A Search Algorithm, And A Prize

_{Gil Dodgen

March 18, 2009

Intelligent Design}

Share: Facebook; Twitter; LinkedIn; Flipboard; Print; Email

There has been some discussion at UD about computational search algorithms, which is one of my specialties.

Just for fun, I’ve included some C source code here (as a .txt file), which is part of a research project. I’ll send a free set of my classical piano albums to the first person who runs the code and publishes the program output in the comments below, along with a correct guess as to what the ultimate purpose of the search algorithm is.

Please provide the following information: CPU clock speed and compiler used.

EIL members are not eligible.

Comments

Adel said, "Providing a better explanation would be considered a positive, not a negative contribution to science, I think." What FSCI does is show that the proposed explanations for the creation on new complex capabilities are deficient and one should look elsewhere. It does not necessarily point directly at an intelligence but that current explanations fail because there are hurdles they cannot get over. So look for another explanation. And in the wings is an explanation that can explain how to get over the hurdles. As obnoxious as this explanation is to many people, it is always there and the more alternatives that go down in flames the more likely it will be offered as a possibility. Never is it a certainty because there is always the possibility that another explanation will arise and it could be the actual one. But all the current explanations are deficient which is why they are having all these conferences and debate within evolutionary biology. So we cannot say they do not recognize the problem, only that they arbitrarily ignore one solution and then tell the public and students that everything is hunky dory.jerry_{March 20, 2009
March
03
Mar
20
20
2009
09:45 AM
9
09
45
AM
PDT}

gpuccio #67 I did not mean to open a controversy about “importance”. I made the example of pi, which is an important mathemathical object, because anyone who knows mathemathics could agree on that. A random number is a mathemathical object, but it has no known function. The point is that a function must be clearly recognizable, as in the case of pi. I agree that some apparently random number could be a special mathemathical object with a function which we still don’t know, but that’s not a problem. In that case, we have a false negative, and we cannot detect design. But if we find the binary sequence of pi, and it is complex enough (let’s say the first 1000 figures), than we can detect design. I see your point but I don’t think this is very useful without a mathematical reasoning behind the importance. Pi is important because it is the ratio of a circle’s circumference to the radius. We have a reason it is important. What is the measure of ‘importanceness’ for non mathematical objects? I don’t think Chesil Beach is a digital sequence bearing functional information. Why not? If I encoded the beach such that a stone’s size and location were represented by a byte of information we’d notice that this string had a pattern, that is a stone’s location was related to the size. Why isn’t this functional information? This is the output of a sorting process no different to a computer’s sort algorithm. Compare it to a ‘normal’ beach if you’d like. I think this is the heart of what I don’t understand about the concept. Do the waves have a high FCSI then not the beach? That’s easy: any random sequence of nucleotides which does not code for a functional protein. You will object: it could have some other functional specification. But again, we are interested only to specifications we can recognize. False negatives are accepted as part of the game. OK I see that. But what about false positives? Here is what you said originally: 4) There must be no known procedure based on laws of necessity which can output that specific sequence in that system. IOW, the observed sequence cannot be explained by any known mechanism based on necessity. This will lead to false positives not false negatives. How do you filter those out?GSV_{March 20, 2009
March
03
Mar
20
20
2009
09:30 AM
9
09
30
AM
PDT}

Adel: "My difficulty with that goal is its negativity. From my limited readings in the history of science, mainly Thomas Kuhn, it appears that theories are overcome by the encroachment of new theories that explain the data better." the design theory does explain the data better. Falsification of the current (wrong) theory is only a part of it. The positive part of ID includes scientific perspectives about the nature of consciousness and intelligence, the complete or incomplete determinism of observed nature, the relationship between computation and creativity, the nature of life, and so on. All of these are very positive problems. ID, like any other scientific theory, starts from the observation of data: designed objects, the process of design, the nature of designers. It is just a first, important attempt at approaching scientifically facts we have always had under our eyes, and whose appropriate intepretation has been prevented by scientific prejudice and materialistic ideology, through the instrumentality of wrong theories like darwinian evolution and strong AI. ID is not the only approach which is contributing to change that scenario of scientific thought, but its role is certainly fundamental. "Since scientific explanations involve causal chains (and since the source [cause] of design in Intelligent Design theory has been ruled out as an object of study), what new purchase on nature does Intelligent Design theory provide to the scientist?" The source of design is not ruled out as an object of study. That is a common misunderstanding. ID is at present mainly a theory about design detection, and therefore can tell us nothing (or very little) about the nature of the designer, his methods and modalities of implementation of design, his purposes and so on. But that does not mean that all those aspects are not open to scientific inquiry. In an ID scenario, all those aspects can and will be investigated scientifically, but that will require new approaches and methods, and certainly new data. Naturally, nobody can now how far the current or future scientific methods can bring us on that way. We have to try to know that. But ID is the necessary premise to do that. "what new purchase on nature does Intelligent Design theory provide to the scientist?" ID is a new scenario, and one which is infinitely richer than the existing ones. In the current scenarios, reality is interpreted as a completely blind and objective dominion, and all subjective realities (consciousness, purpose, design, intelligence, feeling) are just ignored or interpreted as byproducts of strange and undefined mechanisms. In ID, consciousness, intelligence and purpose are empirically observed and investigated for what they are, an essential and common part of reality, detectable both as designers (humans) and designed objects, and a whole field of reality, biological reality, is shown to share the same characteristics as designed objects. That's a completely new perspective (for contemporary science, at least). And one which will change all the future of science. It is a true scientific revolution, in the sense of Kuhn, and one of an extremely universal nature.gpuccio_{March 20, 2009
March
03
Mar
20
20
2009
08:53 AM
8
08
53
AM
PDT}

