29th Oct: I have offered a response to Gpuccio’s challenge below.
I think this is worth a new post.
Gpuccio has issued a challenge here and here. I have repeated the essential text below. Others may wish to try it and/or seek their own clarifications. Could be interesting. Something tells me that it is not going to end up in a clear cut result. But it may clarify the deeply confusing world of dFSCI.
Challenge:
Give me any number of strings of which you know for certain the origin. I will assess dFSCI in my way. If I give you a false positive, I lose. I will accept strings of a predetermined length (we can decide), so that at least the search space is fixed.
Conditions:
a) I would say binary strings of 500bits. Or language strings of 150 characters. Or decimal strings of 150 digits. Something like that. Even a mix of them would be fine.
No problem with that.
b) I will literally apply my procedure. If I cannot easily see any function for the string, I will not go on in the evaluation, and I will not infer design. If you, or any other, wants to submit strings whose function you know, you are free to tell me what the function is, and I will evaluate it thoroughly.
That’s OK. I will supply the function in each case. I note that when I tried to define function precisely you said that the function can be anything the observer wishes provided it is objectively defined, so, for example, “adds up to 1000” would be a function. So I don’t think that’s going to be an issue!
c) I will be cautious, and I will not infer design if I have doubts about any of the points in the procedure.
I am not happy with this. If the string meets the criteria for dFSCI you should be able to able to infer design. You can’t pick and choose when to apply it. At the very least you must prove that the string does not have dFSCI if you are going to avoid inferring design.
d) Ah, and please don’t submit strings outputted by an algorithm, unless you are ready to consider them as designed if the algorithm is more than 150 bits long. We should anyway agree, before we start, on which type of system and what time span we are testing.
I don’t understand this – a necessity system for generating digital strings can always be expressed as an algorithm e.g Fibonacci series. Otherwise it is just a copy of the string. Also it is not clear how to define how many bits long an algorithm is. Maybe it will suffice if I confine myself to algorithms that can be expressed mathematically in less than 20 symbols?
And anyway, I am afraid we have to wait next week for the test. My time is almost finished.
That’s OK. I need time to think anyway. But also I need to get your clarification on b,c and d.
gpuccio’s argument is that a string with all the attributes of “dFSCI” loses that designation strictly because of its origin, in other words, if nature can generate DNA with a necessity mechanism, then DNA has no “dFSCI”.
That makes no sense since that is the whole reason for this debate, to determine the *origin* of the “information” that results in life.
You can’t dismiss the “origin” of “dFSCI” simply because you don’t like the source.
If DNA is the result of a necessity mechanism, then a necessity mechanism is its cause, period.
I agree and I always have, it was gpuccio who disagreed.
If “dFSCI” exists independent of cause, you cannot exclude any cause by definition and that includes a necessity mechanism.
If gpuccio cannot exclude a necessity mechanism, his claim that “dFSCI” is an indicator of an “intelligent design”, fails.
With your simple exclusion of a necessity mechanism for the generation of “dFSCI”, your theory can no longer be falsified by natural causes, simply due to your definition of “dFSCI”.
“dFSCI” becomes an attribute of a designer and thus by definition, non-designers cannot generate it.
Eliminating GAs as models of evolution means eliminating all mathematical models, including probability models. One could reasonably criticize specific models for specific reasons, but I don’t see anyone doing that.
Gpuccio277
I am really struggling to understand your comments (I do begin to wonder if there is a language problem after all). So I will try to be very precise and limited in what I say.
Suppose I
(1) Define a function: “predicts whether monthly London temperature anomalies will be positive or negative.”
(2) Present you with a string of 500 bits. I tell you nothing about its origin except to reassure you it was not in anyway derived from the record of London temperature anomalies.
That is the total of all the information I give you.
On investigation you find that the string does indeed correctly predict London temperature anomalies.
Has the string got dFSCI? If not, why not?
I thought gpuccio might be more reasonable than the average IDer, but his attempted defense of dFSCI is descending into silliness:
Yes, by gpuccio’s own definition.
The metric of interest isn’t dFSCI which can never, by definition, be ascribed to the result of non-intelligent mechanisms. The metric of interest is “functional information”.
