Sometimes very active discussions about peripheral issues overwhelm a thread, so this is a permanent home for those conversations.
The Skeptical Zone
"I beseech you, in the bowels of Christ, think it possible that you may be mistaken."
Sometimes very active discussions about peripheral issues overwhelm a thread, so this is a permanent home for those conversations.
Joe Felsenstein,
They don’t insist it can’t be there. They theorize that if it is intelligent design, it might not be “junk” at all, but a part of the design.
Richardthughes,
No. Not presupposing optimal design; just not assuming that a large portion of it is “junk”.
Precisely. They make a “good design” argument. Otherwise they might just say, “well, maybe the Designer just got frivolous and irritated and tossed in a bunch of junk”. But they don’t say that because to them, as to you, Design must mean good design.
William J. Murray,
Surely ID must predict that it is all functional, or be fine with the current interpretation?
This silly shorthand term ‘Darwinist’ – it would specifically mean someone who considers it true that
(a) Organisms are commonly descended.
(b) The prime mechanism of change in descent is Natural Selection.
In the latter regard, junk DNA was actually resisted by ‘Darwinists’ when first proposed c1972. But the mathematical argument proved compelling, so now it is broadly accepted. But still, many ‘Darwinists’ harbour a suspicion that this stuff is costly, and hence would be eliminated by NS but for its other benefits. A ‘Darwinist’ research program in support of this would be indistinguishable from your ‘ID’ one.
Not really, William. Lack of junk is a positive prediction of ID, apparently, but junk is not a positive prediction of evolution.
Evolution doesn’t fail if there is no junk. ID has a hard job explaining it if there is.
But in any case, it is notoriously difficult to prove a negative, and the whole “junk DNA” story is largely ID hype – evolutionists knew that much non-coding DNA was not “junk” long before anyone used the word “junk”.
Much more to the point is that we know that pseudogenes exist, and those pseudogenes, map beautifully onto phylogenies showing just when and in whom they lost the function they still perform in other species (the broken GULO gene is some primates is an example). Now sometimes a pseudogene may acquire a new function – and perhaps even back-mutate into its old function. But that doesn’t mean it wasn’t “junk” during the time when it was doing nothing (although IDists might think it evidence of front-loading, which would be a reasonable approach for IDists to take to explain apparently non-functional DNA).
But in any case we know there are vast swathes of genomes that do nothing, because if you remove them, it makes little or no difference to the phenotype. That’s why we keep raising the issue of ONIONS.
WJM
Splutter! Now I must retrieve my monocle from my G & T.
I appreciate those of my betters who can hold their noses and do a bit of slumming. Perhaps they will drop a dime in the cup while they are here.
I will only say that I have been posting at sites run by creationists and ID supporters for 15 years. I have been called dishonest, retarded, ignorant, and many other disparaging things, and I have never once flounced off.
Nor have I returned the invective.
My monocle is back in place. But the irony of a poster from the ID constituency complaining about ‘obfuscation of terms’ when people seek clarity is rich. Half my UD life was spent debating definitions, and I didn’t think it was me being picky. Perhaps we need a set of definitions: cintelligence, cinformation, cnatural, cevolution, cinference, centropy etc to avoid anyone thinking that terms are being adopted in the ‘standard’ way. Or maybe it is me.
UD is like a country club you’ve got no chance of getting into, Allan!
Having trouble posting…
William J Murray said previously (at UD) on FSCO/I
“No. It is rigorously defined, and it can be calculated handily. You can find the definition and reference in the FAQ and Glossary on this site, or by googling “kairosfocus FSCO/I” and finding many exhaustive epxlanations and examples on this site and others.”
Would be mind working through some examples here? I couldn’t find anything satisfactory at UD.
If you cannot find anything “satisfactory” at UD, I’ll not be able to satisfy you here.
False dichotomy.
Could you then link to what you think is the most clear and succinct account?