gpuccio, jerry, kairosfocus, Thank you all for your helpful comments. I am a slow thinker, so please don't think I have ignored you if you don't see responses from me quickly. But there is one thing that gpuccio said that is simple enough for me to appreciate, because it so candidly asserts the goal of the Intelligent Design project and the part that FSCI plays in it:
The biggest contribution of the concept of FSCI to progress is/will be to allow falsification of the darwinian theory. It’s no small deal.
My difficulty with that goal is its negativity. From my limited readings in the history of science, mainly Thomas Kuhn, it appears that theories are overcome by the encroachment of new theories that explain the data better. Providing a better explanation would be considered a positive, not a negative contribution to science, I think. Since scientific explanations involve causal chains (and since the source [cause] of design in Intelligent Design theory has been ruled out as an object of study), what new purchase on nature does Intelligent Design theory provide to the scientist? Sorry to keep asking questions, but I would also like to know about any other contributions of FSCI to scientific progress.Adel DiBagno_{March 20, 2009
March
03
Mar
20
20
2009
08:15 AM
8
08
15
AM
PDT}

Adel, The concept for CSI is more vague than FCSI or FSCI because it tries to do a lot more. It is very general concept and an attempt to conclude intelligence in the widest range of phenomena possible. As such, this general measure seems a little vague to most because it apparently can only be expressed in some complicated mathematically formulas. I am on record that for evolution this concept should be abandoned because it is not needed for the much simpler subset which is FCSI or FSCI. While the mathematics for CSI can be daunting, FCSI is quite simple and is a subset of CSI that does not need a mathematical description to understand. It is Information - DNA sequences are an example Complex - not compressible but use your own intuitive understanding of complexity and it will work Specifies - the information specifies something else or designates something else. For DNA it is a protein or a RNA polymer. Function - the thing specified is also complex and has a function. In the case of proteins, these are very functional and work together with many other proteins in a systematic way to perform very complicated processes. For example, how does a protein get made and then get to the exact spot in the cell or organism to be functional. It is a process of several proteins working together with the appropriate code implanted in the genome in the DNA. It is a process that boggles the mind in preciseness but all of it is encoded in the genome as if the genome knew ahead just what was expected. There is a measure called FSC or functional sequence complexity that assesses the complexity of the sequence and essentially its rareness. This part of it is a little complicated but in general the main concept is very simple.jerry_{March 20, 2009
March
03
Mar
20
20
2009
07:39 AM
7
07
39
AM
PDT}

GSV (#58): "Can you tell me what information can’t be encoded in bits as a digital sequence?" Any information can probably be encoded as a digital sequence. What I meant is that the information we observe must "already" be encoded as a digital sequence. You certainly understand that it isn't the same thing. "Who or what defines what is ‘important’? How would we know if something is important unless we know everything about it? For example is the fact that DNA is made up of 4 ‘characters’ important or merely an accident of design?" I did not mean to open a controversy about "importance". I made the example of pi, which is an important mathemathical object, because anyone who knows mathemathics could agree on that. A random number is a mathemathical object, but it has no known function. The point is that a function must be clearly recognizable, as in the case of pi. I agree that some apparently random number could be a special mathemathical object with a function which we still don't know, but that's not a problem. In that case, we have a false negative, and we cannot detect design. But if we find the binary sequence of pi, and it is complex enough (let's say the first 1000 figures), than we can detect design. "For example is the fact that DNA is made up of 4 ‘characters’ important or merely an accident of design?" I have not discussed the genetic code as a designed code, because that would require a different approach. For the moment, I would stick to digital sequences, like protein genes. "Can you calculate this for Chesil Beach? An 18 mile stretch of stones sorted by size." I don't think Chesil Beach is a digital sequence bearing functional information. "Can you give me an example of information that isn’t specified? Often it is a good idea to define something that doesn’t meet the criteria to understand the criteria better." That's easy: any random sequence of nucleotides which does not code for a functional protein. You will object: it could have some other functional specification. But again, we are interested only to specifications we can recognize. False negatives are accepted as part of the game. "What about unknown procedures? Before electricity was discovered lightening was the Gods doing battle!" Again, we are interested only in procedures we know. New procedures will certainly be evaluated when they are known. We are making science, not religion. Weare not looking for absolute truth. we are just looking for best expèlanations. Moreover, a procedure of necessity could not be known in details, but still be conceivable in principle. But when no procedure of necessity can be conceived even in principle, then I am afrais that the vague notion that "maybe some day we could see things differently" is really of no interest to the scientific debate.gpuccio_{March 20, 2009
March
03
Mar
20
20
2009
07:25 AM
7
07
25
AM
PDT}

