On The Calculation Of CSI

_{March 25, 2011

Intelligent Design, Mathematics}

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

My thanks to Jonathan M. for passing my suggestion for a CSI thread on and a very special thanks to Denyse O’Leary for inviting me to offer a guest post.

[This post has been advanced to enable a continued discussion on a vital issue. Other newer stories are posted below. – O’Leary ]

In the abstract of Specification: The Pattern That Signifies Intelligence, William Demski asks “Can objects, even if nothing is known about how they arose, exhibit features that reliably signal the action of an intelligent cause?” Many ID proponents answer this question emphatically in the affirmative, claiming that Complex Specified Information is a metric that clearly indicates intelligent agency.

As someone with a strong interest in computational biology, evolutionary algorithms, and genetic programming, this strikes me as the most readily testable claim made by ID proponents. For some time I’ve been trying to learn enough about CSI to be able to measure it objectively and to determine whether or not known evolutionary mechanisms are capable of generating it. Unfortunately, what I’ve found is quite a bit of confusion about the details of CSI, even among its strongest advocates.

My first detailed discussion was with UD regular gpuccio, in a series of four threads hosted by Mark Frank. While we didn’t come to any resolution, we did cover a number of details that might be of interest to others following the topic.

CSI came up again in a recent thread here on UD. I asked the participants there to assist me in better understanding CSI by providing a rigorous mathematical definition and showing how to calculate it for four scenarios:

A simple gene duplication, without subsequent modification, that increases production of a particular protein from less than X to greater than X. The specification of this scenario is “Produces at least X amount of protein Y.”
Tom Schneider’s ev evolves genomes using only simplified forms of known, observed evolutionary mechanisms, that meet the specification of “A nucleotide that binds to exactly N sites within the genome.” The length of the genome required to meet this specification can be quite long, depending on the value of N. (ev is particularly interesting because it is based directly on Schneider’s PhD work with real biological organisms.)
Tom Ray’s Tierra routinely results in digital organisms with a number of specifications. One I find interesting is “Acts as a parasite on other digital organisms in the simulation.” The length of the shortest parasite is at least 22 bytes, but takes thousands of generations to evolve.
The various Steiner Problem solutions from a programming challenge a few years ago have genomes that can easily be hundreds of bits. The specification for these genomes is “Computes a close approximation to the shortest connected path between a set of points.”

vjtorley very kindly and forthrightly addressed the first scenario in detail. His conclusion is:

I therefore conclude that CSI is not a useful way to compare the complexity of a genome containing a duplicated gene to the original genome, because the extra bases are added in a single copying event, which is governed by a process (duplication) which takes place in an orderly fashion, when it occurs.

In that same thread, at least one other ID proponent agrees that known evolutionary mechanisms can generate CSI. At least two others disagree.

I hope we can resolve the issues in this thread. My goal is still to understand CSI in sufficient detail to be able to objectively measure it in both biological systems and digital models of those systems. To that end, I hope some ID proponents will be willing to answer some questions and provide some information:

Do you agree with vjtorley’s calculation of CSI?
Do you agree with his conclusion that CSI can be generated by known evolutionary mechanisms (gene duplication, in this case)?
If you disagree with either, please show an equally detailed calculation so that I can understand how you compute CSI in that scenario.
If your definition of CSI is different from that used by vjtorley, please provide a mathematically rigorous definition of your version of CSI.
In addition to the gene duplication example, please show how to calculate CSI using your definition for the other three scenarios I’ve described.

Discussion of the general topic of CSI is, of course, interesting, but calculations at least as detailed as those provided by vjtorley are essential to eliminating ambiguity. Please show your work supporting any claims.

Thank you in advance for helping me understand CSI. Let’s do some math!