I must say, the argument that says “all Information (pick your acronym) whose provenance we know comes from Intelligence, and organisms have Information, therefore, organisms come from Intelligence” would seem to violate one of Barry’s Basic Laws of Thought.
Why? If nothing there satisfies you, I certainly cannot do a better job than than the explanations available there.
I mean .. really?
All information of X type that we know the origins thereof are produced by intelligence; therefore, when we find information of X type where we do not know the origin thereof, it is reasonable to infer that intelligence may be the cause.
I’m just asking for a link to the place where you think it is most cogently explained. I’m not saying, a priori, that it doesn’t satisfy me. I won’t know it does until I’ve seen one!
Intelligence in the form of evolution.
The solar system can’t do a random search and come up with life, because there are more possible arrangements of 500 bits than there have been Planck times.
Lizzie has made it clear that she is using Dembski’s definition of information, so speaking of “information of X type” makes no sense. Further, it has been demonstrated in a thread on this very site that CSI, as defined by Dembski, can be generated by known evolutionary mechanisms. Thus, your statement that all instances of information are produced by intelligence is incorrect, unless you accept Dembski’s definition of intelligence that, as Lizzie points out, includes evolutionary mechanisms.
It would be a perfectly good working hypothesis. Compare
All swans I have seen are white.
I can hear a swan
It is probably white.
How would you test it?
Now, apply the same test to your hypothesis about Intelligence and organisms.
William J. Murray,
Hi William – I may have not found the best example. Could you please link to what you consider to be the strongest case?
Many thanks.
No, it’s not like that, Lizzie. If you want to have that kind of example:
1. White swans exist (ID, at least in humans).
2. This other thing (biological features) looks like a white swan (as per near universal agreement on the apparent design exhibited by biological features).
3. It is reasonable to infer that this other thing might be a white swan.
4. How would you test it?
5. Develop a metric for at least a provisional conclusion that this other thing is also a white swan (genetic comparison, FSCO/I of 500+ bits).
The 500 bits thing is irrelevant unless you know the process by which the bits were ordered.
ID has no candidate for a designer, so it is in the position of trying to prove a negative — that evolution can’t do it.
There isn’t any kind of science that can prove a negative.
So why don’t we skip the irrelevant part. Biologists accept the fact that genomes must be the result of an ordering process. Darwin suggested a process, and for a hundred and fifty years, biologists have been studying the details.
OK, let’s try without metaphors:
We observe that many human artefacts have more than 500 bits of FSCO/I
We observe that all biological organisms have more than 500 bits of FSCO/I
We hypothesise that biological organisms are artefacts.
So far so good. Now:
How do we test our hypothesis?
Clearly not by measuring the FSCO/I of the organisms.
So, how?
There reason it should be clear is there is a known process that changes and orders sequences in populations.
The number of bits is completely irrelevant. The only scientific question is the sufficiency of they known process.
All
information of X typebiological species that we know the origins thereof are produced byintelligencehuman beings; therefore, when we findinformation of X typebiological species where we do not know the origin thereof, it is reasonable to infer thatintelligencehuman beings may be the cause.Right, WJM?
This entire argument with WJM is based on nothing more than different definitions of the word “intelligence”. You guys define intelligence such that it is a natural phenomenon; that it is a property of certain systems, including biological organisms.
WJM defines “intelligence” as an aspect of God. He’s already said so. So when he says something like “it is reasonable to infer that intelligence may be the cause.”, he means it’s reasonable to infer that God may be the cause. And indeed, God may be the cause of anything, whether or not it evolved.
Of course there is no way to demonstrate that an invisible omniscient, omnipotent entity isn’t causing everything.
What science tries to do is demonstrate that regular processes can account for some given phenomenon. Sometimes this isn’t easy and takes a lot of time.
You may well be right, but if William means “god” he should say “god” and not “intelligence”. This is why
we can’t have nice discussionsdefinitions matter and aren’t just quibbles over semantics.Incorrect.