Adel DiBagno: kairosfocus has already answered some of your questions. Anyway, I will add my answer too (which is not essentially different), hoping that it may help. Let's go point by point: "That lies well within the Universal Probability Bound of 500 bits." That's true. In my definition, you may notice that I have not given a threshold. I think the threshold should be chosen according to the context. While Dembski's UPB is certainly appropriate for an universal approach to the problem of design detection, in the biological context I strongly believe that we could reasonably assume a more practical threshold: I propose something like 170 bits (a probability of about 10^-50). That should be more than enough for empirical purposes. Anyway, even with that threshold, oxytocin has a definitely lower complexity. "Does it therefore follow that oxytocin is not designed?" No, it just means that design is not objectively detectable with enough certainty by the EF. It could be, in other words, a false negative. The EF is extremely permissive of false negatives, with the aim of practically eliminating false positives. And it should be very clear that very simple things "can" be designed, but that in that case design cannot be reliably detected by the EF. "I was, and still am, struggling to understand what the concept of FSCI brings to science." The concept of FSCI is fundamental if we want to make a quantitative assessment of the probabilistic credibility of darwinian mechanisms. In that sense, it is absolutely necessary for an evaluation of both darwinian evolution theory and ID theory. Which is, I believe, what we are discussing here. "I think a source of my confusion is not having a clear idea of the distinctions between FSCI and CSI. Would you be so kind as to delineate the differences? " Let's say: CSI (generic definition): any kind of information complex enough and which can be specified in one of many possible ways (to be specified), like compressibility, pre-specification or functional specification. FSCI: any kind of information complex enough and which is functionally specified. For my purposes, I limit the present discussion to digital FSCI. FSCI is a subset of CSI. Digital FSCI is a subset of FSCI. "And explain the contribution of each to scientific progress?" The biggest contribution of the concept of FSCI to progress is/will be to allow falsification of the darwinian theory. It's no small deal. "And I wonder whether the FSCI calculation for oxytocin yields a different value from the CSI calculation. If not, why not? If so, why so?" If we make the calculation according to the method I have shown, there is no difference, because what we measure is the complexity, not the specification, in the sense that the specification is 0 or 1 (no specification = no FSCI, specification: FSCI is measured as its complexity). In theory, we could attribute a coefficient for function, but in general I think it is simpler to consider specification as a binary value. In cases of CSI where the specification is not functional, an explicit method must be given in each case to decide if specification is present or not. In the case of biological function, the method I have given in this post is perfectly appropriate.gpuccio_{March 20, 2009
March
03
Mar
20
20
2009
07:08 AM
7
07
08
AM
PDT}

Adel, Read a biology book. FSCI is the same thing as the central dogma or transcription and translation. It was one of the key insights by Francis Crick just after discovering the double helix. It led to the decoding of the DNA and amino acids. It is just making the analogy that this similar process happens elsewhere in language and computer programming. It was this analogy that led Crick to his hunch that what would be seen in genome was the same as was done in cryptography and language. Now when you see FSCI, think Central Dogma or language or computer programming. It is that easy.jerry_{March 20, 2009
March
03
Mar
20
20
2009
07:07 AM
7
07
07
AM
PDT}