If the “functional information” as defined by gpuccio is in excess of 150 bits, the string under consideration is a candidate for dFSCI. If an explanation that does not involve intelligent agency is known, dFSCI does not exist. This makes it clear that dFSCI is only a reliable indicator of intelligent design when intelligent design is known to have occurred.
That’s a pretty useless metric.
Gpuccio 278
I suggest you go back and look at the discussion. In particular I wrote:
None of that was clear to me in your original definition and procedure for dFSCI which gave me the impression dFSCI was a property of a string.
I am sorry – I have not been able to read every comment on this thread. What was Mung’s string?
OK. Let’s change the challenge a bit. All your examples are of things that are known to be man-made for other reasons. We all know that text and code are man-made. Life is not. Give me an example where somebody used dFSCI to deduce design and the string was not known to be man-made (remember we are talking digital).
Incidentally, I did give what I thought was a false positive and you said it didn’t count because you didn’t like the function. Now I am trying to pin down what the rules are with a hypothetical example! (I am not saying that your procedure does not work – but I think if you define it so it does work and it is not circular you will find it does not apply to life.)
Gpuccion 281
And finally I hope that my comment above makes it clear this comment is irrelevant. I was not seeking to prove that the London temperature record had dFSCI.
I cam’t believe we’ve wasted this much time on a measure of ignorance.
Gpuccio 283
I mean “a string that tells us whether monthly London temperature anomalies have been positive or negative without having to look them up”
Because otherwise I might just have copied the London temperatures to make the string.
Fair enough. “Connected” was too vague. Let us just say I want to reassure you that the string was not created by simply copying or recoding the London data.
Identical.
I have no idea what your answer will be – I am that confused about dFSCI. But please not only answer but explain why.
How? String by string: how?
And if you specifically exclude copied strings from your method, how have you determined that none of these strings is linked to any other by a copy chain?
I have demonstrated conclusively that a simple history of variation and selection can quickly obliterate the starting sequence and produce separate lineages that have no sequence elements in common with each other.
You don’t get to pick and choose which mathematics you accept and which you reject.
Me too, many times.
I’ve published the complete history of four lineages diverging from a common ancestor — on this thread. It was probably what killed the site.
All the lineages are responding to the same “oracle,” which never varies. I spent a couple hundred hours building this specifically to respond to gpuccio’s claims.
Awesome. I shall check it out.
gpuccio (#285) quoted himself as saying
IOWs, we must know no necessity mechanism that can explain what we observe, before we can affirm dFSCI.
and added
Where am I talking of “origins”? I am only saying that, if we know a credible necessity explanation for the string, we do not affirm dFSCI.
I am beginning to think that a key issue is the phrase “know no necessity mechanism that can explain …”. Many of the people who have argued that gpuccio’s argument that dFCSI implies design is circular have interpreted “know no necessity mechanism” as ruling out the possibility that natural selection acted.
But if (as I think) gpuccio means by “know” the mechanism that we have a detailed and explicit explanation of exactly how and why natural selection acted in the particular case, then things are not so simple. In general, when I see a phenotype, such as the shape of the legs of some species of mite, I don’t know exactly what genotypes are available or what the fitnesses of the genotypes are. Simply because I have never studied that species, and probably no one else has either.
But I do know that for typical species there are genetic variations, and phenotypes that are that easily visible have noticeable fitness differences. So in that sense I “know” that there are evolutionary forces that can in principle explain the phenotypes.
If gpuccio sees a case where we don’t have a detailed explanation for the phenotype being a result of natural selection and random genetic variation, does he consider that we “know no necessity mechanism”? If so, then he would be at risk of inferring design in such cases, when later they might be shown to be explainable by natural selection and random variation.
Or does he mean by that phrase that we know that no necessity mechanism based on natural selection and random variation is in principle possible?
gpuccio,
Of course it does since you exclude one particular “origin” of “dFSCI”, that being a necessity mechanism.
When you say “specification” is derived from observation of function, you have simply painted a target around your arrow.
If I ask you to “design” a “dFSCI” string but don’t tell you what I want it to do, I haven’t “specified” what function I want you to design.
I challenge you to define a string that will provide the functionality I require without first finding out from me what that functionality is supposed to be.
That is the problem the designer faces, i.e., “What information do I need to put into this string to make it do what I want”.
