Talk:Specified complexity/Archive 1

This is an archive of past discussions about Specified complexity. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

Syllogism(?) of language analogy

Can I say that going from language evolution to biological evolution is syllogistic? I'm not a logician, but it sure seems that, to use a literary analogy, the players have changed even if the plot has not. (That's a syllogism, isn't it?)

A syllogism is a basic deductive process where one deduces that if A implies B and B implies C, then A must imply C. I think what the the transition you're describing is a an example of metaphorical reasoning, which although often effective, has no logical weight behind it. Unknown 22:21, 29 July 2005 (UTC)

Material removed from 'Criticisms' section

I have removed the following from the "Criticisms" section of the article:

Namely, selective pressures overwhelmingly operate not on the specific code an organism has, but on the function that code provides in a given environment. Thus probabilistic treatments of that code are irrelevant without considering the transitional nature of the evolutionary process, the fact that later stages are achieved from previous stages. To assume, as Dembski does, that evolutionary assembly is accomplished as a whole is to assume a non-evolutionary premise.

This argues against something other than what Dembski maintains. He does not assume that "evolutionary assembly is accomplished as a whole". He argues that any natural processes, including evolutionary ones, are fundamentally incapable of generating specified complexity.

This is precisely why it is a criticism. The very foundation of evolutionary theory depends on the generation of complex information. Whether that complex information is confused with complex specified information is the underpinning of any criticism of CSI, except for criticism coming from arguing about his specific numerical methods, which aren't really a criticism of the theory, but only some of the examples in use. --Unknown 15:42, 9 August 2005 (UTC)

Yes, it's criticism, and the basic idea it presents might be sound, but it's nonetheless off-base criticism and so doesn't belong in the article. Dembski doesn't assume that "evolutionary assembly is accomplished as a whole". To include criticism of it would be like including the criticism:

"Due to their relative masses, the earth can't possibly revolve around the moon. This is to assume a non-lunacentric premise" in the article on the heliocentric theory (which doesn't maintain any such thing), or
"Due to the limited nature of the speed of light, the aether can't possibly be moving at an infinite speed relative to the earth..." in the article on the (discredited) theory of the luminiferous aether (which does not maintain any such thing).

--Johnstone 11:45, 19 August 2005 (UTC)

A language analogy is useful: one can communicate "I love you" by saying "I love you" or "I LOVE YOU" or "I <3 U" or "Je t'aime" or any number of methods. In fact, in an environment in which "I love you" is considered to be a default message, no communication at all is needed to convey the intent of the message. The interactions between that which one wishes to communicate, the language or method with which one commuicates, and the interpretations of listeners are similar to natural environments in that language constantly changes and "evolves" to meet the needs of the present. It is for this reason that unused words recede from usage and new ones are coined to meet previously unmet demands on meaning. For example, "Leetspeek" on AOL chatrooms has successfully undermined profanity filters: kids on AOL might understand vulgarity written in language like "ph3aR m3! r0x0rz" but profanity filters do not. Communication is not universally and unchangingly specific; examples of adaptation and evolution in language are commonplace.

The biological analogue of language, i.e. the chemicals, proteins, enzymes, foods, environments, etc. required by life, are not specific, just as the words of language are arbitrary. For example, aerobic organisms by definition need oxygen to live, though there is little specification in general regarding how that oxygen must be supplied. For small aerobic organisms, no real circulation system is required at all, since diffusion works well enough. For larger aerobic organisms, whose inner cells would not receive adequate oxygen via diffusion alone, a circulatory system is needed with special carrier molecules for that oxygen. Though the circulation system is highly complex, it is not in and of itself evidence against evolution.

In the discussion page for the intelligent design article, the consensus is that this analogy with language does not belong, because either it or its criticisms are invalid.