Adel: First, I appreciarte that you are clearly a serious person, so I will take time to help. Go back to Orgel et al, 1973 on. [Cf the UD WACs and glossary.] that is, complex specification distinguishes biomolecules from crystals [high order] and random tars [random, non-functional sequences]. The specification in the original cases is of course both informational and in the context that these molecules do the work of life in the cell. CSI is actually a GENERALISATION from FSCI, i.e it asks about other kinds of specification and the underlying issue of specification in general. That is how we come to Trevors and Abel on three kinds of sequence complexity. It will help you to read this paper. And look very carefully at the 3-d illustrative graph:
(i) orderly sequences are highly algorithmically compressible and show little complexity or algorithmic functionality: e.g. . . . DDDDDDDDD . . . (ii) random sequences are very incompressible algorithmically [you have to repeat them to get them correctly, e.g. the lottery winner combi], but are highly complex [take up a lot of bits to say them], and have low algorithmic functionality: e.g. . . . yhr6edbktyd . . . (iii) Functional sequences a re a bit less complex [languages have in them some redundancy], and are a bit less resistant to compression than random sequences [sometimes you can shorten the program] but are high on algorithmic function: . . . START, aa1, aa2, aa3, . . . ., aa100, STOP, express, fold, and presto you have Cytosine C, say for a frog. [I know I am leaving off important bits and pieces of the story]
Now, too, we have bits as a unit of info capacity. We know that beyond 500 - 1,000 bits, the whole observed universe acting as a search engine is not going to be able to search at random or equiv enough of the configs to make it plausible that chance-based searches are able to get the high degree of contingency implicated. For oxytocin etc, you will need to ask wha tis the info capacity, and so fart we work on GP's numbers, giving 27 bits of functional storage capacity -- a short string length. So, we look at hat in context and say this part is too short for us to say on functional specification + complexity alone, IT is specifically designed. However, it works in a context 9funcitonality is like that: parts come together and work together to give a whole that functions], and that context has so many other units of high complexity and specificity, that we can see the whole is designed. It is likely that oxytocin is simply a small cog in the overall system. As to metrics of CSI, FSC etc, the point is that length has different units, which may for practical pur5poses say the same thing. Since FSC and FSCI are subsets of CSI, metrics to these will be metrics of CSI. The general metric proposed by Dr Dembski in 2005 is applicable, though a bit cumbersome. Durston et al have done a metric on FSC that looks at populations of functional proteins etc, and then uses the h metric of info theory to deduce a number of functional bits. beyond a reasonable threshold, we may infer that the search space is beyond reach of chance processes on the gamut of our observed cosmos. At a much simpler level, we folks here at UD have simply observed that we often use functional bits in our digital age. For instance the screen you are reading probably has 24-bit or more pixel colour depth [8 bits each R, g, B], and is probably far beyond 800 x 600 pixels [480 k pixels]; which together give a working bit capacity on tehorder of 111 mbits. Similarly, a DVD has 4.5 or so gigabytes. Now, just 1,000 bits opens up a config space of 2^1,000 or, over ten times the SQUARE# of the number of quantum states of our observed cosmos over its working life. |the observed cosmos working as a search engine could not sample enough of the space to make us confident that random searches could reasonably get to the shores of the islands of relevant function. But of course designers like us routinely generate that many functional bits and more. So, we have used a simple -- non info theorist's --rule of thumb for hte rest of us. if you see at least 500 - 1,000 bits,a nd you observe a specific function, and you see that his is not credibly a forced outcome, then the best conclusion is that this is functionally specific information of sufficient complexity that you can comfortably infer that the object manifesting it is designed. (The threshold is actually very -- GP would for good reason say excessively -- stringent. The number of C-atoms and other requisites of complex cell type life in our cosmos is a lot smaller than the number of atoms of H and He, and the habitable zones in which searching for e.g. first life could happen are a much smaller sector of the relevant spiral and barred spiral galaxies. Ellipticals are out, irregulars are worse out, and you have to be in a just right zone to have the heavy elements with enough "peace" to have long term environmental stability sufficient for life etc. 75% of atoms are H, and most of the rest is He. And the heavier stuff is not evenly distributed by any means. As to criteria for getting to life like ours on a planet suitable to support hat, the odds get longer and longer as we go along. And that is before you get to first life, which on reasonable amalyses will require 600 k bit6s or so of DNA style information, apart from the info stored in the structure of the cell -- DNA is more a database than a computer.] Trust that helps. GEM of TKIkairosfocus_{March 20, 2009
March
03
Mar
20
20
2009
06:41 AM
6
06
41
AM
PDT}

And I wonder whether the FSCI calculation for oxytocin yields a different value from the CSI calculation. If not, why not? If so, why so? That might help.Adel DiBagno_{March 20, 2009
March
03
Mar
20
20
2009
05:41 AM
5
05
41
AM
PDT}

gpuccio:
ID does not deduce function. We “observe” function in existing sequences. For instance, if we have the sequence of a protein in a genome, and we know from empirical observation that that sequence codes for a specific protein, and that that protein has a specific function, which we can observe in a specific living cell, then we say that the gene is a functional sequence.
Please forgive the ignorance that prompted my question. I was, and still am, struggling to understand what the concept of FSCI brings to science. I think a source of my confusion is not having a clear idea of the distinctions between FSCI and CSI. Would you be so kind as to delineate the differences? And explain the contribution of each to scientific progress?Adel DiBagno_{March 20, 2009
March
03
Mar
20
20
2009
05:30 AM
5
05
30
AM
PDT}

