,1,33,5,0,0);
constraints.anchor = GridBagConstraints.EAST;
response4 = new Button(" Quit ");
constraints.fill = GridBagConstraints.NONE;
gridbag.setConstraints(response4, constraints);
add(response4);
constraints.anchor = GridBagConstraints.CENTER;
and
buildConstraints(constraints, 0,4,1,1,0,20,0,0);
constraints.anchor = GridBagConstraints.SOUTHWEST;
startgame = new Button(" Start ");
constraints.fill = GridBagConstraints.NONE;
gridbag.setConstraints(startgame, constraints);
add(startgame);
startgame.setEnabled(false);
They could be called homologous using the same criteria that is used with say proteins, yet they both were written separately and both
belong to separate but "homologous programs." Also, both return positives from the filter. So, not only would the finding of a homologue, not signal a false positive, it wouldn't surprise me at all. This type of slight modification is normal in programming.
Anyway, back to RBH:
I can't agree with you that the only kind of "reliable design detection " we are offered is that kind, it's been noted that the flagellum has undergone testing and produced a positive result. And it's well known that life is full of CSI so don't give up on this worthless filter yet.
Pim, maybe I spoke too soon when I said I haven't seen Frances around? Nobody has tortured my simple handle (Cre8) as much as Frances, until you. I have to wonder why it bothers you so? There're a couple of funny commercials making the rounds at movie theaters about some self-made Internet hip-hopper named Brian who insists on being called Delux247, but his neighbor refuses to use the handle, preferring to call him Brian, making for some really comical scenes. But you feel free to continue in your torturing of my handle (Crea8), I really don't mind. (Insert Howard Dean scream here) ![[Big Grin]](biggrin.gif)
quote:
The cereal bowl example however does not seem to use the Design Filter to reach a
conclusion of 'design'. So I am not sure why we are moving back and forth between a
variety of examples, none of them too relevant to the issue at hand
Because, if you have a candidate for the filter, it does not have to reach a conclusion of design to be evaluated.
In fact, that would completely negate the need for the filter wouldn't it?
As to the flagellum example, Dembski goes out of his way to be fair with his numbers and it really doesn't take much to see that the bacterial flagellum exhibits SC.
Hopefully, we can use this forum to springboard into examples which satisfy both sides and see the filter used on some "non-trivial" examples.........Cre8
[ 04. February 2004, 08:23: Message edited by: Cre8ionist ]
IP: Logged
|
|
Matthew J. Brauer
Member
Member # 819
|
posted 31. January 2004 10:17
Cre8,
First, I wish you'd stop talking about "shell games" and "sleight of hand" in this context. Some of us are trying to figure out the answers to real objections to the operation of the filter. If every challenge is taken personally, the cooperative exploration process won't work.
Second, if this design concept is to be scientific, it needs to provide some objective criteria for assigning its terms. If something is "specified", it should be apparent to everyone why that is so. We should all be able to follow all of the steps used to determine that it is in fact specified.
Finally, if the filter is not applicable to some problems, there needs to be an objective means for determining this. "It would not seem applicable" is not good enough.
My frustration in all of this is that not only the amateurs (meant in the positive sense) but the professionals (Dembski, Richards) seem to be unable or unwilling to provide the step-by-step description of what needs to be (if it is science) an objective process.
I don't understand some of the steps (why are solar eclipses not amenable to the analysis, for instance?) so it's not up to me to demonstrate the utility of the method. Someone who understands the method needs to demonstrate it.
I've done a lot of reading of the primary literature on the design inference. I have yet to see a single compelling example worked through thoroughly.
(I'm not alone in this criticism. Del Ratzsch -- no ID critic! -- has made similar objections that remain unanswered by Dembski.)
[typo edits]
[ 31. January 2004, 22:34: Message edited by: Matthew J. Brauer ]
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 31. January 2004 16:38
Cre8: As to the flagellum example, Dembski goes out of his way to be fair with his numbers and it really doesn't take much to see that the bacterial flagellum exhibits SC.
I am not sure what you mean by 'fair with his numbers'? What I am pointing out is that Dembski's calculation may have little relevance to biological concepts. Does the bacterial flagellum exhibit SC? Perhaps but if SC is not a reliable indicator of intelligent design or in other words, if, as the data suggest, natural mechanisms can create SC then we have to conclude that the appearance of SC in the flagellum based on a non-biological scenario is of no relevance to inferring intelligent design.
Cre8 talking his screenname: I have to wonder why it bothers you so?
What makes you thinks it bothers me? I have no issue with your screenname. If my confusion as to the exact spelling is disturbing to you then I apologize. I am focusing on the scientific arguments and hope that we can thus focus on these concepts rather than on this minor issue?
So lets return to the issues which are basically as Bauer explains:
quote:
I've done a lot of reading of the primary literature on the design inference. I have yet to see a single compelling example worked through thoroughly.
