(5th April, 2013: stickying this, for a bit, as it has come up. Mung might like to comment).
Time to look at this in detail, I think 🙂
His definitive paper to date on CSI, is Specification: The Pattern That
Signifies Intelligence. It is very clearly written, not very mathy, but, by the same token, a paper in which it is easy (IMO) to see where he goes wrong.
Here is the abstract:
ABSTRACT: Specification denotes the type of pattern that highly improbable events must exhibit before one is entitled to attribute them to intelligence. This paper analyzes the concept of specification and shows how it applies to design detection (i.e., the detection of intelligence on the basis of circumstantial evidence). Always in the background throughout this discussion is the fundamental question of Intelligent Design (ID): Can objects, even if nothing is known about how they arose, exhibit features that reliably signal the action of an intelligent cause? This paper reviews, clarifies, and extends previous work on specification in my books The Design Inference and No Free Lunch.
It’s in eight sections, and the argument in brief goes like this:
We cannot conclude that just because a pattern is one of a vast number of possible patterns that it was Designed; to conclude Design, it has to be one of small subset of those patterns that conforms to some kind of specification. Fisher proposed that if a pattern of data was one that fell in the tail of a probability distribution (aka Probability Density Function, PDF) of patterns of data under some null hypothesis, we could reject the null, but he didn’t give a clear rational for the cut-off point. Dembski suggests that if we have sequence data which a very large number of sequences would be possible under a non-Design hypothesis, and those sequences are binned according to their “compressibility” (ease of description) then they will form a Probability Density Function in which there is a tail consisting of a small subset of easy-to-describe pattern. Under the non-Design null hypothesis, these will happen rarely. If, therefore, the number of opportunities for them to happen is low enough, we can reject non-Design. And if the number of opportunities required to give them a sporting chance of happening at least once in the history of the universe is fewer than the number of events that have occurred in the universe, then we can confidently reject Design.
If any ID proponents think I have mischaracterised Dembski’s argument, I welcome your comments. But, assuming I have this broadly right, here are the problems as I see them:
- He does not attempt to characterise the probability distribution of his compressible sequences under his “non-Design” null, and simply assumes that only Design processes could reliably result in highly compressible patterns that would be improbable under a process that assigned each element in the sequence independently from any other – he does not attempt to argue why this should be the case, and it demonstrably is not.
- He does not show how compressibility should be measured (in fact Hazen et al, as discussed here IMO do a much better job, by substituting functional efficiency for compressibility, but their paper does not help Dembski’s case)
- He ignores the fact that the very easiest-to-describe sequences (e.g. ranked order sequences) are readily produced by non-Design sorting processes, yet can be highly “complex” i.e. one of a vast number of sorted and unsorted sequences), e.g. Chesil Beach.
Now, there may be various other ID papers proposing some kind of alternative to CSI that tackle some of these problems, but my point is that these three objections are fatal flaws in Dembski’s concept, and that therefore any improvement has to tackle all three. But it would be interesting to see if there is any disagreement about whether I have his argument right, and what the flaws are.
I agree with Norm here. ID/creationists have always attempted to drag a “debate” onto their territory using their “definitions” while expecting their opponents to “prove” things using those fuzzy ID/creationist concepts.
Chemistry, physics, and biology have managed just fine without using any of the ID/creationist’s muddy concepts. The minute anyone attempts to describe chemical, physical, and biological systems in terms of ID/creationist definitions, all they get is infinite mud-wrestling as the words slip and slide gradually among emotionally charged definitions.
Hermeneutics, exegesis, etymology, and generalized word-gaming are not the methods of science.
As Joe Felsenstein points out, CSI could apply to fitness. However, fitness is already already a well-defined concept in biology. Why muddy it up with CSI? CSI carries too much baggage that makes people think they are talking about something when it fact they are not.
Dembski brings his own misconceptions into any discussion; and he has not done a very good job of defining what he means without propagating a lot of misrepresentations of science as he does so.