gpuccio on the FSCI of oxytocin:
...the FSCI will be about 27 bits.
That lies well within the Universal Probability Bound of 500 bits. Does it therefore follow that oxytocin is not designed?Adel DiBagno_{March 20, 2009
March
03
Mar
20
20
2009
05:20 AM
5
05
20
AM
PDT}

kairosfocus: I obviously agree with you on the points you raise. I have made it clear that I was giving specific restraints in my definition so that the discussion could be simple and explicit. I prefer to discuss digital information because analogic information poses different problems, and in the end we are interested here especially to genomic information, which is digital. Linguistic information is certainly a form of functional specification (the function is to convey a meaning in a specific language). But, again, genomic information, at least at the level of protein coding, which is the best known, is not linguistic, but conveys purely functional data (the primary structure of the protein). There is probably a linguistic aspect in the genetic code, but for the moment I am not dealing with it here. So, as you know, I always keep in mind the biologic model, because it is simple, clear and absolutely relevant to ID. But again, the definition I gave above is in no way universal, and is absolutely targeted to the biological-genomic context.gpuccio_{March 20, 2009
March
03
Mar
20
20
2009
04:19 AM
4
04
19
AM
PDT}

PS: Chesil beach is great for kotching smooth hounds mon [just like the Palisadoes tombolo . . . see if you can spot where that is . . . which also has some shingling, can be great for jacks and the odd snook or two], but is a manifestation of beach dynamics [chance + necessity working on particulates or bigger fragments in a watery, wavy medium near land], not of FSCI. FSCI is about FUNCTIONING information, and shingle rocks are not driving algorithms, or expressing information in languages or functioning as a purposeful structure that requires precise and otherwise improbable specification of the elements and their relationships. Can you show your working here please? One could argue that life is merely chemical dynamics to counter this assertion of yours, the only way to show this is wrong is to calculate the relevant FCSI.GSV_{March 20, 2009
March
03
Mar
20
20
2009
03:01 AM
3
03
01
AM
PDT}

gpuccio I appreciate the reply. A few questions for you: 1) It is a digital sequence (IOW, it can be read as a sequence of digital values, like the sequence of the results of the tossing of a coin, or the sequence of nucleotides in DNA). Can you tell me what information can’t be encoded in bits as a digital sequence? 2) It is functionally specified. IOW, the sequence, in the appropriate context which can read and use it, can convey the information necessary to realize a well recognizable function. For instance, a sequence of 0s and 1s (as can be read in the tossing of a coin, or any other system which can be read as a sequence of bits) is functionally specified if, say, it corresponds to some important mathemathical object, like pi, or to a sequence of bits which, when inputted in a specific computer, behaves as a functional software. Please note that the function must be explicitly definable, and the context is essential to define it. Who or what defines what is ‘important’? How would we know if something is important unless we know everything about it? For example is the fact that DNA is made up of 4 ‘characters’ important or merely an accident of design? 3) It is complex (unlikely), in the Dembski sense of the word. IOW, the probability of attaining that sequence randomly, in the system where the sequence is found, must be extremely low, lower than a threshold which can be conventionally defined in various ways, in different systems. The important point is that the threshold must be low enough to make the random hypothesis really unacceptable from an empirical point of view. The probability is meant as the probability of the functional set (all possible sequences which retain the defined function). If we can assume an uniform distribution, the probability is simply the ration between the number of functional sequences and the total search space. . Can you calculate this for Chesil Beach? An 18 mile stretch of stones sorted by size. In order to keep the length of my posts acceptable, I would stop here for the moment. I want to remark that the above definition is relative to a very specific subset of the more general concept of CSI. CSI is a concept referring to “any” information which can be in “any” way specified. In my definition, I have given many additional restraints (the information has to be digital and not analogic, the specification must be of the functional type), because those restraints allow us to go with very simple, objective and quantitative definitions, and they apply perfectly to the object of our discussion (biological information). So, in the biological field, nothing important is lost by those restraints, and much is gained in clarity and objectivity. Can you give me an example of information that isn’t specified? Often it is a good idea to define something that doesn’t meet the criteria to understand the criteria better. 4) There must be no known procedure based on laws of necessity which can output that specific sequence in that system. IOW, the observed sequence cannot be explained by any known mechanism based on necessity. What about unknown procedures? Before electricity was discovered lightening was the Gods doing battle!GSV_{March 20, 2009
March
03
Mar
20
20
2009
02:57 AM
2
02
57
AM
PDT}