But that’s not what actually happened. Dr. Liddle. What actually happened is the way I ordered the sequence, but since what actually happened doesn’t fit your desired goal, you’ve put the developed testing method (FSCO/I) in for the initial observation of apparent similarity.
IOW, okay they look similar, so how do you test?
Grade its FSCO/I.
Okay, the FSCO/I indicates intelligent design. Now, how do you test?
Etc.
Still hopeful. Thanks!
William J Murray
[…]
[…]
[…]
[…]
Argumentum ad ideologicalblinkersum has to be one of the most feeble, and ultimately self-defeating, weapons in the ID-er’s arsenal. Honest people can simply come to different conclusions. Could I respectfully ask that you give it a rest, already?
test
eta: Hmmm. I can post to Sandbox, but am ‘Forbidden’ in other threads. Try again later.
Over at UD Dennisjones has put forth gene knockout experiments as a killer argiment against evolution. I have a few questions.
How long before a UD regular points out that knockout doesn’t simulate an evolutionary scenario?
How long before a UD regular can understand and explain. why?
Would this make a good TSZ topic?
This used to be one of Behe’s killer arguments.
Start one. You are listed as having author privileges.
I would like to see such a thread.
This approach assumes that FSCO/I is a coherent concept with a mathematically rigorous definition that can be reliably and consistently calculated for an artifact by any objective observer. It further assumes that this metric is an indicator of the involvement of an intelligent agent. None of those assumptions have yet been supported, either by yourself or any other IDC proponent.
If you disagree, please provide links to specific instances of such support.
I endorse this request.
The anti-ID equivalent is the argumentum ad religousblinkersum, but I don’t see as much of it here, frankly, as the other one.
None of each would be nice.
I put them in that order, because you had omitted to provide an operational definition of the feature they share – the features that “look similar”.
You wrote:
And, as I said, I was trying to get rid of the metaphors, the swans and stuff, and drill down to the nitty gritty.
As what is commonly, in ID circles, taken to be the point of similarity between organisms and (some) human artefacts, is their degree of specified complexity, I substituted what is commonly taken to be a measure of this for the whiteness of your swans.
In other words, I revealed what looks like hidden circularity in your logic.
To avoid this charge, please explain in what way you think [some] human artefacts resemble biological organism, and why, specifically, this common feature is the feature that indicates their artefactual origins. If it is their “specified complexity”, you are guilty as charged 🙂 If it is something else – please say what it is.
However, even if is the colour of their feathers, you are still up a gum tree, because:
Because the reason FSCO/I (were it to be calculable, and I have seen no evidence that it is) is supposed to be a “test” for “intelligent Design” is that it is based on the assumption that the only way of producing a thing with FSCO/I is intelligence, and even Dembski concedes that at least for his metric, CSI, an evolutionary algorithm with the right kind of landscape parameters, can produce it. And anyway, I’ve demonstrated it
That leaves open the question of whether such an algorithm can be produced by a non-Intelligent agent, but that’s a different story.
Since I’m not the only person to note this issue today, I would like to take this opportunity to ask William to create a new topic (assuming that Lizzie has been as generous in offering such privileges as she usually is) on FSCO/I.
Richard Hughes has quoted William as claiming, on UD:
That suggests that the material is readily available. Having a clear exposition, with example calculations, all in one place would be a valuable resource and basis for more constructive discussion.
What are your thoughts, William?
P0ker! I’d said sodding p¬ker! I’ve been intermittently trying to post This all bleedin’ day! Then the penny dropped.
P~ker. P~bloody~oker. 😕 What is WP’s beef with that word?
eta: Link no worky!
I didn’t apply my intelligence to locate the ‘anti-target’ in the space of all 5-letter words!
Please. I don’t want to attack a strawman, I’d like to see what an ID proponent regards as a “rigorously defined” and “calculated handily” example, an “exhaustive explanation”, if you will.
I agree, Richard. That’s exactly what I’m asking William to provide as well.