Comments

I wish to propose a rigorous method to calculate the "increase in CSI" represented by the original four proposals. The principle is general and applies equally to all four. In each case you must identify the minimum set of instructional changes in the "source code" or "instruction set" necessary to move the solution from a state lacking the ability to produce the specification to the state containing that possibility. The CSI is the amount of change required, measured in bits of "atomic" changes in the code; for instance in the genetic code this would be any kind of mutation that can occur as a single-step operation, e.g. substitution, deletion, addition, maybe even duplication of section of code. The "bits" of change is the chance of the particular change required occuring out of all the possible atomic changes; equal to the negative log to the base two of the probability of that change out of all possible changes. For the first case, for instance, it may turn out to be just a single point mutation that is required to produce the increase in production of the protein. What are the odds of obtaining that particular mutation out of all possible single-action mutations that could occur? Once that is determined, you can calculate the addition of CSI represented by the change. Maybe that occurs 1% of the time, and represents about 6.5 bits of CSI. It is the choice of change, and the resultant elimination of possibilities that is the hallmark of information. The total CSI is the measure of the accumulation of the choices required to produce your target "specification". Whether it "does" occur, of course is determined by the generation time and number of processes, e.g. the probabilistic resources, and the chances of the survival of intermediate states long enough to get to the required final state. For the other cases, "before" and "after" can all be identified, and the unit of functional change identified, (and is likely defined as part of the program), with probabilities available for the implementation of each type of change at each stage. For digital code this is much easier to do, as you can actually track it step-by-step.SCheesman_{March 24, 2011
March
03
Mar
24
24
2011
04:57 AM
4
04
57
AM
PDT}

F/N: PAV, translation is of course an example of functionality. Even if your string were say the output of an encrypting machine, once we were able to make it function, providing it passed a threshold, it would be FSCI. But, a truly random unspecified string will not generally have linguistic or more specifically prescriptive algorithmic function. DNA has both: it is translated into the mRNA dialect, then used to assemble an AA string in the ribosome, and finally the resulting protein is folded, often agglomerated and activated, then put to work. Again, notice how I refer to empirical reality first and foremost as a basis for being meaningful and accurately connected to the real world.kairosfocus_{March 24, 2011
March
03
Mar
24
24
2011
04:03 AM
4
04
03
AM
PDT}

Does someone think comments are closed? I wouldn’t be here if they were.
I think markf was referring to another thread.Heinrich_{March 24, 2011
March
03
Mar
24
24
2011
03:50 AM
3
03
50
AM
PDT}