PS: Chesil beach is great for kotching smooth hounds mon [just like the Palisadoes tombolo . . . see if you can spot where that is . . . which also has some shingling, can be great for jacks and the odd snook or two], but is a manifestation of beach dynamics [chance + necessity working on particulates or bigger fragments in a watery, wavy medium near land], not of FSCI. FSCI is about FUNCTIONING information, and shingle rocks are not driving algorithms, or expressing information in languages or functioning as a purposeful structure that requires precise and otherwise improbable specification of the elements and their relationships.kairosfocus_{March 20, 2009
March
03
Mar
20
20
2009
02:31 AM
2
02
31
AM
PDT}

GP; I add that sometimes, function is structural or linguistic, though the algorithmic or data structure forms are the easiest to understand. In the linguistic case, as has already been repeatedly raised, say 143 or more ASCII characters in contextually responsive, more or less correct English would be a reasonable threshold for FSCI. A good structural case -- and note, life for biology is "defined" in light of known major cases and family resemblance thereto [so much for the "that's woolly" and refusing to look at descriptive statements tied to key cases as yardsticks objections . . . -- would be a digital form of the specification for say a functioning stone arrowhead. This case brings out the observation side: arrowheads must be shaft-able, must not overbalance the arrow, and must nor derange flight [e.g. by being excessively asymmetrical or planing in the air]. On hitting say a deer they must be capable of crippling wounds. An arrowhead is not usually designed digitally, but the information is digitisable. And, the odds of a natural stone meeting the above criteria is utterly so remote as to be practically impossible, though logically possible. many shapes are possible of course, but hey will all fit into an island or at least an archipelago of function, even flu- flus for birds. [Birds being fragile you use a blunt ended arrow, often with a puff at he trailing end not normal fletching.] The island or archipelago of function is tiny in the sea of possible configs that are produced by natural forces of chance + necessity. but, archaeologists, professional or amateur, routinely identify arrowheads from natural stones of similar size. And, if we find an arrowhead on Mars, we will instantly recognise it form its functionally specific complex information linked to the achievement of an intelligent purpose. GEM of TKIkairosfocus_{March 20, 2009
March
03
Mar
20
20
2009
01:58 AM
1
01
58
AM
PDT}

Sal: we have been here before. I can only answer you as I have always answered. 1) It's not me who am setting intelligent designers loose to recognize intelligent design. Intelligent designers, in the form of humans, have been there for a long time before I was born, and have always recognized intelligent design. Intelligent designers are an empirical fact, and not a philosophical abstraction. That's, probably, why you don't like them. But I suspect that you are an intelligent designer too, though in disguise as an intelligent deviser of abstract speculations. 2) I have no circularity. And I do have a definition. My definition is empirical, and can be applied to empirical facts. You know, that's what we do in science. I know you have not a high opinion of science, and we have many times shared common reservations about the extreme and blind faith in science which has polluted our culture, but as you probably know I retain a great respect for good science, and good science is what we are trying to do here. If you are not interested in that, I have not much to add. 3) The definition is a definition. It is not circular. I have been accused of circularity many times about my definitions, but as soon as we go into details about where that circularity would be, no one has ever been able to show it. So please, as I have great respect for your intelligence, do try. The definition is not circular because it starts from empirical concepts, each one as explicitly defined as possible, and it defines a specific form of observable information. I suspect that you think that my definition has metaphysical implication, or that in itself it should be evidence of intelligent design. That is not true. My definition of FSCI is exactly what I have said it to be: a definition, and operational tool. It serves to be used in a more general argument, which I have not yet made on this thread. That more general argument is the basis for the design inference. And, if and when we arrive at it, you will see that it is not circular, too. So, to sum up: my definition is only a definition, a tool to recognize what complies with the definition and what doesn't. The only critic you can do is to show that, as an empirical tool, it does not work: for instance, that there are empirical objects, of the form requested by the definition, which cannot be categorized according to the definition itself. If you can do that, please be my guest. And if you are still convinced that there is circularity in the definition, please show where it is. We are here to discuss.gpuccio_{March 20, 2009
March
03
Mar
20
20
2009
12:05 AM
12
12
05
AM
PDT}

gpuccio, So, as I suspected, intelligent design is in the eye of the intelligent designer. Dembski has the sense at least to try to define specification in terms of restricted agents.
Since the semiotic agents that formulate patterns to eliminate chance are finite rational agents embodied in a finite dimensional spacetime manifold (i.e., our universe), there are at most countably many of them, and these in turn can formulate at most finitely many patterns. Accordingly, the totality of patterns that S’s cognitive apparatus is able to distinguish is finite; and for all such rational agents embodied in the finite dimensional spacetime manifold that is our universe (whether it is bounded or unbounded), we can represent the totality of patterns available to such S’s as a sequence of patterns T1, T2, T3, ... This sequence, as well as the totality of such agents, is at most countably infinite. Moreover, within the known physical universe, which is of finite duration, resolution, and diameter, both the number of agents and number of patterns are finite.
I believe there is paradox, even in Dembski's approach, because he gets compact descriptions from agents (each has a cognitive apparatus and is rational, though he eschews mention of intelligence) which must be of high CSI. He does, however, try to define the effect of a higher type of intelligence in terms of a lower type. It seems that you feel no compunction about predicating the existence of intelligent design from the get-go, and then setting intelligent designers loose to recognize intelligent design. You have only circularity and begging of the question -- no definition.Sal Gal_{March 19, 2009
March
03
Mar
19
19
2009
06:33 PM
6
06
33
PM
PDT}