Regarding this, William:
I see you wrote, at UD:
I do find it extraordinary that those who think they’ve found a flaw in evolutionary theory so regularly demonstrate just how little they understand the theory they are attempting to critique. In fact, I assume that the realise they think it’s a flaw is because they’ve missed the point.
And here is a case in point.
No, William, I have not “smuggled in” anything. It is right there, hiding in plain site, for you to take note of, as indeed you did.
What and what “it” is, this thing that I blatantly stuck in there, with no attempt to “hide” it at all, was the fitness function.
Of course I “deliberately chose it”. What else did you think I was going to do? Get gang of monkeys to type my script?
An evolutionary algorithm is one that requires a fitness function – a criterion, or set of criteria by which the algorithm decides what to keep and what to cull.
They are called “evolutionary algorithms” because they are work exactly like evolution, except that instead of the environment providing the fitness function (eagles will eat the least well camouflaged chicks), we write in one ourselves (the computer will eat the virtual organisms with the lowest products).
If I didn’t “smuggle” such a fitness function into my evolutionary algorithm, it wouldn’t be an evolutionary algorithm!!!!! The Fitness Function is the Natural Selection part! Not much point in trying to model evolution if you leave out NS!
It’s like you guys (not to mention your elks) are saying: hey, evolution doesn’t work unless you have Natural Selection! Well, duh, sorry guys, but Darwin got there first 🙂
Note that my his fitness function, unlike Dawkins Weasel, is not identical to the most fit solution. In fact, it doesn’t resemble it at all – in fact, it took me a while to work out (separately) what the most fit solution would be, although I expect most of the mathematicians could see the solution fairly easily.
Morever, as in nature, the link between genotype (the sequence of heads and tails) and phenotype (the virtual organism product of runs-of-heads) is not one-to-one. Different genotypes similar phenotypes, and similar genotypes have rather different phenotypes.
So it’s really a rather nice little algorithm – it works exactly as evolution does, it starts with a randomly generated string with a small amount of Specified Complex Information, and evolves one that is specified and low probability, and therefore High Information, and from its rarity as a specified Target, we can work out how many Bits it has, and demonstrate that it exceeds the threshold for CSI.
In other words, evolutionary assisted search can, as Dembski fully agrees, increase the amount of Complex Information in the Genome, and produce CSI.
That is, of course, because of the information contained in the Fitness function, both in nature and in silicon. That’s why Dembski asks: but where did the Fitness function come from? Well, it comes from the environment. So then he asks: well where did the algorithm come from? And we say, well, we don’t know yet, but Szostak has some good ideas. And then he says, in effect, but Szostak’s scenarios wouldn’t work without organic chemistry, so where did that come from? You need fine tuning yadda yadda.
In other words, there’s nothing wrong with the fitness function – it works, it increases Complex Information in the genome, produces CSI, and is found in nature. That means, that, given self-replication with heritable variance in reproductive success, Information will increase, and CSI will emerge – it doesn’t take Intelligence, it just takes self-replication with heritable variance in reproductive success.
So let’s have less anti-Darwin. Take pot-shots at OOL if you like, but poor old Darwin, who made it perfectly clear that he didn’t know how the first life forms originated, was absolutely right, and my little exercise shows perfectly clearly that evolutionary search can increase the Complex, Information In The Genome. It starts with a little, ends with a lot, and I smuggle in nothing that does not have a direct counterpart in natural evolutionary processes – namely, a “Fitness Function” aka your old nemesis, “natural selection”.
*raises a glass to Charles*
Eric chimes in:
I generate exactly what Dembski specifies. I don’t pretend to generate anything else. But his definition, given in Specification, is perfectly clear – a string of minimum Shannon Complexity, which is one of specified subset of strings from a much larger set of strings, such that the probability of finding it by chance is less than the UPB.