MG: I have a crisis coming to a head to deal with, and other matters, so I will pause to simply underscore what I have already said, only to see it again brushed aside without serious consideration. Pardon me, therefore, but I think you are putting the cart before the horse. Pardon, too, the resort to caps to draw attention to the most salient point: 1: UNLESS YOU CAN RECOGNISE AND ACCEPT THE EMPIRICAL REALITY OF FSCI AS OBSERVED, IDENTIFIED AND DESCRIBED IN THE ACADEMIC TECHNICAL LITERATURE BY ORGEL, WICKEN AND OTHERS 30+ YEARS AGO, YOU CANNOT CONSTRUCT OR UNDERSTAND VALID MATHEMATICAL MODELS THEREOF. (And if you cannot see that Orgel and Wicken -- not the easily rhetorically dismissed undersigned -- were not making "meaningless" noise, then no further progress is possible.) 2: Similarly, unless you can accept and understand that information measured in functionally specific bits is a commonplace of life in a digital age [think of the size of say YouTube Videos or the 143 character limit on tweets, i.e 1,000 bits if we work on ASCII 7-bit characters such as are used to type text in posts here at UD and elsewhere], you will be blind to that reality when it confronts you in the guise of DNA, and as the measure of DNA protein coding sub-strings. 3: Long since, too, I have pointed out that we must understand that there are several valid approaches to definition, and that to absolutise one particular abstruse and derivative approach, mathematical modelling, is apt to misconceive and mislead. 4: In short, mathematical models answer to empirical reality; not the other way around. 5: And, I have in fact given a specific, commonplace approach to mathematical measurement and calculation of functionally specific, complex information, without resorting to an infinite regress of demanded further mathematical definitions, and/or circularity. Namely, that for the simple heuristic X = C*S*B:
a: once one can identify a semiotic agent to recognise objectively observable functionality, one can assign a binary value 1/0 to S. (Text in English and functional DNA strings readily meet that criterion. That is we identify a functional macrostate that is met by a cluster of associated microstates.) b: Similarly, on the classic and generally used negative log information storage capacity metric, we can measure degree of complexity, as for n bits, there are 2^n possible configs and with > 1,000 bits, we are beyond a reasonable threshold for successful search by random walks and associated trial and error to detect functionality, on the gamut of the observed cosmos. So, our semiotic agent can count up the number of bits and if beyond 1,000, the complexity criterion is passed, C = 1; not 0. c: Already, we have identified the specific information storage capacity in number of bits, so we have n bits. d: To get the FSCI metric, X functionally specific bits, on this first level model we simply multiply: X = S*C*B. e: For your post at the head of the thread, that is 31, 283 functionally specific bits, and for a typical 300-AA protein we have 1,800 bits. There being of course thousands of such proteins in a typical cell. (The regulatory networks will of course have additional FSCI.)
6: if you cannot acknowledge that the metric X = S*C*B is defined on generally accepted scientific and engineering praxis as well as a wide field of current technology, and provides a simple way to identify and measure FSCI [one that is readily made and allows us to see that FSCI in our direct observation routinely and only comes form intelligence], then there is no basis for discussion of more complex metrics and cases. G'day GEM of TKI PS: I note that the design inference on FSCI or other metrics of Orgel-type observed CSI is an inference on ROOT cause, not on immediate source. This was already discussed by Paley in 1806, when he suggested a self-replicating watch as a further evidence of design above and beyond that which would be evident in the structure and function of a watch: ________________ >> Suppose, in the next place, that the person who found the watch should after some time discover that, in addition to all the properties which he had hitherto observed in it, it possessed the unexpected property of producing in the course of its movement another watch like itself -- the thing is conceivable; that it contained within it a mechanism, a system of parts -- a mold, for instance, or a complex adjustment of lathes, baffles, and other tools -- evidently and separately calculated for this purpose . . . . The first effect would be to increase his admiration of the contrivance, and his conviction of the consummate skill of the contriver. Whether he regarded the object of the contrivance, the distinct apparatus, the intricate, yet in many parts intelligible mechanism by which it was carried on, he would perceive in this new observation nothing but an additional reason for doing what he had already done -- for referring the construction of the watch to design and to supreme art . . . . He would reflect, that though the watch before him were, in some sense, the maker of the watch, which, was fabricated in the course of its movements, yet it was in a very different sense from that in which a carpenter, for instance, is the maker of a chair -- the author of its contrivance, the cause of the relation of its parts to their use. >> ________________ I think that this has much to say to the onward debate on gene duplication etc, but unless there is a willingness to face the implications of the existence of complex organised function and associated information that can be measured in simple and generally acceptable ways, there can be no progress on more complex matters. If one already stumbles at the starting gates, one is not in contention for the prize at the finish line.kairosfocus_{March 24, 2011
March
03
Mar
24
24
2011
03:49 AM
3
03
49
AM
PDT}

Does someone think comments are closed? I wouldn't be here if they were. There is a glitch in the system about which I must speak to our tech man where for some reason you will get that message now and then. Anyway, our policy is to auto close comments after thirty days, and it can't have been twenty-four hours yet .O'Leary_{March 24, 2011
March
03
Mar
24
24
2011
02:53 AM
2
02
53
AM
PDT}

PaV @66 (& SCheesman @58) -
And, a little farther down: Observing some outcome and calling it a specification is a category error. This is what I have been saying for some time, though not as pithily and clearly as SCheesman.
But isn't this what IDers do? How is the specification for the bacterial flagellum defined without observing the outcome? Why isn't any mention made of, for example, cilia? They carry out the same role (making the bacterium mobile). How can an a priori specification be made without observing the outcome, when it's the outcome (e.g. the bacterial flagellum or Mount Rushmore) that is under investigation?Heinrich_{March 24, 2011
March
03
Mar
24
24
2011
02:09 AM
2
02
09
AM
PDT}