Adel DiBagno: As you can see, I did my homework before reading you request :-) You ask: "And if I wouldn’t be presuming too much, I would like to know how does ID deduce function given only a specific sequence?" I don't know if I understand you question correctly. ID does not deduce function. We "observe" function in existing sequences. For instance, if we have the sequence of a protein in a genome, and we know from empirical observation that that sequence codes for a specific protein, and that that protein has a specific function, which we can observe in a specific living cell, then we say that the gene is a functional sequence. We have to give an explicit definition of the function and a threshold for it only in order to be able to quantify the FSCI, because we have to define quantitatively the functional subset. Obviously, our measurement will depend on our definitions, but given a certain definition, the measurement is explicit (although not necessarily easy).gpuccio_{March 19, 2009
March
03
Mar
19
19
2009
02:56 PM
2
02
56
PM
PDT}

Arthur Smith: "An example or two to help us join up the dots would be helpful. For instance, can youmeasure how much FCSI there is in a molecule of the 9 amino acid polypeptide, oxytocin?" Well, while I wait for comments on the theoretical part, I would like to answer that just to jump to a practical application. I cannot answer your question with absolute precision (I would need to do extensive research to do that), so I will give an answer based on some assumptions, which can be true or not, but which can certainly be verified. 1) We have to define explicitly a function for oxytocin, and to provide a measurement of that function and a threshold based on that measurement in order to define if the function is present or not in a specific system. Let's assume, for instance, that we define as function the ability to determine a certain level of uterine contraction in a certain percentage of patients, at a certain dose. 2) We have to know how many sequences bear that function according to our definition of it. To make calculation easier, I will give here another simple restraint: we will consider only sequences of the same length as oxytocin. 3) The functional complexity of oxytocin will be the ratio of that number of sequences to the total number of sequences of that length. We assume here an uniform distribution of the random generation of aminoacid sequences in a biochemical system. That would require further discussion, but for the moment I will go on. So, if we assume that the only sequence of 9 aminoacids which retains the function as defined above is the molecule of oxytocin (which may be true or not), then the probability of oxytocin will be: 20^-9, which is about 2^-27 If we express FSCI as -log(probability) then the FSCI will be about 27 bits. (I apologize for my goofy mathematics, I hope the concept is clear enough).gpuccio_{March 19, 2009
March
03
Mar
19
19
2009
02:50 PM
2
02
50
PM
PDT}

gpuccio, Thank you for your responses. No rush. Good work requires time. Now, would you be so kind as to respond to Arthur's request:
For instance, can youmeasure how much FCSI there is in a molecule of the 9 amino acid polypeptide, oxytocin?
And if I wouldn't be presuming too much, I would like to know how does ID deduce function given only a specific sequence?Adel DiBagno_{March 19, 2009
March
03
Mar
19
19
2009
02:43 PM
2
02
43
PM
PDT}

Oops! I am afraid I forgot one of the points which make the definition complete, and which has been well established by Dembski in his EF. I add it here: 4) There must be no known procedure based on laws of necessity which can output that specific sequence in that system. IOW, the observed sequence cannot be explained by any known mechanism based on necessity.gpuccio_{March 19, 2009
March
03
Mar
19
19
2009
02:23 PM
2
02
23
PM
PDT}

