At Uncommon Descent, William Dembski’s and Robert Marks’s coauthor Winston Ewert has made a post conceding that using Complex Specified Information to conclude that evolution of an adaptation is improbable is in fact circular. This was argued at UD by “Keith S.” (our own “keiths”) in recent weeks. It was long asserted by various people here, and was argued in posts here by Elizabeth Liddle in her “Belling the Cat” and “EleP(T|H)ant in the room” series of posts (here, here, and here). I had posted at Panda’s Thumb on the same issue.
Here is a bit of what Ewert posted at UD:
CSI and Specified complexity do not help in any way to establish that the evolution of the bacterial flagellum is improbable. Rather, the only way to establish that the bacterial flagellum exhibits CSI is to first show that it was improbable. Any attempt to use CSI to establish the improbability of evolution is deeply fallacious.
I have put up this post so that keiths and others can discuss what Ewert conceded. I urge people to read his post carefully. There are still aspects of it that I am not sure I understand. What for example is the practical distinction between showing that evolution is very improbable and showing that it is impossible? Ewert seems to think that CSI has a role to play there.
Having this concession from Ewert may surprise Denyse O’Leary (“News” at UD) and UD’s head honcho Barry Arrington. Both of them have declared that a big problem for evolution is the observation of CSI. Here is Barry in 2011 (here):
All it would take is even one instance of CSI or IC being observed to arise through chance or mechanical necessity or a combination of the two. Such an observation would blow the ID project out of the water.
Ewert is conceding that one does not first find CSI and then conclude from this that evolution is improbable. Barry and Denyse O’Leary said the opposite — that having observed CSI, one could conclude that evolution was improbable.
The discussion of Ewert’s post at UD is interesting, but maybe we can have some useful discussion here too.
Behe’s argument is at least testable. Unfortunately for him, it has been tested.
It does speak directly to the question of whether evolution is sufficient.
If Dembski’s “specification” criterion is in fact NOT question-begging, it needs to be explained in a better manner than I’ve seen him do. (I admit here that I haven’t and am unlikely to ever read any of his books.) But, e.g., in something I HAVE read, he writes,
A bit of terminology will prove helpful here. The “good” patterns will be called specifications. Specifications are the non-ad hoc patterns that can legitimately be used to eliminate chance and warrant a design inference. In contrast, the “bad” patterns may be called fabrications. Fabrications are the ad hoc patterns that cannot legitimately be used to eliminate chance.
He here uses both “non-ad hoc” and “legitimate” to tell us when “patterns” are “good” and doesn’t attempt to define either of those terms: he just gives an example of a case in which he thinks we’ll all agree (presumably “legitimately”) that a particular “pattern” was clearly ad hoc. But that’s just moving the hard question from the determination of a design “pattern” to the question of “ad hocness” (“legitimately” discovered). No real help there, I wouldn’t think.
In the matter of design v. no design in evolution there obviously is no agreement on on these matters, and his determinations are apparently entirely subjective. That being the case, this particular “filter” would seem to have to be discarded. And, as the burden is here on the design theorist (just as it is in all the non-science cases he describes–Caputo, copyright, etc.), the result would seem not to be favorable to the case he is trying to make.
[I apologize in advance if I am neglecting some non-question-begging criteria he has provided elsewhere.]
What’s left in the ID cupboard?
I should compliment Keith S on drawing a comment out of Winston Ewert. Compliments also to Winston Ewert for making it. Perhaps also to Barry Arrington for leaving the door opening for long enough for this clarification to result.
The trouble is that CSI doesn’t really point to anything at all. It exists in designed objects, in the evolved, and, using Dembski’s criteria, in crystals and the like. You can look at various objects and ask if the evidence of design exists in them, if evidence of evolution exists in them, or if an ordering process such as crystallization caused them.
When ID is involved, that’s the problem, for life has evidence of evolution and not evidence of design. Pretending that CSI is evidence of design is the sciency way of fooling some people (often themselves, I’ll wager) into believing that there is evidence of design. Otherwise they’d have to just admit that it’s god of the gaps all of the way.
But of course it is.
Glen Davidson
By the way, what the heck does “specified” mean in specified complexity?
Sometimes it seems to me that it means “functional” but I don’t see why that should not be the product of evolution.
But most of the time, it seems that by “specified” they mean “I assume it designed”.
Wouldn’t it be much easier to understand ID if they talked about “I assume it designed” complexity?