UB #31
So the question remains: Does the output of any evolutionary algorithm being modeled establish the semiosis required for information to exist, or does it take it for granted as an already existing quality.
I see that you are returning to your argument that the decisive criterion for information is not the improbability of natural causes but the presence of symbols. You raised this in the previous thread and I think we jointly came to the conclusion that therefore proteins do not contain information because they do not contain symbols. My final comment was a couple of questions:
1) So can you confirm that as far as you are concerned the only part of life to contain information and therefore CSI is DNA. The bacterial flagellum and the immune system are not examples of CSI. 2) Assuming that is true, I assume the symbols in a string of DNA are the bases. What do they symbolise?
Which you never had a chance to answer because comments closed. May be you can pick that up now?markf_{March 24, 2011
March
03
Mar
24
24
2011
01:10 AM
1
01
10
AM
PDT}

2 more cent. Gene duplication may double the amount of some existing CSI. However, simple reason shows that nothing new & novel is produced. To produce CSI, it needs to be more than what you started...and not repeats. I guess one can maybe imagine this as a packet of CSI being value X, and a new packet of equal but differents CSI would be: X^Y (where Y is a positive number). But adding (duplicating) seems like it would maybe something that built up CSI slower like: X+(X/1)+(X/4)...+(X/N^2) I know this seems contradictory to my previous post, but in the previous post, the ring is not a repeat of a smaller CSI packet. It's all just brain storming anyway.JGuy_{March 24, 2011
March
03
Mar
24
24
2011
01:08 AM
1
01
08
AM
PDT}

My 2 cent. I am not fully tracking this discussion yet, but I thought it might be interesting to ask this...and who knows, the concept behind the question might help open up some ideas: When considering an object for CSI, does size factor? ..and.. ifso, how might that help one to approach (think calculus perhaps?) any useful measure of CSI? Example 1: A shiny smooth metal ring that looks like it was a perfectly sliced crossection from a metal cylinder. Imagine finding one that was the size of a pea, then imagine finding one the size of a car... Wouldn't size affect how you perceived the cause of it's origin? Despite bieng perfectly similar, somethigns small seems like it would have less parts to arrange...but somethign large seems even more unlikely. This example just remionded me of the monolith in "2001 A Space Odessey" - one could imagine a "small monolith" as bieng a possible product of nature...but(!) exactly WHY?? And can that reason be used to approach a better way to measure CSI? Thoughts anyone?JGuy_{March 24, 2011
March
03
Mar
24
24
2011
12:48 AM
12
12
48
AM
PDT}

MathGrrl: The answer that SCheesman gives is splendid and spot on. In particular, he says: . . . the examples specified, though appearing to be framed in terms relevant to CSI, are in fact not what CSI measures in the first place, but rather contingent phenomena of more basic processes which more properly might be examined for changes in CSI. And, a little farther down: Observing some outcome and calling it a specification is a category error. This is what I have been saying for some time, though not as pithily and clearly as SCheesman. Namely, and here I'm referencing Dembski's paper on Specification, most critics of ID confuse a "prespecification" with a true, and precisely defined "specification". CSI involves a pattern which "induces" a rejection region. One then applies a probability distibution which captures the circumstances of that pattern, and calculates the level of improbability. If it exceeds the UPB, then you have CSI. In the case of Tierra, e.g., what "specification", i.e. 'pattern' are we dealing with? None. All Tierra succeeds in doing is to be able to use a very simplified assembly-language program, and get the computer to keep computing with it halting. So what. And the only thing that seems to happen is that the length of the computer program lengthens; nothing new, nothing of value, is produced. And the added length seems to come entirely from the original inputted programs, which is kind of like "gene duplication", which, as has been pointed out, is not added CSI. Here's a link to a paper from 2003 that uses a variant of Shannon and Kolmogorov complexity to calculate the increase of complexity over increasing computer time used. Net result. There is some initial complexity built up (60 bits is the maximum info content, well below the UPB), and then NO MORE complexity. I've come to the conclusion recently that the best way to know that you're dealing with CSI is that true CSI is translatable, or convertible. So, e.g., the above sentence, which everyone knows is the product of an "intelligent" agent, and thusly CSI, is translatable from English into Italian. How about this: icks thekskd tj sB Jks t8e gjsklclt shjtle;s';s sleltj a skt elsl;t e;sjsowje ty;slbjhslt;s 'atjelejmt ;sofo. I just typed all of this at random. Can you translate all of that into Italian? So, getting back to Tierra, what can the output of the program be "translated" into? I can't think of a thing, other than to refer it back to the original input. But, then, it becomes a situation wherein no new CSI has been generated. ( And even the original input might not be long enough to constitute CSI. But we can assume so.) Put another way, FIRST a pattern is detected, and then the mathematics applied to the situation surrounding its formation. In the case of biological function, a cell can "translate" DNA into a protein. Hence, since the "pattern" of DNA (i.e., the particular sequence coding for the protein) can be translated into a protein (by, guess what? A translation process), then as far as the cell is concerned, the pattern is "specified". Then, one calculates the improbability. Since it is known that chemically, physically, one nucleotide base is indistinguishable from another, and likewise for amino acids, then one simply takes the length of the sequence of a.a.'s, and the improbability is 20^N, where N = the number of a.a.'s in the sequence. The UPB is exceeded after, what, about 80 to 90 amino acids. Hence, proteins, and the protein sequence of DNA is CSI. Dembski has a sample calculation in his NFL.PaV_{March 24, 2011
March
03
Mar
24
24
2011
12:02 AM
12
12
02
AM
PDT}