GSV, Adel DiBagno, Arthur Smith: Sorry to be late in answering. About FSCI. Just a summary, to begin (the subject has been debated many times, but I am ready to discuss any details you like). I will give a very restrictive and explicit definition, which can be applied to biology. We define FSCI any information which has the following characteristics: 1) It is a digital sequence (IOW, it can be read as a sequence of digital values, like the sequence of the results of the tossing of a coin, or the sequence of nucleotides in DNA). 2) It is functionally specified. IOW, the sequence, in the appropriate context which can read and use it, can convey the information necessary to realize a well recognizable function. For instance, a sequence of 0s and 1s (as can be read in the tossing of a coin, or any other system which can be read as a sequence of bits) is functionally specified if, say, it corresponds to some important mathemathical object, like pi, or to a sequence of bits which, when inputted in a specific computer, behaves as a functional software. Please note that the function must be explicitly definable, and the context is essential to define it. 3) It is complex (unlikely), in the Dembski sense of the word. IOW, the probability of attaining that sequence randomly, in the system where the sequence is found, must be extremely low, lower than a threshold which can be conventionally defined in various ways, in different systems. The important point is that the threshold must be low enough to make the random hypothesis really unacceptable from an empirical point of view. The probability is meant as the probability of the functional set (all possible sequences which retain the defined function). If we can assume an uniform distribution, the probability is simply the ration between the number of functional sequences and the total search space. In order to keep the length of my posts acceptable, I would stop here for the moment. I want to remark that the above definition is relative to a very specific subset of the more general concept of CSI. CSI is a concept referring to "any" information which can be in "any" way specified. In my definition, I have given many additional restraints (the information has to be digital and not analogic, the specification must be of the functional type), because those restraints allow us to go with very simple, objective and quantitative definitions, and they apply perfectly to the object of our discussion (biological information). So, in the biological field, nothing important is lost by those restraints, and much is gained in clarity and objectivity. I will be happy of any comment or discussion on these points,, and then I would go on discussing the application of the above concepts to the quantification of FSCI in functional proteins.gpuccio_{March 19, 2009
March
03
Mar
19
19
2009
02:19 PM
2
02
19
PM
PDT}

tribune7:
We can falsify descriptions of protein sequences and DNA, and we can falsify claims that only things of known design have the same traits as those of protein sequences and DNA.
Do it!Adel DiBagno_{March 19, 2009
March
03
Mar
19
19
2009
01:15 PM
1
01
15
PM
PDT}

Gpuccio Yes please. I too have been asking what is FCSI. Is it a measrable quantity? What units? etc? An example or two to help us join up the dots would be helpful. For instance, can youmeasure how much FCSI there is in a molecule of the 9 amino acid polypeptide, oxytocin?Arthur Smith_{March 19, 2009
March
03
Mar
19
19
2009
12:00 PM
12
12
00
PM
PDT}

David --Sal Gal’s point, I believe, is that Hamlet and Polonius are engaging in a bit of design detection. They are imaginatively supplying designs that are not there — like, Sal Gal suggests, IDers in general. Sincere thanks for the lesson in Shakespeare but the analogy compares an ephemeral opinion to the concrete. You can't falsify what one sees in the clouds. We can falsify descriptions of protein sequences and DNA, and we can falsify claims that only things of known design have the same traits as those of protein sequences and DNA.tribune7_{March 19, 2009
March
03
Mar
19
19
2009
07:39 AM
7
07
39
AM
PDT}

gpuccio:
I think you know even too well that many of us actually work with a very clear and detailed definition of FSCI...
It would be a kindness, sir, if you would provide that definition for those of us who need to be reminded.Adel DiBagno_{March 19, 2009
March
03
Mar
19
19
2009
06:17 AM
6
06
17
AM
PDT}

gpuccio: “I think you know even too well that many of us actually work with a very clear and detailed definition of FSCI” There is a beach I saw on a recent BBC program called Coast called Chesil Beach. This is an 18 mile arc of steep shingle and at the western end (Bridgeport) there is fine shingle and as you travel east the shingles get larger until you reach Portland where we have stones that are fist sized. What is the FSCI of this beach? As I see it this beach has a high FCSI which is odd as this beach was not formed by an intelligent agent. Can you help me here?GSV_{March 19, 2009
March
03
Mar
19
19
2009
04:03 AM
4
04
03
AM
PDT}

Gil: “Check out DNA, mRNA, ribosomes, and protein synthesis. If the correspondence isn’t obvious, I can’t help you.” I would like to point out this isn’t a very good way to get your message across, RTFM without even pointing out a manual is annoying. Here are your original questions about your program: “How much time, and how much in the way of probabilistic resources would be required to evolve this program by random variation and natural selection, in a step-by-tiny-step fashion? Could this be done? What is the likelihood of improving it by randomly altering the code?” Taking them one at a time: I have no idea do you? How could you calculate this? How does one measure ‘probabilistic resources’ for this program? What does evolve mean for non replicating streams of 1s and 0s? What is the means of evolution for this program? Does the environment, effectively the OS in the computer, change also? Artificially yes of course it easy to write some code that changes, say, one character at a time and then tries to compile the code, any compilation errors result in non fit code. Anytime the code does compile we have a fit program for some measure of fitness, by checking the output we could measure if it was fitter. Given an infinite amount of time and a measure of fitness after each random alteration that discards the less fit and keeps the most fit the likelihood is 1 you will get an improved program.GSV_{March 19, 2009
March
03
Mar
19
19
2009
03:36 AM
3
03
36
AM
PDT}

Prev 1 2 3 4 5 Next

You must be logged in to post a comment.

Leave a Reply