Combine this with the issues of false positives (Del Ratzsch, Perakh), the non biology relevant calculations (Dembski for the flagellum), and the various other problems with the filter and the observation that the filter has yet to be shown to be useful to any non-trivial example may indicate that while the filter in 'principle' could infer design, in actual applications it fails to be useful to infer new design (Del Ratzsch). There is nothing wrong with that because as Del points out Dembski's argument is not necessarily about the practical nature as much as the logical possibility of infering design.
Now that I have corrected crea8 to read cre8 I hope we can refocus?
[ 31. January 2004, 16:48: Message edited by: Pim van Meurs ]
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 01. February 2004 11:36
Cre8:
"Manufactured by Og Nanobiotics, Inc., on Beta Centaurus. Patent applied for October 22, 4004 BCE" comprises 96 characters. With an alphabet of approximately 40 total characters (including spaces, numerals and punctuation marks) for a "chance" probability of 40^96. That's plenty improbable to pass the UPB of 1 in 10^150.
Cre8 wrote quote:
And it's well known that life is full of CSI so don't give up on this worthless filter yet.
Well no, actually it's not "well known." It's not "known" at all. Complex Specified Information is composed of two numbers, Specification and Complexity. Specification (S) is 1 or 0; a specification either exists or doesn't exist. Complexity (C) is a probability with a value between 0 and 1. "Information" is a null word - it plays no role in the formal calculation/measurement of CSI.
So CSI = S*C
Can one point to published values of CSI for a range of objects that support the assertion "And it's well known that life is full of CSI ...". As far as I know there are none.
As I reread this in draft, it struck me to ask whether anything in the universe is not "specified." Can anyone name an object or process in nature for which one cannot identify a specification? Does S ever take the value 0?
RBH
IP: Logged
|
|
Erik
Member
Member # 160
|
posted 01. February 2004 12:10
Writing CSI = S * C is potentially a bit misleading, because C depends on the specification (what's the value of C when S = 0?).
quote:
As I reread this in draft, it struck me to ask whether anything in the universe is not "specified." Can anyone name an object or process in nature for which one cannot identify a specification? Does S ever take the value 0?
All events can be trivially "specified" by a constant rejection function. This produces a specification that encompasses the entire sample space.
However, we are encouraged to find a specification that is as tight as possible (because this will minimize the calculated probability of the specification). The specification that contains the entire sample space is extremely loose. So the answer to your question is that all events are specified, but some specifications are tighter than others and only the really tight specifications tend to be interesting. (All this should be understood under the assumption that Dembski's definition of "specification" is coherent, of course.)
Erik
[ 01. February 2004, 12:18: Message edited by: Erik ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 01. February 2004 12:35
Specification should also be real-valued from 0 to 1. Actually, specifications are never really quite "zero", as the universe can be finitely discretized, so the specification of "any possible state of a discrete model of the universe", which fits every possible observation, only specifies a finite number of things.
We should have something like S = (probability of observed event meeting our specification) / (probability of any event that meets our specification)
This is not mathematically correct, but it illustrates why S should be real-valued and never quite reach zero.
IP: Logged
|
|
Cre8ionist
Member
Member # 140
|
posted 01. February 2004 12:46
RBH, not a good weekend for me timewise. Perhaps tonight. I want to jump in a little on Rex's thread but still mulling it over.
On the S of SC point. It was Leslie Orgel who said quote:
In brief, living organisms are distinguished by their specified complexity. Crystals are usually taken as
the prototypes of simple well-specified structures, because they consist of a very large number of
identical molecules packed together in a uniform way. Lumps of granite or random mixtures of polymers
are examples of structures which are complex but not specified. The crystals fail to qualify as living
because they lack complexity; the mixtures of polymers fail to qualify because they lack specificity.
[Leslie Orgel, "The Origins of Life", Chapman & Hall, 1973.]
Apparently he thinks that random mixtures of polymers lack specificity. Now if you're looking to understand specificity, understand it in the way that Orgel means it. I believe that's the same way Dembski means it. More later..........Cre8
[ 01. February 2004, 12:49: Message edited by: Cre8ionist ]
IP: Logged
|
|
Erik
Member
Member # 160
|
posted 01. February 2004 13:15
quote:
Rex Kerr: Specification should also be real-valued from 0 to 1. Actually, specifications are never really quite "zero", as the universe can be finitely discretized, so the specification of "any possible state of a discrete model of the universe", which fits every possible observation, only specifies a finite number of things.
A "specification" is not a number -- it is an event (i.e. a subset of the sample space). RBH's variable S is not a "specification", it is defined to take the value 1 if a specification exists and the value 0 otherwise. With a suitable understanding of the definition of C this makes expression "CSI = S * C" correct. Changing the definition of S in the above fashion will make the expression incorrect, so why do it?