vj #39 You seem to be the only person in this lengthy discussion actually addressing Mathgrrl's challenge and you have put a lot of work into it. Thanks. The smiley face is clearly a calculation! But is it based on principles that can be even approximately repeated in another context? There seem to be quite a lot of arbitrary decisions and no guidance as to how to make similar decisions elsewhere. For example: Why only two colours? Why 128 by 128? When calculating the probability why do you assume that all colours are equally likely for each "pixel" and more importantly that the probability of one pixel taking a colour is independent of the probability of its neighbours taking that colour? But most importantly when calculating the probability why did you choose the probability of something that looks like a Smiley Face? You could equally justifiably choose: * looks like a face (smiley or otherwise) * looks like part or all of a human being * looks like part or all of any animal * looks like something we come across regularly in everyday life etc You haven't addressed the problem of an objective basis for choosing a specification which is the whole reason Dembski introduced Kolmogorov complexity in his paper.markf_{March 24, 2011
March
03
Mar
24
24
2011
12:00 AM
12
12
00
AM
PDT}

niwrad (comment #37 above) wrote: >> About gene duplication allegedly increasing CSI >> I would reply with a simple question: >> “when one sends an email twice, do you think the receiver >> gets additional information respect the single mail?”. >> It is obvious there is no additional information. Depends what you mean by "sends an email twice". I'm an IT guy. Duplicate emails are not usually identical. I can examine headers to see whether they were sent by the same mail client, from the same server, with the same unique MessageID, passed through the same set of mail servers, etc etc. Lets say I am able to determine that there is no obvious reason for the duplication (everything looks identical) ... in that situation I am likely to conclude that there is a communications breakdown causing the re-send, or maybe a bug in the software that some developer has written. In other cases I might conclude that the author deliberately re-sent the message ... perhaps an unexpected error at the sending end prompted this intelligent intervention ... or perhaps if the content of the message was such that the sender was clearly anxious, a re-send might indicate to me an increasing level of anxiety on the part of the sender. So perhaps the sending of an email, which is known to be the product of intelligence (as I know of no naturally occurring email-sending systems) is not a suitable candidate for comparison with a CSI candidate where the authorship is disputed.Spiny Norman_{March 23, 2011
March
03
Mar
23
23
2011
11:57 PM
11
11
57
PM
PDT}

Upright Biped @31: “information – any information – only exists by means of a semiotic convention and rules”. Me @62: “I would say that the utilization of semiotics is without question a hallmark CSI.” But I think I just thought of an instance where semiotics is not used in CSI… 3D RNA structures like tRNA, ribosomal RNA and micro-RNA, where the physical medium that stores the information folds into functional structures that are utilized as logic operators or machinery components. In these instances, the base pairs do not need a decoder to be considered CSI. The sequence of the base pairs is still critical (highly specified), because that is what dictates the resultant 3D shape. And the resultant structure is critical in ensuring the maintenance of the system (functional). Hence, we have FSCI, but no semiotics. Although I suppose it would be important to point out that the tRNAs and ribosomal RNAs are actually mechanical components for the translator, or rather the physical manifestation of the semiotics to which you refer! Personally, I find that causing the physical medium that carries the code to serve as both code carrier and components for machinery is SCARY genius. It causes me to join in with the Psalmist and say “we are FEARFULLY and wonderfully made”!M. Holcumbrink_{March 23, 2011
March
03
Mar
23
23
2011
11:34 PM
11
11
34
PM
PDT}