I start with high probability strings, generated randomly by coin-tosses, then subject them to random mutation and natural selection according to an environmental fitness function. The fitness function, unlike WEASEL is not the same as the fittest string, in fact it doesn’t resemble any solution, it isn’t even a string, and doesn’t tell us what the fittest strings are like. It is thus the precise equivalent to some environment challenge, such as camouflage from predators.
Sure. But that’s not how Dembski defined CSI – he defined it in terms of probabilities.
Dawkins didn’t make a mistake, he just made an algorithm that was trivial, in which the fitness function WAS the solution. Mine isn’t even the same sort of beast, and certainly doesn’t tell us what the fittest solutin is.
Of course I’m rewarding sequences that score highest on the Fitness function (or rather penalising sequences that score low. That’s how Fitness functions, as in nature, work. You are the most conspicuous chick, you are the mostly likely to get eaten. Your less conspicuous siblings are the ones that will go on to spread their less-conspicuous-making genes. But I am not “moving” anything “forward”. In any generation, the offspring are as likely to be less succesful than their parents as more. In fact, as in nature, the fitter the population becomes, the more likely it becomes that any offspring will be less fit than the parents, because there are then far more ways of being less fit than your parent than more. So nothing is being “moved” – all that is happening is exactly what happens in nature – the kids that score lowest on the fitness function are culled, leaving the fitter survivors to breed.
Of course you do. That’s why evolution works – except you don’t have to specify a single target (I didn’t) and you can have a vast number of fitness functions (camouflage, speed, size, beak depth, eggshell hardness, you name it), each with their own “Target” region). The “target” in nature is intrinsic – it’s whatever best ensures survival. And, as in my sim, the population will converge upon it. The reason our in silico models work is because the system works.
Nothing is smuggled, and what you think is “smuggled” IS natural selection. Natural selection is the fitness function – its the criterion by which an individual is selected, and it’s “natural” in nature because it’s whatever best ensures survival in the natural environment. With artifical selection (breeding) the fitness function is the breed standard. If you, Eric, really don’t know that natural selection and the fitness function are the same thing, then perhaps you should stop telling people what natural selection isn’t until you do!
Of course. And not only the fitness function came from me, so did the landscape architecture – the various mechanisms of mutation, for instance, the method of reproduction, the starting population. All that came from me, the programmer. And all, apart from the starting population, is exactly analogous to what happens in nature. Sure, there’s information, but that information is easy to track – the fitness function is right there in the environment, killing conspicuous chicks, rewarding well-camouflaged ones.
Tough. Sure it’s no use to anyone, but then what use is a e-coli bacterium to anyone except itself? And Dembski’s definition of CSI doesn’t include a condition that the pattern has to “mean anything”.
It’s a bunch of highly non-random numbers. It’s so non-random that I can predict it with quite a high degree of confidence.
Not by “happenstance” but by systematic measurement against a fitness function in exactly the same way as organisms are measured against environmental challenges and found adequate, or wanting.
Let me summarise just how well it resembles biology:
The genotype is subject to mutation, using point mutations, insertions, deletions, and duplications.
On the contrary, the whole attempt demonstrated very nicely what Dembski has in any case conceded, that evolutionary algorithms are perfectly good at finding Targets. And this has been a very good opportunity to explain yet again, why this accusation that in silicon evolutionary models “smuggle” in anything is so ridiculous – all they “smuggle” in is the very thing that Darwin discovered – the fitness function aka natural selection, in other words, a naturally occuring criterion that affects whether an organism breeds successfully or not. Obviously in silicon models we write them. Nature is perfectly capable of providing them.
As your friendly neighbouring eagle.
PS: I notice that CentralScrutinizer objects that:
Well, of course it is, yes. That’s because I wrote it. But it is the exact counter-part to natural selection, just as Darwin pointed out that artificial selection of farm animals by farmers led to certain traits emerging, and that there was a natural counterpart to this in nature. But obviously nobody’s going to write an evolutionary model that they didn’t, um, artifically write.
And lastyearon provides this (is lastyearon a spoof? I’m never quite sure)
Which is sort of funny.
Spotters badge?