You have stated that a string generated by a necessity mechanism, does not have “dFSCI”, therefore DNA would not contain “dFSCI” , if its origin was a necessity mechanism.
Either “dFSCI” is dependent on its source or it isn’t.
Joe I agree that this phrase is key. Clearly it is a phrase that is open to interpretation. How plausible does a necessity mechanism have to be and how much detail has to be conceived before you can say you know of a necessity mechanism.
In his most recent example Gpuccio has indicated that he knew of a necessity mechanism:
“because it could be copied from the string for London, or be connected to it in some other way”.
without any further detail. This is so vague it applies to life as well and indeed many of the proposed examples of things that have dFSCI. The string of text could have been copied from the Lord’s prayer!
If that’s what he believes, he’s delusional. He’s assuming he has proved a negative in an area where actual research consistently supports the opposite.
So there we have a question for gpuccio:
In deciding whether we “know no necessity mechanism” for an adaptation, must we rule out the possibility of RV+NS? Or must the details of the RV+NS be established before we can decide that a “necessity mechanism” is “known”?
If we have an adaptation sufficiently complex to be a candidate for dFCSI, then
1. In the former case the inference of design is then effectively already made.
2. In the latter case the inference to design is vulnerable to having it later be shown that RV+NS is possible, so the inference to design is not sensible.
I think GP probably feels secure in the knowledge that an historic path through a serial stochastic, chaotic process will not be repeated this side of the heat death of the universe. We are attempting to follow the paths of molecules – admittedly ones which leave a trace, but only amongst survivors. An empirical demonstration of the capacity of RV + NS to generate a specific biological string is practically impossible – there will never be that ‘later’. That does not make dFSCI any more respectable, but it seals it off for its main purpose, which I think is to convince GP.
I’d like to see the methodology Gpuccio applied to Durston’s protein sequences to determine they were designed.
dFSCI (or any variant is never mentioned in the Durston’s paper at all. I’ve asked KF why that is and he made up some blah about it being the same thing anyway. If it’s the same thing why invent another name KF?
The odds against any particular mutation happening are not as great as creationists think. It is quite possible for a large population to test all possible point mutations. And for sexually reproducing organisms, that population includes sperm.
The Lenski experiment suggests that a sequence of mutations can be replicated.
There is also evidence that there are many synonymous or equivalent sequences.
Gpuccio 292
……..
There seems to be serious confusion here. I take it you are referring to my hypothetical example
First I am not sure what the difference is between a data string and any other kind of string. My function was simply “predict whether the London temperature was above average or below”. Would it still be data if I was predicting whether oxygen would be carried in the blood stream or not?
Second the necessity clause does not apply to the London temperatures. It applies to a string which I have told you nothing about. It might be a string of amino acids coded in binary for all you know. So how can any special conditions about data (whatever that means) possibly apply to it?
More confusion. The fact the Lord’s prayer was designed was immaterial. We are talking about whether a string that was created and happened to correspond to the Lord’s prayer was designed. It may help if I change the example.
Suppose you are confronted with a string of digits that happens to be the numbers of the most recent winners of the UK lottery in order. Would you say that string was designed?
Perhaps it would just be easier for Gpuccio to pick an example and show how he worked it out!
Two comments:
The ability of mutations in sperm to “try out” new mutants is not much more extensive than for eggs — for most multicellular organisms the genes would not express themselves unless the sperm succeeded in fertilizing an egg. So the number of trials is the same for mutants in eggs as for mutants in sperm.
As for the dFCSI argument convincing gpuccio, that is not what we are concerned with here. It is whether gpuccio’s argument can be stated in a way that the assessment of dFCSI can be made by anyone else, and whether the argument is then convincing. So far I am still unclear on how gpuccio would have us assess whether dFCSI is present.
The mutation rate for sperm is much higher than for eggs due to many more copy cycles. And purifying selection eliminates anything that is not favorable, neutral or nearly neutral. Certainly this would be true for mutations that affect basic metabolism. You can see this just by looking at he behavior of sperm cells.
There is this basic level of expression prior to fertilization. I’m certainly no expert on this, but I have read that most mutations carried originated with sperm cells. I think there have been several recent papers on this.
I keep bring it up because I would like to see some discussion of it. I’m quite capable of being wrong.