Upright Biped @31: “information – any information – only exists by means of a semiotic convention and rules”. I think I would disagree with that. It seems to me that Shannon information is still information regardless of the existence of a code to make it useful. But I would say that the utilization of semiotics is without question a hallmark CSI. UB: “So the question remains: Does the output of any evolutionary algorithm being modeled establish the semiosis required for information to exist, or does it take it for granted as an already existing quality.” I never thought of that before. Talk about cheating! I think I might be more inclined to believe that evolutionary algorithms have any relevance whatsoever the day they illustrate their efficacy towards the random development of a working language FIRST. That’s the first time I have ever seen that point being made. In light of that, evolutionary algorithms now seem super duper silly to me (whereas before they were just super silly). That aside, I like what Johnson pointed out in Programming of Life: “Ludwig sponsored an “Artificial Life” contest to find the shortest self-replicating program, with the winning program having 101 bytes. The probability of this program arising by chance is 256^-101 or 10^-243. If 10^8 computers each make 10^7 trials/sec for 3x10^22 trials/year, a solution becomes probable after 10^220 years. If a suitable program were half as large, “only” 10^99 years of processing would be necessary to make probable a self-replicating program by chance. It should be noted that any prescriptive program still requires an operational platform on which to execute. These programs, like all computer programs, were designed and were executed on designed platforms. Information, not random data, caused solutions”. I suppose MG is right to tell you that this you have pointed out does not satisfy her original request, but then again, what’s the point? I agree with Joseph (@2) that if chance and necessity can in fact generate CSI, however you quantify it, then ID can be put back on the shelf for most people. But it does not make it moot. Choice is still a very real and likely possibility.M. Holcumbrink_{March 23, 2011
March
03
Mar
23
23
2011
10:17 PM
10
10
17
PM
PDT}

Mathgrll, "To be precise..." Well, of course. You take it for granted because you must. It's a no winner any other way. I had sorta hoped you might actually address the issue though, so I won't ask again. It's kinda like 'piss on the truth' isn't Mathgrrl. As long as one remains satified with their ability to deny the observation, right? Cheers...Upright BiPed_{March 23, 2011
March
03
Mar
23
23
2011
10:12 PM
10
10
12
PM
PDT}

When I was getting my degree in psychology each of my professors stressed that psychology "is a science!" They would say it over and over again. And then they would show how to quantify the amount of self-esteem someone has using a questionnaire. And then when they find that people with high self-esteem are more likely to engage in such and such behavior. While "self-esteem" was a squishy concept, measuring it was useful in predicting future behavior. If CSI is not mathematically rigorous (although I think it is in the end) it still can be useful in detecting design, whether in DNA, computer programs, artificial life, SETI etc.Collin_{March 23, 2011
March
03
Mar
23
23
2011
07:24 PM
7
07
24
PM
PDT}

The central problem of biology is therefore not simply the origin of infrmation but the origin of complex specified inforaion. - page 149 of "No Free Lunch" (bold added)

Algorithms and natural laws are in principle incapable of explaining the origin of CSI. further down on the same page
Joseph_{March 23, 2011
March
03
Mar
23
23
2011
05:36 PM
5
05
36
PM
PDT}