“This feature has high “I assume it designed” complexity. It can’t have evolved because it is very complex and I assume it designed”.
Also, did you notice how specified (intended) complexity in shitty phallacious explanations is much higher than in scientific theories? Mmm, why would that be?
GlenDavidson,
I beg to disagree. If we used CSI the way Dembski did from 1998 to 2006, you would be right. Or rather, the way we all assumed he did. But he has now clarified his argument (or perhaps its changed his argument). Now we see that CSI is only present when you have ruled out all the normal processes of evolution, deterministic or stochastic.
So seeing CSI is prima facie evidence of Design, because you have to have already ruled out anything but Design in order to call it CSI.
Both opponents of ID, and many of its most ardent adherents, did not understand this condition. I wrote that CSI could be observed. Mea culpa. But so did most of the pro-ID commenters, including Barry A, Denyse O’Leary, and virtually all of the people who comment at UD. Now it has been clarified that it cannot be observed unless you have also ruled out normal evolutionary mechanisms.
That condition was pointed out by many people here and at PT. Now, finally, Winston Ewert is admitting the correctness of those posts.
Guillermoe,
The best way of rescuing the notion of “specified” is to make specification be fitness. We are, after all, talking about the ability of natural selection to bring about adaptations.
But I wasn’t discussing Dembski’s position in particular, rather how CSI is typically used by IDists. I mentioned Dembski’s “criteria,” since he calls simplicity “complexity” in some cases (and most IDists seem to follow with that), yet I’m hardly interested in wading through his various attempts to ignore the evidence.
And yes, CSI is typically just used as an attempt to define the apparent complexity of life as “designed,” whatever twists and turns one particular apologist’s thought takes.
Glen Davidson
Great, great point.
It’s weird to call something prima facie evidence if its not actually observable at all but is deemed to be present as a result of an inference. That’s precisely the opposite of the textbook understanding of what prima facie evidence for something can be. There’s no “first glance” here, but a conclusion.
Anyhow, I think the concept of “specification” that is being depended on by Dembskiites isn’t itself careful specified. As indicated above, it seems to me to be question-begging–at least as described in the (admittedly brief and early) Dembski writing I quoted.
Prima facie is my wording, not Dembski’s. Let’s modify that to “strong evidence”.
They are not assuming the conclusion. They are just assuming that some other means (undiscussed) has allowed you to draw the conclusion that standard evolutionary mechanisms have an extremely low probability P(T|H) of giving rise to an adaptation at least as fit.
So they’re leaving all the heavy lifting to occur offstage, and not saying how it’s done. Then they drag in CSI at the end, making it a redundant flourish.
(By the way, could one of the administrators here please assign this thread the categories “Intelligent design” and “Evolution”? Also please correct the typo in the original post “P(T}H)” to be “P(T|H)”. I lack the authority to do those.)
Joe Felsenstein,
I corrected the typo and assigned categories. I also changed permissions for you so you no longer lack authority. 🙂
I see on UD that Winston Ewert says explicitely that the calculation of CSI presumes no evolutionary pathway, and that if you can demonstrate irreducible complexity, then CSI is superfluous.
Thanks. At last, unlimited power! (Rubs hands, eyes gleam maniacally, fiendish grin).
IC is a different argument than CSI. Perhaps they can be combined, but Dembski’s original argument concerned CSI and his Design Inference contained no mention of IC. We can discuss the uses or non-uses of CSI without discussing IC.
Barry Arrington has weighed in at UD, with a post “Back to Basics” which declares that the critics of CSI
Let’s be clear. CSI is not a circular notion, whichever of the two ways that it seems to have been defined. The most recent way, or the clarified way, requires that we have eliminated the possibility of the adaptation being brought about by normal evolutionary processes. It assumes that if you see CSI, you have first used some other means to rule out normal evolutionary processes such as natural selection. OK, maybe you can do that (though how is not explained).
So it is not CSI itself, but the argument that observing CSI rules out normal evolutionary processes that is circular. Circular because we are only allowed to declare CSI to be present when we have first ruled out those normal evolutionary processes.
Does anyone use CSI in that way? Here is one example of its use:
(CSI is only present if one has already ruled out chance or mechanical necessity). Would Barry agree that this statement makes a circular argument? The author of that 2011 statement is Barry.