Erik
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 01. February 2004 14:41
Cre8:
Now if you're looking to understand specificity, understand it in the way that Orgel means it. I believe that's the same way Dembski means it.
And what about complexity?
quote:
We can now see that Dembski's "specified complexity" is merely a label we apply to an object for which we do not currently have a viable detailed explanation, i.e. one which is detailed enough to allow us to calculate a probability and where the probability turns out to be reasonably large. The "specified" part of the label arises because Dembski is summing the probability of all outcomes matching a "specification". It might be argued that Dembski's meaning of "specification" is similar to that of Orgel and Davies. However, his meaning of "complexity" is quite different from theirs. As we have seen, it is merely improbability in disguise, and must be calculated with respect to all the chance hypotheses we can think of. Orgel and Davies, on the other hand, use "complexity" to mean Kolmogorov complexity, i.e. incompressibility, which does not depend on probability at all and so is independent of any chance hypothesis. Clearly, Dembski's "specified complexity" is quite different from that of Orgel and Davies
Metanexus
In fact I would argue that specificity and specification may not be the same concepts.
quote:
One way around this problem is to assume that inorganic catalysts were available to ensure that only the correct nucleotides formed. For instance, when the components of nucleotides became adsorbed on the surface of some mineral, that mineral might have caused them to combine only in specific orientations. The possibility that minerals served as useful catalysts remains real, but none of the minerals tested so far has been shown to have the specificity needed to yield only nucleotides having the correct architecture.
Orgel
and
quote:
Wa¨chtersha¨user argues correctly that the constraints imposed on the orientation of molecules when they are adsorbed or synthesized in place on a surface should often lead to an increased specificity in their reactions. However, the very limited studies in prebiotic chemistry relevant to this hypothesis have not uncovered any highly regiospecific reactions.
Self-organizing biochemical cycles, Leslie Orgel, PNAS 2000 vol. 97 no. 23 12503–12507
Or
quote:
We have used amino acids activated by carbonyldiimidazole to study the enantiospecificity of peptide elongation in aqueous solution. Peptide 'primers' Glu10 and Ala3Glu10 were elongated with the enantiomers of arginine, glutamic acid, asparagine, phenylalanine, serine and valine. The homochiral addition was always the more efficient reaction; the enantiospecificity was large in some cases but very small in others. In every case Ala3Glu10 was elongated more efficiently than Glu10.
The specificity of peptide chain extension by N-carboxyanhydrides. Wen KE, Orgel LE. Orig Life Evol Biosph. 2001 Jun;31(3):241-8.
and
quote:
"There is no agreement on the extent to which metabolism could develop independently of a genetic material. In my opinion, there is no basis in known chemistry for the belief that long sequences of reactions can organize spontaneously -- and every reason to believe that they cannot. The problem of achieving sufficient specificity, whether in aqueous solution or on the surface of a mineral, is so severe that the chance of closing a cycle of reactions as complex as the reverse citric acid cycle, for example, is negligible." Leslie Orgel, 1998 (The Salk Institute for Biological Studies).
[ 01. February 2004, 14:57: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 01. February 2004 19:11
Erik points out that
quote:
A "specification" is not a number -- it is an event (i.e. a subset of the sample space).
Usually sample spaces have events as elements (so a specification is not an event but a set thereof), but the point is still valid: specifications are not numbers.
However, to calculate the complexity, one needs to calculate "specificational resources", which one can introduce in the S term of the S*C equation. If it is already included in the C term, then there is no point in having an S at all, as you cannot compute specificational resources without a specification!
[ 01. February 2004, 21:41: Message edited by: Rex Kerr ]
IP: Logged
|
|
Erik
Member
Member # 160
|
posted 02. February 2004 06:45
quote:
Rex Kerr: Usually sample spaces have events as elements (so a specification is not an event but a set thereof), but the point is still valid: specifications are not numbers.
The elements of the sample space are called outcomes. A set of outcomes is called an event (in advanced treatments one imposes the restriction that only certain kinds of sets of outcomes are events). Events are therefore subsets, not elements, of the sample space. A "specification" in Dembski's sense is an event.
There is probably little to gain in making RBH's equation exactly correct, but the specificational resources should come in as a cut-off for C, not as a multiplying factor. E.g.,
C = H(-log2(Pr(specification) * spec.res. * repl.res.)) =
= H(-log2(Pr(specification)) - log2(spec.res. * repl.res.))
CSI = S * C
where H is the Heaviside step function. With 2^-500 as a universal probability threshold this reduces to
C = H(-log2(Pr(specification)) - 500)
quote:
Rex Kerr: If it is already included in the C term, then there is no point in having an S at all, as you cannot compute specificational resources without a specification!