And as far as this paper of Dembski’s is concerned, it is simply another example of ID/creationists taking simple ideas explained in introductory textbooks – on statistics in this case – and muddying up the concepts with lots and lots of words that try to make the paper appear profound.
CSI has no use in physics, chemistry, and biology that is not already accomplished far better with the known concepts already taught in beginning courses. Worse, CSI reinforces the notions by ID/creationists that atoms and molecules are just inert things sitting around waiting to be selected and placed in highly improbable configurations that are specified ahead of time.
Evolution and natural selection don’t just apply to biological systems; they apply to almost all systems that are evolving in complexity while being sorted by the prevailing environment. No CSI is required.
Explain how CSI applies to the formation of crystals or any other kind of condensed matter.
If it doesn’t apply, then explain to us just where along the chain of increasing complexity that CSI has to replace the laws of chemistry and physics.
Yes, they have, Joe. I can only conclude that you do not understand them.
Well, go with Mayr and Dawkins if you like, but right now you are talking to me. I agree with Dawkins that if organisms were designed by an intentional agent they would probably look different from the way they do. And both teleology and teleonomy are “allowed” in biology.
Well, try re-reading it.
It’s more that I am not in agreement with your conception of “the Darwin, Dawkins, mayr or the current theory of evolution” which I think is a straw man.
Although there are somethings that I disagree with Darwin about, and of course many things I disagree with Darwin about.
Not really Joe. It’s about information that is complex and specified. The fact that he doesn’t happen to type them in that order is neither here nor there. And didn’t he write the UD FAQ?
Exactly. The referenced paper is an update on his earlier treatments.
I propose that complexity in a species’ genome would increase largely by mutation and drift when the population is well adapted to its environment and exists in large numbers for many generations. Increasing “complexity” in this case means increasing variation (an increased number of alleles).
If the population comes under environmental pressure leading to a bottleneck situation, complexity in the group would decrease, and “specification” would increase (via natural selection).
If the population divides into two daughter groups, there is speciation, and both groups are successful, then we would see an increase in both complexity and specification, and therefore a definite increase in CSI. An example would be getting brown bears and polars from an original group. There are two environmental specifications for the bears, instead of one, and more variant alleles than there were in the one original group if the two new ones are considered as a whole.
So, I think that mutation, drift, selection and speciation must be able to increase CSI in a genome, however we measure it.
Therefore, it’s not an indicator of intelligent design.
Somehow, I think I will be told by I.D.ers that “specification” does not mean environmental constraint.
You are confusing the “target” with the fitness landscape.
Not surprising as, atypically, in “weasel” the “target” and the fitness landscape happen to be identical.
Weasel is thus a very bad example of a genetic algorithm, and you can have “cumulative selection” without a target. Not without a fitness landscape, of course, because then you’d have no selection to accumulate.
Genetic algorithms have nothing to do with blind and undirected processes- Hazen never demonstrated blind and undirected processes can produce anything and Lenski sure as heck didn’t demonstrate such a thing.
Well, you could actually make your case as opposed to just falsely accusing me.
He said that?
Citation, please- one must wonder what all the fuss over ID is, then.
And yes I read Joe F’s post- it doesn’t make any sense, doesn’t have anything to do with functional information and still doesn’t have any real-world support.
Re-reading it ain’t going to help, Liz- it is bogus from the start- as I have said, and proven, CSI pertains to origins. Not my fault that Joe F quote-mines “No Free Lunch”.
LoL! The “strawman” erected by evolutionary biologists- look I have supported my claim wrt the theory of evolution, many times, with references to evolutionary biologists. OTOH all I get in return are false accusations.
Elizabeth,
What is a “fitness landscape” in the real world, Liz? It is a made-up construct for virtual worlds, Liz. It is purely artificial.
Joe, all posts “see the light of day”. I just move a few of them to a different location.
Well the “S” in CSI pertains to “specification”, Liz, so it would have some relevance
– and CSI only pertains to ORIGINS, Liz- Dembski said that many times yet you feel like you can just ignore that
Wow, just wow- what country do you live in?