Winston Ewert, in a new post at Uncommon Descent, argues that Dembski has not changed the way he defines CSI or uses it in his Design Inference. I am not convinced. “keiths” gives some citations arguing against Ewert’s position.
In particular, Dembski had a Law of Conservation of Complex Specified Information (e.g. in his 2001 book No Free Lunch: Why Specified Complexity Cannot be Purchased Without Intelligence). It is unclear why one would need that Conservation Law, supposed to show that normal evolutionary processes could not produce CSI, if you had defined CSI as not present unless those normal evolutionary processes had already been ruled out. Which is to say you have ruled out in its very definition that Specified Complexity could be “purchased” by normal evolutionary processes.
I hope some of you will take a look at that UD thread and let me know whether you understand Ewert’s position.
I do congratulate Ewert for openly grappling with these issues.
Joe, this remark of Ewert’s:
Dembski doesn’t say that life is improbable because it exhibits specified complexity. Rather, he says that life is improbable because of Behe’s argument of irreducible complexity. It argues that it exhibits specified complexity as a consequence of that probability. He doesn’t regard specified complexity as an established fact. Rather he explicitly stats the “jury is still out.” The specified complexity argument is explicitly contingent on the irreducible complexity argument.
contains mass of epistemologically confusing matter, centering on the use of “because.” I don’t think it’s possible to evaluate statements like that, because the writer has not made clear whether these becauses refer to why something is happening (ratio essendi) or how we come to know something is happening (ratio cognoscendi). Ewert’s troubled remarks over the years on this issue betray his confusion (and probably Demski’s too).
This isn’t an issue of probability: it’s a matter of insufficient clarity, and, I think, too few philosophy courses.
Over at UD in the “Back to Basics” thread there is a big discussion where Barry Arrington and the usual crew of commenters (Mung, Joe, Kairosfocus, etc.) are arguing that Dembski’s definition of specified complexity is basically the same as Leslie Orgel’s 1973 definition.
“keiths”, “r0bb” and “learned hand” try to correct this. Dembski’s useage (at least after 2006) is different, because it adds an important condition. Namely, that specified complexity is not present if normal evolutionary processes have a high probability of producing the adaptation. The UD commenters, with the exception of Winston Ewert, have not noticed this point.
Dembski’s concept is clearly derived from Orgel’s. But Dembski will not compute Specified Complexity, or count it toward declaring CSI to be present, if “chance” or “necessity” can explain that adaptation. You will not find any such condition in Orgel’s discussion.
On this critical point Barry Arrington has still not realized that his own 2011 challenge to show that natural selection can explain any instance of CSI conflicts with Dembski’s recent definitions of CSI. Winston Ewert understands that, but the news has not reached Arrington.
Dembski’s
Barry is attempting to clarify on a new thread:
Is “clarify” a synonym for “obfuscate” now? I can’t keep up with American English!
I see Barry Arrington is continuing his detailed and incisive argument for an ID paradigm.
presumably…conceding that using Complex Specified Information to conclude that evolution of an adaptation is improbable is in fact circular.
link
Yet Joe quotes Ewert. We can all read it in the OP.
I’d love to play poker with Barry. I’d take all his money.
Moments before being shot.
He’s certainly making it clear that a layperson shouldn’t try to argue probability with people who know wtf they’re doing.
If any of you are interested in joey g’s incomparable expertise on the subject of CSI, go here:
http://theidiotsofintelligentdesign.blogspot.com/2014/11/here-are-some-posts-by-muslim-yec-joe.html#comment-1697107738
P.S. I am ‘The whole truth’.
Slightly OT – I don’t have much to say about attempts to bit-map biological polymers, other than that this is a poor representation of ‘function’ – but this from ‘mahuna’ caught my eye:
Or: replicating clouds.
Yeah, a few sporadic junkyard-747s would hardly be notable. Unless one of them had “METHINKS IT IS LIKE A WEASEL” painted on the side.
FWIW, the problem I’ve had with my own serendipitous, tornado-constructed junkyard 747 is that it tends to run a little loud. My mechanic can’t seem to figure out why, but he suspects the old beer keg comprising part of the high-bypass turbofan engine.
Barry’s latest:
Is that a notadmission of error?
1. He has not admitted error. He approves of HeKS’s statement that it is not that CSI is defined as present only when the adaptation cannot be produced by natural processes. HeKS says this is an empirical observation.