Yes, the factor S is potentially misleading, as I noted in my first reply to RBH. I wrote: "Writing CSI = S * C is potentially a bit misleading, because C depends on the specification (what's the value of C when S = 0?)." The question in the parenthesis was supposed to call attention to the fact that C is undefined when S = 0.
Erik
[ 02. February 2004, 07:01: Message edited by: Erik ]
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 02. February 2004 08:03
Erik wrote quote:
Yes, the factor S is potentially misleading, as I noted in my first reply to RBH. I wrote: "Writing CSI = S * C is potentially a bit misleading, because C depends on the specification (what's the value of C when S = 0?)." The question in the parenthesis was supposed to call attention to the fact that C is undefined when S = 0.
I wasn't so much trying to write a mathematical formalism as to highlight the fact that a "specification" either exists or doesn't (again assuming that "specification" is a coherent notion). As Erik points out, there is a conditional lurking: absent a specification, one doesn't actually know how to calculate C, since the specification must tell us what the relevant components are for the calculation. Note, however, that Dembski in his flagellum example in NFL provides a sort of specification at one level - some kind of vaguely described outboard motor - but does what calculating he does at a different level, the level of individual proteins.
A specification is a attribution of class membership to a particular instance, and identifying the appropriate class is where a whole lot of fuzz enters the analysis.
RBH
IP: Logged
|
|
Cre8ionist
Member
Member # 140
|
posted 02. February 2004 18:20
Pim, on the fly again, Dembski does write of Chaitin-Kolmogorov-Solomonoff in section 2.4 of NFL (Case Study: The Compressibility of Bit Strings)
Yes, the C in SC is different for Orgel and Dembski.
Dembski's version has to admit highly compressable strings due to the fact that they can signal a design inference, as in the Caputo case.
quote:
To sum up, the collection of algorithmically compressible (and therefore nonrandom) sequences has small probability among the totality of sequences, so that observing such a sequence is reason to look for explanations other than chance.
Point being made though was not really that life is full of specified complexity ala Dembski, but that life is well known to contain specified complexity (CSI). A more generalized statement, not really aimed at Dembski only. I think that whether you're talking about Dembski, Orgel, Thaxton, Davies, Eigen, Kuppers, or yes, even Dawkins himself, they all talk about the highly specified and complex arrangements of biological material that life displays in their writings. Therefore, there'll be plenty of opportunities to test the filter, which was actually the point of my statement.
Perhaps if Dembski had just called it Specified Improbability we could have skipped half the arguments....One question for you, exactly when does something cross the line from non-complexity to complexity in your world?.........Cre8
[ 02. February 2004, 18:29: Message edited by: Cre8ionist ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 02. February 2004 22:35
Erik's correction of my terminology is, of course, accurate. I'm not sure whether I managed to misremember it, or whether I'd recently read something that had events, not outcomes, as the elements of a sample space. In any case, I'm wrong.
Sorry 'bout that.
Erik's formula also looks correct if CSI is supposed to be a classifier (i.e. 1 when there is CSI and 0 otherwise). I guess I'd agree that there's little to be gained that way, as looking at the output of a binary classifier isn't very informative when one is wondering about the robustness of a classification.
In any case, back on the game, I can't think of a good way to compute the probability of those crystals, so I'm going to repeat my previous assessment of "cannot conclude design" since we know of natural processes that produce crystals sort of like that. (They could be artificial ones, but I can't tell from the photo. Heck, they could be rendered with some 3D rendering package and not exist at all....but it still isn't safe to conclude design.)
In the smiley-face moon example, I think if you find the density of pointy-things and curvy-things you'd find that the probability that two points and one curve fall into an otherwise fairly flat area in an orientation that is conducive to seeing a face is fairly high. E.g. if we chop the moon up into 10^30 pieces and have only two possible "eyes" and one "mouth", and demand that each be in the proper place, the probability is still roughly 10^-90. So we can't conclude design there, either, using the filter.
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 02. February 2004 23:59
Cre8: Point being made though was not really that life is full of specified complexity ala Dembski, but that life is well known to contain specified complexity (CSI). A more generalized statement, not really aimed at Dembski only.
This does not make sense, how can one claim tjhat life is well known to contain CSI when the definitions of such are so different? That sounds like equivocation to me. In fact, the attempt appears to be made to suggest that CSI in nature may be 'well known' but as people have argued quite well on these threads, it is hardly clear that CSI really exists (CSI in the Dembski sense that is). By mixing definitions such as complexity, information, probability, specificity, any conclusion seems suspect and in need of further evidence.
Cre8: One question for you, exactly when does something cross the line from non-complexity to complexity in your world?.........
Complexity, especially the version of Dembki's has no real black and white complexity/no-complexity but rather complexity comes in gradations from low to high.
IP: Logged
|
|
|
Printer-friendly view of this topic