Your determination is based on the requirement of a complete genealogy for every molecule back to the beginning of life. Even if we have a protein with a novel function, and we can trace the evolution of that protein to its precursor, that wouldn’t be sufficient. This renders dFSCI useless to determine whether or not evolution created the sequences at issue.
That’s the whole point of course. Changes in genes result in selectable phenotypic differences.
This conveniently allows you to avoid dealing with phenotypic evidence.
gpuccio and Joe,
gpuccio, while I don’t agree with everything Joe says on this topic, I believe we both have a base agreement on one issue, and that is that “dFSCI”, regardless of its “parent”, exists as a characteristic of a string of information.
That characteristic is simply stated as being a “string of digital information that is complex and contains specific functionality”.
Joe, am I right so far, up to this specific point, that we agree?
Now Joe believes that “dFSCI” has never been demonstrated to be the result of a non-design source while I believe it has, but whatever its reason for coming into being, that specific string, still contains the same “specific functionality and complexity”, that it did before the source of that information is taken into account.
Joe, am I right?
gpuccio, you are clearly stating here that “dFSCI” is NOT intrinsic to the “functional specified complexity” of a string.
If string A has “dFSCI”, then string B, an identical copy of string A, MUST contain “dFSCI”, if “dFSCI” is a characteristic of a string.
Only if “dFSCI” is dependent on its origin/source/generator, can you claim that B does not contain “dFSCI” even though string A does.
gpuccio said:
IOWs, if you have a path for RV + NS, sufficiently detailed and tested, I will accept it gladly. If nobody has such a thing, I will consider that no necessity mechanism is known.
OK, that clarifies things. In cases where RV+NS is not ruled out, but still possible, but where we have not investigated enough to say much further about whether RV+NS accounts for the adaptation, gpuccio considers “that no necessity mechanism is known”. So if the adaptation is sufficiently complex and there is some knowledge of its “function”, gpuccio is then willing to declare dFCSI to be present.
It sounds to me as if petrushka is right: dFCSI is a measure of ignorance. If so, then this is just a Designer Of The Gaps argument.
So if we discover more about a particular case and now find that for a gene for which dFCSI has been declared to be present, RV+NS can account for its evolution, what would gpuccio do?
(1) Would gpuccio then concede that dFCSI is not a good indicator of Design?
(2) Or would gpuccio declare that in that case, dFCSI turned out to have been incorrectly assessed, and that it was still be the case that correct assessments of dFCSI did indicate Design?
There must be something here that I do not understand.
No, but a whole series repeating itself precisely, when we factor in the stochasticity of selection/drift, is negligible. It’s also worth dispensing with a ‘point-mutation’ view. A population can only explore very local space by such means. But grosser mechanisms – particularly recombinations – make much greater leaps in phase space. And again, it becomes impossible to reconstruct the start point. When Mother and Father are long in the grave, along with their relatives, how would you determine which of your genes were recombinant, in the gamete pair that made you?
This is the nature of GP’s challenge to reconstruct RV/NS – the RV part, anyway. But NS is no more tractable. If every individual is equally camouflaged, the gene responsible has no selective advantage. If there is variation linked to survival, selection will operate. But there have to be those ‘others’. How can you reconstruct them – the dead ones – in order to determine selective advantage? Which, as I say, is all very convenient for the professional denier.
I can’t see any scenario in which gpuccio would concede.
From previous discussions at UD and at Mark’s blog, I think gpuccio believes the designer twiddles the bits in a way that is indistinguishable from naturalism.
In other words the dice are occasionally, but not always, loaded. the history we infer is the correct history, but occasionally key mutations are forced by an immaterial designer.
GP is free to correct my understanding of this.
I think I follow the recombination and genome mutation line of reasoning, but the sequences themselves — even the bits and pieces that get recombined — are the result of point mutations. Presumably the point mutations accumulate in idle regions that are not under selection.
It’s true that more mutations occur in sperm than eggs, probably due to the greater numbers of copy events in the chain, and the greater age at which meiosis takes place. A female has done most of the meiosis she is ever going to do while she is still a foetus.
The mutations that we see coming from fathers are the ones that have been through the primary filter of purifying selection on ‘bad sperm’. This tests a relatively minor fraction of genes, since most genes are silent in sperm. But there is another filter: females. We males probably probe a great deal of space, but since we only get to fertilise one egg with our 200-million-sperm ejaculate, this profligacy is a bit irrelevant to the ‘population’ argument.