But Dembski did define CSI as only present when the specified complexity could not be produced by “chance” (normal evolutionary processes). He didn’t make this definition conditional on the outcome of some more observations. He just didn’t.
2. Barry seems to have not commented on whether his 2011 call for examples of CSI produced by natural processes was vacuous (it was, given the post-2006 version of Dembski’s definition of CSI).
3. Winston Ewert expresses approval, in the comments on the new post, of Barry’s and HeKS’s positions, even though they contradict Ewert’s previous posts.
4. Keith S. gives a good summary of “what we have learned”. Now all that is necessary is for Barry and friends to admit that this is what they have learned. Alas, Barry has now “gone back to work”.
How hard is it to realize that so far only evolution is an actual description of a mechanism producing new biological features?
Doesn’t the challenge start there? Aren’t the IDiots who, to be up to the challenge, should describe how “design” produces biological features?
The best thing keiths has done there, IMHO, is point out the clear errors by ForJah and vjtorley regarding Dembski’s def. of CSI. That mistake is obvious and incontrovertible.
I’d say the credit for the dismemberment of Arrington should really go to Learned Hand (who also happens to be less obno/chest thumpy about his work).
In the “Thanks for the CSI Debate” thread at UD, there are still some interesting comments appearing. “Moose Dr.” (here) has imagined a random-mutation-plus-natural-selection model that can generate 500 bits of “CSI like information”. “Moose Dr.” has self-identified as in the ID camp, but is clearly worried about why the definition of CSI has to rule out RM+NS by definition.
A couple of comments about that: Under the earlier Dembski CSI definition (or perhaps people’s misinterpretation of it) CSI was declared if the adaptation was so far out on the fitness scale that its probability of occurrence by random mutation was less than 2-500. Then the issue was whether natural selection could get you that far out. Dembski had a Conservation Law that seemed to say that no natural evolutionary force could do this. Various people, including Elsberry and Shallit and me punched big holes in that Law’s ability to say any such thing.
In fact, at this site (here) I gave precisely that, a simple model calculation of natural selection that could result in that earlier sort of CSI. Well, actually my model only had 100 loci so it couldn’t quite do that, but if you straightforwardly extend it to more than 500 loci it certainly could.
Some ID folks seemed a bit startled by this, but most of the ID or creationist commenters on that 2012 thread tried to divert the discussion to the Origin Of Life or else tried to declare that since the model was intelligently designed (by me) that it was not a model of natural processes. Which is silly — it is like saying that since a computer simulation of rocks falling down a hillside in a landslide is intelligently designed, therefore so is a landslide.
Anyway, I hope “Keith S.” or one of his allies calls that TSZ thread to the attention of “Moose Dr.”
Also to Gary Gaulin, who asks if anyone is interested in making a computer model of CSI.Oh pray Gaulin never finds here. You can wash your website 100 times and never get him out.
Richardthughes,
OK, I’ll take that part back — you know more than I.
What is interesting in all this is that yes, there are quite a few denizens of UD (Barry, Moose Dr., even Joe) who take their definition of CSI from Dembski’s No Free Lunch: Why Specified Complexity Cannot Be Purchased Without Intelligence. And they interpret that definition as many of the rest of us did: CSI could be present even if natural selection has not been ruled out. I used the full title of Dembski’s work above because even the title seems based on the assumption that we need to figure out what can “purchase” CSI — that would make no sense if it was built into the definition of CSI that it could not be “purchased” without intelligence. The book would then be rather short.
Other commenters there (including, I believe, KF) also declare that CSI is essentially the same as the concepts used by Orgel, by Hazen, and even by Kirk Durston. Correct me if I’m wrong, but none of those folks put a condition into their definition that it wasn’t specified information if it could be put into the genome by natural selection.
What they haven’t admitted is that Dembski’s definition has changed away from theirs (or maybe that it has been clarified so that we all now see that we were mistaken). What seemed to be a useful definition which could lead to a proof that CSI could not arise by natural selection and mutation, is now an afterthought of no importance. Dembski seems to have given up on using his Law of Conservation of Complex Specified Information to show that CSI cannot get into the genome by natural selection. Now he just doesn’t call it CSI if natural processes can get it there.
Dembski seems to have adopted the semi-Deistic view of Michael Denton’s “Nature’s Destiny.” Which harkens back several centuries. the great clockwork metaphor.
Presumably, humans have a magical little guy inside that is exempt from determinism. But some theists believe in predestination.