In reverse order- positive evidence for Intelligent design has been presented. OTOH there isn’t any evidence for materialism’s “explanation”, which BTW makes the claim that you said no scientist makes-
If no scientist made the claim- the claim that matter, energy, necessity and chance are all that is required- then, again, what is all the fuss about ID?
We exist Liz. There is only ONE reality behind that existence and so far materialism has struck out. You want to gripe about CSI, well, what is materialism’s metric- IOW how do you know that there are plausible alternative mechanisms? And what is your evidence for these mechanisms?
You are welcome, William, but there is a problem with it 🙂
Natural selection may well select for decreased CSI (using Dembski’s definition) on occasions, if the CSI is calculated on, say the genome sequence. It may even select for decreased functional complexity, if the functional complexity is simply calculated on the basis of, say, what proteins are coded for, if the protein in question is not of functional utility – or is actively disadvantageous – to the phenotype. However, that doesn’t mean that natural selection can’t add CSI. You’ve done a bit of a bait and switch there, by slipping in the word “search”. Nobody is claiming that evolution is a process of “searching” for CSI. What evolution “searches” for (if we must use that potentially misleading and baggage-laden term) are phenotypes that thrive in the current environment; possibly even populations that adapt to changing environments. This will often entail an increase in CSI. It may not.
Evolution doesn’t have a net “upward” trajectory, it only “goes” where reproductive success lies. There are more roads in the direction of increased complexity than reduced (because complexity has a lower limit but not an upper limit) so you’d expect more CSI in general, over time, rather than less, but that doesn’t mean it anything is “searching” for increased CSI.
Please cite where Dembsi said that.
Hi Mike- why isn’t the onus on you to provide positive evidence for matter, energy, necessity and chance being all that is required to account for our existence?
Ya know if you could just demonstrate that “the laws of chemistry and physics” could produce what we call CSI, then you would have refuted a major tenet of ID.
But right now every time we have observed CSI and knew the cause, it has always been via some agency- always, 100%- almost law-like 🙂
It’s a description of what happens in the real world, Joe. For example, a peppered moth population inhabits a fitness landscape in which, along the patterning dimension, better camouflage has a peak.
I posted it in this thread, Liz- page 149- the sentence starting the very next paragraph after what Joe F quoted
No Free Lunch- page 149:
Please cite where Dembski said that.
Liz, whatever survives to reproduce, survives to reproduce- what is “good enough”- for whatever reason- usually gets to do so- your fitness landscape is never going to take a peppered moth and make an eagle out of it, no matter how many generations it has to work with- an eagle, amybe a hawk or falcon, would be king of the fitness landscape the moths reside…
Let me use an apt metaphor that applies to you.
You are standing on the New Jersey side of the Hudson River with your back to New York City denying the existence of New York City.
People try to tell you to turn around and look, but you refuse to do so; denying all the while that New York City exists.
The evidence is all around you, even as you bang on your computer keys denying that it is there. You just have to look.
Nobody in chemistry and physics uses the notion of CSI because there are far better concepts that describe what is going on. But if you don’t learn them, you’ll never know.
Forget CSI; it is a useless idea in science.
Well, CSI, per se, is a rather intractable value to compute as there is no straightforward way of computing “compressibility”. However, I’m happy to accept it in principle, and that it could, in principle, be applied to a genome. I think the Hazen paper does a better job of something equivalent (and actually applicable to biology).
Well, conceptually, it obviously can. It’s just you’d have to operationalize it more carefully than Dembski does to measure it. That’s exactly what Hazen et al did, and demonstrated that AVIDA (an NS system) could increase it.