MathGrrl: I don't believe it is possible to come up with any mathematical description of the CSI in the 4 examples you give, but not because the concept of CSI is incoherent, but rather because the examples specified, though appearing to be framed in terms relevant to CSI, are in fact not what CSI measures in the first place, but rather contingent phenomena of more basic processes which more properly might be examined for changes in CSI. 1. Change the focus from duplication of a gene to yield some concentration of a particular protein to the changes necessary in the genetic code or epigenetic action required to accomplish it. It may well be that a single point mutation produces this outcome, in which case the change in CSI is precisely zero, as the before and after state contain the same amount of information. 2. Any computational algorithm, be it ev or Tierra, produces a fixed outcome depending on the initial conditions. Given the same initial conditions, you get the same solution. Observing some outcome and calling it a specification is a category error. All the information, all the CSI is bound up in its instruction set. Anything occurring inside the computer (e.g. the changing world of genomes) does not alter the CSI contained in the program itself. All the possible outcomes, all the possible genomes, were already specified by the program itself before it was even run - the space of possible solutions is not altered by the actual playing out of the code in one or a thousand simulations. You cannot say that sampling the possibilities inherent in the program has added one iota to the CSI of the program itself, nor can they change a byte of its code. Change a byte of the code, though, and the universe of possibilities changes; sometimes not much, sometimes a great deal. Change your scenario to examine the computer code itself. How many extra instructions are necessary to get it to grow, evolve, or destroy a genome, or the modify its interactions with its digital world? That is where you can measure changes in CSI. 3. See Number 2. It may take 1000s of generations to evolve the solution you are looking for, but the evolution is pre-determined by the computer code and the initial conditions of the simulation. The only "specification" in your example is that you happened to pick a pattern you want out of the possible outcomes. By the way, who decided that a parasite was important? Did the computer programme? Of did the programme simply produce this phenomenon and you have decided it is important? Was the programme designed to produce parasites, or was this just an interesting outcome observed after the fact? The Mandlebrot set produces interesting patterns, but they contain no specified information above and beyond the original simple formula and colouring algorithm that produces them. 4. Same issue as 2 and 3. The original code contains the specification and instructions, carefully assembled to produce the effects desired. Conclusion: In order to talk about CSI in a useful way, you need descend to the lowest instructional level, the kernel of information that drives the outcomes observed. None of your examples do this. My recommendation is, if you really want to understand CSI better, refocus your search to the original instruction sets that produce these phenomena. Rigorous mathematical calculation of CSI in that context may not be easy, but it is no longer incoherent or intractable.SCheesman_{March 23, 2011
March
03
Mar
23
23
2011
04:58 PM
4
04
58
PM
PDT}

I'm no mathematician, but I don't think one needs to be one in order to see that gene duplication as an evolutionary mechanism for the origin of information is problematic. Where did the first genes come from? What were they duplicated from? If gene duplication leads to fitness and new novel functions over time, why evolve repair mechanisms that actually try to prevent errors like gene duplication? Why is it, when we see the rapid appearance of new enzymes like in the case of the nylon eating bacteria, we see that gene duplication was not at work?Polanyi_{March 23, 2011
March
03
Mar
23
23
2011
04:30 PM
4
04
30
PM
PDT}

:oops: should be "built-in responses to environmental cues"....Joseph_{March 23, 2011
March
03
Mar
23
23
2011
04:27 PM
4
04
27
PM
PDT}

Intelligent Design is OK with mutations. We just say that some or even most are directed, not random with respect to anything. The point being is that mutations are directed pretty much the same way computer programs direct an output, as spell-checker is so directed. Dr Lee Spetner wrote a book (1997) titled "Not By Chance" in which he discusses the role of "built-in mechanisms to environmental cues" as part of his "non-random evolutionary hypothesis".Joseph_{March 23, 2011
March
03
Mar
23
23
2011
04:23 PM
4
04
23
PM
PDT}

Darwinism, Design and Public Education page 92:
1. High information content (or specified complexity) and irreducible complexity constitute strong indicators or hallmarks of (past) intelligent design. 2. Biological systems have a high information content (or specified complexity) and utilize subsystems that manifest irreducible complexity. 3. Naturalistic mechanisms or undirected causes do not suffice to explain the origin of information (specified complexity) or irreducible complexity. 4. Therefore, intelligent design constitutes the best explanations for the origin of information and irreducible complexity in biological systems.
Joseph_{March 23, 2011
March
03
Mar
23
23
2011
04:11 PM
4
04
11
PM
PDT}

Origins- CSI is about origins. MathGrrl:
That is not reflected in Dembski’s paper referenced in the original post of this thread and it does not help to demonstrate how to calculate CSI.
1- It is reflected in "No Free Lunch" and other ID writings 2- Until we get past your road-blocks it is difficult to continue 3- In comment 12 I linked to a paper discussing what you asked for. Did you read it?Joseph_{March 23, 2011
March
03
Mar
23
23
2011
04:08 PM
4
04
08
PM
PDT}