Wow. There are now two brain dumps by Kairosfocus up at UD (here and here) that not only have some of the same diagrams he has been posting a lot in his FYI/FTR rants, but one of the diagrams is repeated in both posts. The posts are FYI/FTR which of course means Comments Off.
That’s a lot to read. Can anyone point out to me where in all that is the discussion of Dembski’s post-2006 condition that the pattern is not CSI if the probability of getting to the target set T given a hypothesis H of natural forces is not small. In other words, where he discusses whether Dembski’s CSI (not KF’s FSCO/I) requires that the “chance” and “necessity” mechanisms that you have to rule out first include natural selection? I don’t see it.
Isolated islands. There’s the only remaining ID argument. KF has been pushing that since forever. No evidence of isolation. It’s just obvious.
Two more FYI/FTR rants today by Kairosfocus. Many of the same figures. The topmost four threads at UD are now these Comments Off rants. I still don’t see any discussion of the issue of whether natural selection has to be ruled out before you declare CSI to be present.
Where will it end? When will that central issue actually be dealt with?
When Jesus returns.
Hi Joe,
When were you published in a peer reviewed publication establishing the circularity of using CSI to conclude design?
Or you had to wait for keiths to point it out to you?
But you will now publish this astounding finding and include keiths as a co-contributoror?
No?
I have published over 100 peer-reviewed papers. And you?
However of course almost all of this discussion has not been peer-reviewed. Doesn’t worry me. And when I have written on Panda’s Thumb or on TSZ about this I have tried to credit people before me for pointing out the issue. I didn’t discover it by myself, but I hope I have written clearly on it.
Out of curiousity, Mung, do you see that including the term P(T|H), where T is the tail probability (or target set) for CSI, and H is the hypothesis of normal evolutionary processes including natural selection, means that Dembski will not declare CSI to be present unless the target T is considered on other grounds to be very improbable under that evolutionary process?
Some people in the UD discussions do see that, do you? I’m impressed that these others are honest about seeing it. Are you with them in that?
And if you agree with “Keith S.” and with others including me on that, doesn’t that mean that saying that having CSI rules out NS+RM is a rather uninteresting and trivial statement?
I’m curious as to what your take on this is.
The only people interested in such have already read it here and on the threads at UD. You over-estimate the interest ID has in the wider world.
ID died at Dover. Dembski noticed this and changed his tactics (he’s not been on UD much lately, notice that? Too busy writing books for ) but it’s business as usual for the rest of the crowd at UD.
Man, the Earth is round and nobody would publish a research paper proving that. When something is so obvious, it’s very hard to publish it.
If you don’t agree with Joe, you could prove him wrong very easily. Just choose a real biological example and prove it’s designed using CSI. We’ll tell you what we think about it.
No?
mung, I too would be interested in seeing your published, peer reviewed ID papers along with the published, peer reviewed ID papers by other ID-creationists “establishing” the NON-circularity of “using CSI to conclude design”. Can you produce any such papers? And by “using CSI”, do you mean calculating CSI, measuring CSI, computing CSI, imagining CSI, or what?
Since published, peer reviewed papers matter so much to you (and joe g, etc.) let’s take a look at the peer reviewed publication record of the most vociferous CSI-dFSCI-FSCO/I pusher at UD: kairosfocus.
Here is gordo’s entire publishing history:
http://books.google.com/books/about/A_thirty_two_channel_integrated_voice_an.html?id=R0BnNwAACAAJ
Is it published in a peer reviewed science journal? Does it establish “using” CSI-dFSCI-FSCO/I as a NON-circular method to “conclude design”? Does it establish ID in regard to biological evolution and/or the origin of life?
I have a suggestion: you, joe g, kairosfocus, gpuccio, torley, phoodoo, dembski, uprightbiped, arrington, luskin, gauger, axe, mclatchie, egnor, meyer, hunter, o’leary, klinghoffer, and all of the other ID-creationists should get together, pool your massive intellects, co-author a paper that establishes “using” CSI-dFSCI-FSCO/I as a scientific, reliable, accurate, well-defined, astounding, NON-circular method to “conclude design” of life, evolution, and everything else in the cosmos, and submit the paper for peer review and publication in a legitimate science journal.
My comments are still being held up by moderation.
Creodont2,
Yes, I see. I just approved them. Try a test post in the “moderation issues” thread and I’ll try and resolve it from there.