Sequences that are made from elements that could theoretically be rearranged in a vast number of ways, if each element was selected by some process completely independent from any other (as in a series of coin-tosses, or throws of a die), but which is one of a much smaller subset that has some identifiable (“compressible” in Dembski’s formulation, or “functional” in Hazen’s)kind of pattern
Well, as Joe F said, he, and I, disagree that CSI is not a potentially useful concept. So neither of us are trying to “have it both ways”. And I’d certainly be willing to accept, for example, that DNA sequences that code for proteins is one of a small subset of the vast number theoretically possible DNA sequences that don’t. That seems a perfectly reasonable “specification” to me, and I fully agree with Dembski that such sequences are vanishingly unlikely to be due to “Chance” if “Chance” means “by a sequence of independent selection processes”. I just disagree that such a “Chance” hypothesis is necessarily a “no-Design” hypothesis.
In other words I think he has (inadvertently) equivocated with two meanings of “Chance”.
I agree, Joe. But I was responding to your point that a fitness landscape doesn’t exist in reality. It does. I just gave you an example. It’s a term we use to describe natural selection and drift. Both are perfectly real phenomena.
Fine. If what he mean by “origins” is “the origin of complex specified information” nothing I have said contradicts that. NS results in new (i.e. original) complex specified information.
You just refuse to deal with the reality that AVIDA does not represent NS nor anything in the real-world.
Liz, you don’t appear to undersatnd NS- it is just a result- and it does not apply to origins- IOW NS starts with existing CSI , the CSI that has to be explained before you can go using NS to explain something else.
The whole point behind CSI is right now every time we have observed CSI and knew the cause, it has always been via some agency- always, 100%- almost law-like
So when we observe CSI and didn’t directly observe it we infer some agency was involved.
And to refute that inference all one has to do is step up and demonstrate blind and undirected processes can produce CSI starting with no SI at all.
I’ve made the case many times, Joe, but you seem extraordinarily resistant to the concept. This is evident by your repeated raising of “weasel” as an example. In all practical genetic algorithms the “target” is not the same as the “fitness landscape”. In “weasel” it is.
I work in neuroscience, which is a branch of biology, and teleology is allowed. And take genetic engineering – it’s allowed there too. Also in breeding programs, or in evolutionary studies of domesticated animals. And it would be a perfectly viable hypothesis to propose that living things were intentionally engineered by aliens, as Dawkins himself has stated (in Expelled, no less), if there was any evidence. There’s nothing taboo about teleology in biology in principle, as long as it’s supported by evidence.
Well, your understanding of evolutionary theory is certainly very different to mine. I think mine is more current, and more accurate.
You have directly contradicted yourself here, Joe!
I do, Joe. I understand it quite well. Consider, please, the possibility that the error of understanding may be on your part. In fact, scroll up to the top of this page, and read 🙂
It does represent NS. That’s exactly what it does. Read the paper, download the program, and get back to me when you’ve played around with it for a bit.
Then explain and verify where your supposed “existing CSI” came from. But first explain exactly what the words complex, specified, and information mean, in ID world, and demonstrate exactly how they are measured in living and non-living things in nature. Also, demonstrate what the term CSI is useful for, and show evidence (not appeals to authority or bald assertions) that supports your claims and definitions.
Natural selection doesn’t add anything; it only redistributes and fixes what is already there. However, even if we go by Joe F’s redistribution and fixation method of “acquiring” CSI, I didn’t say that NS couldn’t “add” CSI, only that it hasn’t been demostrated that it, in principle, adds more than it subtracts.
No, you’re just doing the usual semantic obfuscation, red herring, straw man and obfuscation routine.
I fully contextualized in this thread that I mean “search” in the descriptive and not prescriptive sense. Perhaps this is once again a case of you responding to posts I’ve made out of context with previous posts, but seeing as I’ve already pointed out that problem to you, it seems to me to be more of a willful attempt at derailing the debate down out-of-context rabbit-holes.
One would think that the simple principle of charity would dictate that when I use the term “search” in an evolutionary sense, you assume I don’t mean that evolution is actively searching for a specific target, and that when I use the term “target”, I mean that in the general descriptive sense and not in the prescriptive sense.
Nor have I said they were; what is at issue is is whether or not NS adds anything significant to the equation of how CSI comes into existence (search in the post hoc, descriptive sense, not in the prescriptive sense).