Congratulations to the powers that be at UD for uncircling the wagons a little.Zach Bailey_{March 23, 2011
March
03
Mar
23
23
2011
04:05 PM
4
04
05
PM
PDT}

QuietID, Is programmer intervention required for spellchecker to work? Is programmer intervention required every time a program runs into a decision?Joseph_{March 23, 2011
March
03
Mar
23
23
2011
04:04 PM
4
04
04
PM
PDT}

kairosfocus, That was a lot of words, but nowhere in there did you provide a mathematically rigorous definition of CSI nor did you show how to calculate it for the four scenarios I described in the original post. Could you please do so?MathGrrl_{March 23, 2011
March
03
Mar
23
23
2011
04:00 PM
4
04
00
PM
PDT}

Niwrad at 37, I must disagree with the view that duplication adds no new information. It often does. Let me tell you a wheeze from the mid-twentieth century that perfectly illustrates that fact: A young lady went into the telegraph office and asked to send a telegram. She gave the operator a piece of flowered stationery with one word on it: Yes The operator explained: Miss, it'll cost you $2.00. You can have ten words for $2.00. She replied, "Certainly not! Nine more yesses will make it sound like I am too anxious."O'Leary_{March 23, 2011
March
03
Mar
23
23
2011
03:59 PM
3
03
59
PM
PDT}

vjtorley,
Your comment on probabilistic complexity was as follows:
This is another term that is impossible to calculate, although in this case it is a practical rather than a theoretical limitation. We simply don’t know the probabilities that make up PC… Computing PC based on known processes and assumed probabilities will certainly lead to many false positives. This version of CSI is therefore more a measure of our ignorance than of intelligent agency, just as Dembski’s is.
In reply: the fact that we don’t know what the probabilities are doesn’t mean that we can’t put an upper bound on them, by computing the probabilities for a wildly optimistic scenario.
Actually, it does mean exactly that. The discovery of a new mechanism could change the probabilities you are calculating to 1. An upper bound implies knowledge that is simply not available, so it remains a measure of our ignorance. From an earlier section of your post:
In your response, you wrote several comments:
While I understand your motivation for using Kolmogorov Chaitin complexity rather than the simple string length, the problem with doing so is that KC complexity is uncomputable.
Quite so.
I find the thought process behind your metric interesting and hope to discuss it further with you. On this thread, however, I am focused on CSI as defined by Dembski and as claimed by ID proponents to indicate intelligent agency. Your metric isn't the CSI that is claimed to support ID. That CSI is supposed to be computable. I would very much like to learn how to compute it. (I will certainly go through the detailed calculation you posted later this evening. I appreciate the significant effort you are dedicating to this discussion.) On a personal note, I hope you and yours are safe and sound.MathGrrl_{March 23, 2011
March
03
Mar
23
23
2011
03:59 PM
3
03
59
PM
PDT}

niwrad,
About gene duplication allegedly increasing CSI I would reply with a simple question: “when one sends an email twice, do you think the receiver gets additional information respect the single mail?”. It is obvious there is no additional information.
The difference is, as noted in the original post and in my post 6, a duplicate gene can lead to an increase in production of a particular protein, with significant impact on the subsequent biochemistry. Such a change in protein production can even enable or disable other genes. The analogy to email or books is fatally flawed.MathGrrl_{March 23, 2011
March
03
Mar
23
23
2011
03:58 PM
3
03
58
PM
PDT}

PaV, I have read Specification... several times. It does not provide sufficient mathematical detail or examples for me to understand CSI to the extent required for me to measure it objectively. If you are able to do so, based on that paper or your other reading, please define CSI with some mathematical rigor and demonstrate how to calculate it for the four scenarios I detailed in the original post.MathGrrl_{March 23, 2011
March
03
Mar
23
23
2011
03:40 PM
3
03
40
PM
PDT}

Prev 1 … 11 12 13 14 15 Next

You must be logged in to post a comment.

Leave a Reply