When surviving sperm and egg fuse to make a diploid genome, some stretches are more likely to contain mutations than others. But this is true within the whole chromosomes from one parent, as well as between the chromosome copies from each – mutation rates are not constant across the genome. The diploid genome simply has x total mutations, unevenly scattered. But when this individual makes gametes, it will recombine these chromosomes, creating chromosomes ‘averaged’ between the paternal and maternal contributions, plus an extra layer of more or fewer additional mutations from the individual itself, depending on whether it is itself male or female.
Weeeell … It’s useful to distinguish mechanisms of variation – but I think it important to realise it’s all just one big shake of the ‘variation tree’. The bits that get recombined are bits that have themselves been recombined, as well as accidental substitutions.
SNP-type point mutation involves a simple substitution, so – say – ACCGGTACG becomes ACGGGTACG. But you could get a single-base insertion or deletion: ACGGTACG or ACTCGGTACG. These are pretty close but also a long way apart. And because they cause a frame shift in protein coding regions, the result can be a major change in the amino acid sequence of the peptide. So tiny non-substitional changes can probe distant space.
Then when we get to the recombinational changes, again we could see ACCGGTACGTACG – a recombinational repeat – or even get a bit of the antisense complement stuck in ACCGGTACG to give ACACCGACG. It’s ‘only’ a recombination, since all that happened was that the strand next to CGGT – GCCA – got flipped round (to preserve 3′ 5′ polarity) and pasted in. Is that a recombination or 4-bases worth of ‘point mutation’? OK, strictly it’s the former, but hopefully you get my drift.
As I understand it both 1 and 2 because dFSCI is a temporary state relative to a function and an observer’s knowledge at that time. Let t0 be the time before the RV+NS was found to be the case and t1 be the time after.
1) Then the gene had dFSCI relative to Gpuccio’s knowledge at time t0 but the string was not designed
2) At time t1 the gene no longer has dFSCI relative to Gpuccio’s knowledge because string was not designed.
This is a very odd concept and easily abused – but it sort of makes logical sense.
What it does mean is that if inspecting strings for which the origin is already known then by definition any string with a known natural origin has not got dFSCI at the time of inspection. So any attempt to correlate design with dFSCI would have to somehow imagine whether the string had dFSCI before the origin was known. But this is an almost meaningless exercise.
Gpuccio
I don’t think there is a “necessity clause” in Durston’s definition which means, unlike dFSCI, his definition is not to relative an observer’s knowledge at a given time.
Mark,
gpuccio is saying that Durston’s FSC is the same as dFSI, not dFSCI. The “necessity clause” applies to dFSCI but not to dFSI.
Perhaps it would be clearer if it were called dIC. Digital Irreducible Complexity. I think what gpuccio is arguing is there are sequences for which there can be no incremental history. Perhaps he avoids making this expllcit because it removes the argument from the realm of mathematics and into the realm of chemistry.
Of course Darwin addressed this possibility explicitely by saying that his theory would be defeated if you could find a feature that couldn’t be reached incrementally.
What I can’t understand about gpuccio is why he thinks he has evidence for this. Lots of people have demonstrated how complexity can arise. And how the history of accrual ccan be erased.
Gpuccio’s “evidence” comes after he has settled on his viewpoint, not before. It did not create his viewpoint, it was created by his viewpoint.
He’s decided that incremental pathways are not available, therefore the evidence he allows himself to see supports only that.
Morton’s deamon on steroids.
I don’t dispute gpuccio’s right to form a hypothesis before looking for supporting evidence. What I object to is the apparent willful disregard for disconfirming evidnce.
He disregards inner ear evolution because we don’t have the molecular history. He disregards detailed molecular histories because they aren’t sufficiently complex. He disregards simulations because the undermine the mathematical basis of his belief system.
Very odd.
You are not applying the label “dFSCI” if the string is the result of certain forbidden processes.
That means “dFSCI” is NOT an attribute of a string, it is instead, according to you, the “intended output” of only “certain processes”.
Without considering anything else about a string except it’s length in bits, I can do the same thing.