Unless you can show that what NS produces as it differentiates between success and failure in practice is in principle a deviation from the mean in favor of increased CSI, then NS is useless as a mechanism for explaining increased CSI. It might increase it (according to Joe F’s method), but it doesn’t necessarily game evolution in favor of it, because, theoretically, it could just as easily reduce CSI.
In order to explain increase in CSI above the mean provided by chance mutation, NS must game the system in favor of increased CSI.
Unless NS actively skews progress towards increased CSI (by inserting pro-CSI-relevant active or oracle information), we can only expect those roads to be explored as far as chance can provide.
And, to avoid your usual obfuscation, I mean “chance” in the euphemistic sense, not as a literal entity named “chance”.
Once again, you will first have to provide us with a recipe of measuring or calculating CSI in an organism. No ID proponents have done that so far. You will be the first.
Look guys, this is the most simple and basic logic.
You (the general you in the anti-ID arena) have been claiming for a long time, arguing over and over about how we (ID proponents) don’t understand how it’s not a “chance” process and that it’s not just “chance mutations”, but that when you throw in natural selection as a filtering process, ****that**** is what makes the difference and turns evolution into a non-chance process that explainsthe highly complex, interdependent features (or CSI) in question.
Now, however, challenged with the argument that NS can also as easily (and perhaps even more easily) decrease the CSI in the genome, you stop arguing that it is NS that makes the difference between low and high CSI and start claiming that there was nothing that NS was required to explain in the first place – now, apparently, more specified, complex functional features don’t even need NS to explain their existence, because they aren’t anything special requiring anything other than chance* to explain their existence (*euphemism).
If what we were arguing about before isn’t even a real thing that requires NS to explain, why even bring up NS as if it is the difference-maker? Once again (modified for clarity):
Unless you can show NS to select more for increased CSI than for decreased CSI, you have no logical basis for the claim that NS adds anything significant to chance* as an explanation for increased CSI in the genome.
Do you ever apply the principle of charity, Liz, or do you just jump on every opportunity to derail, obfuscate, and rabbit-hole? Do you really not know what Joe meant there?
Well, this is simply the usual problem with mud-wrestling over words that have no particular meaning. What is the purpose in that?
Since you are an advocate of CSI, perhaps you would lay out an unambiguous definition of the idea, tell us how to calculate it, and tell us where it applies.
Start with something simple. Does it apply to crystals? If so, tell us how to calculate it.
Do you or do you not believe that crystals evolve and are subject to natural selection?
If you don’t think it applies to crystals, where does this notion of CSI become relevant in the hierarchy of complex evolving systems?
It seems to me that you and Joe G are doing what ID/creationists have done since the 1970s. You want the science community to accept your “concepts” even though you yourself can’t define or explain them; and then you ask members of the science community do disprove something you have made up and can’t articulate.
I would think that it would be your responsibility to establish the notion of CSI firmly in science. Dembski hasn’t done it yet.
Why is anyone obligated to “refute” something that doesn’t even have a definition?
William J Murray. Do you think CSI is calculable?
Look, ma, no hands! 🙂
We are saying that an interplay of random mutations and feedback increases fitness of organisms. CSI is a wholly creationists’ invention. It goes way back to Henry Morris who termed it organized complexity. It was later rebranded by Dembski. None provided a recipe for measuring or calculating the thing in a biological context.
Explain how to measure or compute CSI, WIlliam. Otherwise we are left to guess what it is, which is what Joe F. did in the opening post. Reread it.
If there is someone who derails, obfuscates, and rabbit-holes, and doesn’t apply the “principle of charity”, it’s Joe G and the rest of you IDists.
WJM, in case you missed it the first time:
Which has more CSI, a mouse, a turtle, a butterfly, an octopus, a human, a whale shark, or an elephant?
(continued)
Since creationists are not giving us a recipe for computing or measuring the CSI, we are left to guess what they want. Joe F’s opening post is an attempt to decipher Dembski’s vague ideas and to come up with a plausible guess of what CSI entails in a biological context. He did just that and showed that CSI — as he understands it — can be increased through natural selection.
If you don’t like Joe’s supporting arguments, go ahead and criticize them. If you think his formulation of CSI is wrong, formulate one yourself.
You can’t have your CAEK and not measure it, William.
So now you want me to calculate the value of a commodity about which you have argued unrelentingly for years that NS was sufficient to explain the existence of, only that I was too ignorant of biology to understand why it was sufficient?
Oh, my. Let’s look at what we have here now:
(1) If CSI doesn’t have a calculable value, or isn’t a meaningful commodity, then your arguments about NS being sufficient to produce it are revealed as nothing but rhetoric;
(2) If CSI does have a calculable value, or is a meaningful commodity, then you must show that NS skews evolutionary processes in favor of increasing it or have NS revealed as inconsequential as an explanation for it.
Either way – if CSI is a calculable and/or meaningful commodity, or if it is not – the argument that NS explains it is revealed as nothing but rhetoric unless its proponents can show that CSI is a meaningful commodity that NS skews evolution in favor of.
See, there really isn’t any deep math or biological education and experience necessary to get to the heart of most of these debates.
I’ll help you make your observations less selective, and more complete. You could have said that every time you have observed CSI and thought you knew the cause, it has always appeared to come from a living material intelligent agency for whom CSI would be a prerequisite. You have never observed an intelligent designer who did not owe its existence to pre-existing CSI, and therefore, it does not make sense to infer that intelligence is the best explanation for the existence of CSI of unknown origin.
Isn’t that better?
..and the designer has always been ‘people’. why not infer ‘people’? It’s more accurate.
The Hazen paper tells you how to calculate biologically functional information- what is your problem?
Let’s try to avoid being disingenuous, William. This has been asked quite a bit – here, at AtBC, at UD, at Pandas, and by a variety of critics on several book sites. It’s a well-established issue for the concept of ID – none of it’s proponents have been able to give it a definitive mathematical definition.
As for the the insinuation regarding NS, nobody is claiming that NS requires CSI, however if CSI is actually a viable characteristic of certain objects, there is no reason to think NS cannot account for it. Thus several folks have demonstrated how it could – hypothetically – account for this mystery characteristic, providing said mystery characteristic falls within a very narrow definition. However, since no ID-advocate seems to be able to actually demonstrate what CSI actually is, it’s a rather moot hypothesis at this point.
Perhaps we could ask which has more CSI:
http://upload.wikimedia.org/wikipedia/commons/7/72/Stone_sphere.jpg
or
http://photography.nationalgeographic.com/staticfiles/NGS/Shared/StaticFiles/Photography/Images/Content/dew-spider-web-698720-ga.jpg
or
http://image.shutterstock.com/display_pic_with_logo/60660/60660,1279953410,3/stock-photo-massive-cathedral-termite-mounds-nasutitermes-triodae-kakadu-national-park-northern-territory-57713866.jpg
or
http://www.susancohangardens.com/blog/wp-content/uploads/2010/02/birds-nest.jpg
or
http://1.bp.blogspot.com/_Q_G8tOtOEpg/TOzTr3cwNrI/AAAAAAAAAKA/m3qQmUUDu0M/s1600/beaver-dam-insurance.jpg
No, we are not asking you to calculate the value of CSI. You need to provide a recipe for calculating or measuring it. ID proponents have not done that yet.
As I explained above (for the umpteenth time!), ID proponents do not know themselves how to compute or measure the value of CSI even for the simplest biological systems. What are we to do in this situation?
(1) Do nothing until ID proponents get their act together and figure out how CSI can be either calculated or measured. You could help them along, perhaps.
(2) Make some reasonable guess that CSI is a proxy for fitness. That is what Joe F did in the opening post of this thread. We know that natural selection can increase fitness, so if Joe’s guess is right then CSI can be increased through natural selection. If you don’t like Joe’s approach, fine. Feel free to present your own vision of CSI and tell us how it can be measured or calculated.
LOL.