Any “pre-specified bit pattern” larger than UPB bits, is unlikely to occur all at once, in less than 2^UPB events.
Why do I even care about the string’s purpose if I really only care about its origin?
Once you disregard all the evidence for evolution, there is no evidence for evolution left.
Therefore ID!
Joe, do you still believe this?
Do you think it is fair for gpuccio to exclude “necessity mechanisms”?
I also don’t care what gpuccio’s motivation is, the issue is whether gpuccio has a method (dFCSI) that can be carried out by other people, and that is evidence that RV+NS cannot explain an adaptation.
So far there are problems with dFCSI.
I am still wondering what gpuccio would have us do when a sequence is designated as having dFCSI because there is not a sufficiently detailed explanation of it by RV+NS, but later such a detailed explanation is found. As I asked:
(1) Would gpuccio then concede that dFCSI is not a good indicator of Design?
(2) Or would gpuccio declare that in that case, dFCSI turned out to have been incorrectly assessed, and that it was still be the case that correct assessments of dFCSI did indicate Design?
It will be interesting to see which of those gpuccio would choose.
In the meantime gpuccio’s dFCSI appears to be an attempt to formalize Michael Behe’s irreducible complexity argument in terms derived from William Dembski’s CSI argument. The attempt does not seem to me to be successful. There are important unanswered questions — so far no one here on TSZ can claim that they know how to apply the dFCSI concept.
I don’t think that dFCSI adds anything to Behe’s argument, and I don’t think we’ll be seeing Behe switch to describing his argument in terms of dFCSI.
But you have to “include” necessity mechanisms in order to prove your argument that non-design mechanisms fail at your specific “dFSCI” test.
We agree that regardless of source, “dFSCI” is an intrinsic property of a string containing “specified complex functionality”.
If you want to exclude “non-design mechanisms”, what is the point of the test?
That’s a good idea.
Firstly, does a string contain “dFSCI” because of its configuration, or does something external to the string determine “dFSCI”?
In other words, is “dFSCI” something a string can possess regardless of its origin?
Joe,
If you exclude nature generating a “specific” pattern out of billions of others, what is the point of gpuccio’s test?
We’re supposed to determine whether two competing agents, X and Y, can generate Z.
gpuccio is by definition excluding X.
Guess who’s going to win! 🙂
I’m reading and digesting this. I still think sperm fecundity exists because it has some function, and I still think it should be investigated as a surrogate for reproductive fecundity.
Probably not the simple thing I presented, though.
Certainly it has a function – or rather several – but I would doubt that mutational exploration is one of them. The issue for any given gene outside the Y is is that it spends pretty much half its time in male bodies and half in female. While in males, it is more likely to suffer a mutation. The question would be whether that is an ‘active’ mechanism – the generation of extra mutations is the reason why there is a mutation-elevating character to male gametogenesis – or a ‘passive’ one – the nature of male gametogenesis is such that elevated mutation rates are an unavoidable byproduct.
I’d go for the latter. DNA polymerase has a per-base error rate, and the more bases you copy, the more errors you make. It is possible that organisms could elevate the number of copies to increase the likelihood of error, but they could just as easily reduce the fidelity of DNA polymerase.
Yet … the mutation rate actually goes down as sperm move along the gametogenetic pathway. That is, some mechanism (and I’m not sure if anyone knows what) is in place to reduce the mutations in germ cells, suggesting that germline mutation is selected against. My own guess is that there is some kind off ‘cross-reference’, perhaps with the Sertoli cells. It’s quite a remarkable thing, that we detect and eliminate a proportion of the tiny amount of amount of mutation in a 6 billion bp genome, but something like that must take place to account for the data.
Nonetheless, the general asymmetry of sex is a really interesting feature of most eukaryotes. There is no genetic asymmetry to speak of, but massive morphological asymmetry. It appears to be the norm that, in multicellular forms, producing many small motile gametes and few large static ones, by binary switch, is more stable than symmetric isogamy. And having gendered individuals producing solely one gamete type is more stable than hermaphrodity. Either way, most genes invest equally in both strategies, alternately or simultaneously.
And obligate diploidy means that ‘pseudo-fecund’ sperm still have to pass through a bottleneck dictated by the availability of eggs, and population limits on diploid somas, which constrains the amount of evolutionary exploration they can do.