--Johnstone 01:06, 5 August 2005 (UTC)

Removal of information compression analogy, and antifreeze protein example

I have removed the following:

The following examples demonstrate the concept of specified information:

*High information, low specificity. For example, the 10-letter structure "dkownl el." According to Shannon’s theory of information, a random string of letters contains the highest possible information content, because it cannot be compressed into a smaller string. However, the random nature makes the string without meaning, and thus non-specified according to Dembski. (Note that “meaning” does not play a role in Shannon information theory.)

*High specificity, low information. For example, the 10-letter structure "aaaaaaaaaa." The sequence has low information because it can be compressed into a smaller string, namely “10 a’s” . However, because it conforms to a pattern it is highly specified.

*Specified information. For example, the 10-letter structure "I love you". This has both high information content, because it cannot be compressed, and specificity, because it conforms to a pattern (grammar and syntax). In this case, the pattern it conforms to is that of a meaningful English phrase, one of a selection of strings, which together make up a small fraction of all possible arrangements. Similarly, in living things, the “patterns” of functional biological molecules make up only an infinitesimal fraction of all possible molecules.

These examples appear to be contrived; as mentioned in the section above, it has been agreed in the discussion page for the intelligent design article that this is not a valid example.

Cheng, in 1998, described the evolution of various antifreeze proteins in diverse animals, one of which was co-opted from a digestive enzyme called trypsinogen.

This lacks a reference, and may not be appropriate if it depends on co-optation, which is accounted for in Dembski's work.

--Johnstone 11:50, 6 August 2005 (UTC)

A "fraction" is one integer divided by another. As the number of possible molecules is infinite, and infinity is not an integer, the portion of molecules that are functional is not a fraction, miniscule or otherwise. The only way it can be given any meaning at all is through a measure. As all measures are ultimately arbitrary, any argument based on it would be fatally flawed, which is a major problem with CSI; it pretends that arbitrary determinations have some objective basis.Flarity 06:16, 1 July 2006 (UTC)

References are misdirected

The references need to be fixed, recent edits shuffled them around and the WP footnote mechanism is not very robust.--CSTAR 01:23, 17 August 2005 (UTC)

In addition now the concept of pattern in regarde to the reference space Ω is defined after its first use.--CSTAR 01:25, 17 August 2005 (UTC)

incomprehensible

Is there no way to shorten this to a simple explanation, some quotes, and some straightforward criticism? My eyes glaze over after the first couple of sentences. If nothing else, I'll see if I can tighten the criticism. It is extremely vague. A summary and a few good quotes should do nicely. FuelWagon 00:22, 18 August 2005 (UTC)

Could you be a little more specific aboout "incomprehensible"? I would be very happy to say the truth, i.e. the whole enterprise of specified complexity is a crock, but I'm sure that qualification won't last long. As far as explanation one does need to say clearly what Dembski claims he is accomplishing, in as close a language as possible to the language he uses. You can't do this without covering the technical material.

The alternative, to mention criticisms (vague or sharp, it doesn't make any diffference)without an attempted rigorous explanation of the general goals, will invariably lead to yet one more article submerging into the Wikipedia mass of mediocrity.

I hope you don't consider the intelligent design article a good example, with the IDists challenging every adverb (most, many, all scientists, community of scientists etc etc).---CSTAR 00:50, 18 August 2005 (UTC)

An observation on Dembski's approach: while it might not be appropriate to call it "incomprehensible", it certainly can be said its validity cannot be judged by a lay audience -- the most they can do is nod their heads. This is not a problem as such, since science journals are full of such articles -- in evolutionary science, the average formal paper on population statistics is a pretty good example. (In this context the concept of "fog factor", the number of years of education needed to follow along, holds.)

The issue is: who is Dembski's target audience supposed to be? As a layman -- electrical / electronic engineering degree -- Dembski's arguments seem to be targeted towards the community that can understand them, and of no use to anyone like me who can't. Has Dembski ever clarified: Who is the target audience for his pitch? MrG 4.225.211.216 19:21, 8 June 2007 (UTC)

I suspect that "obfuscatory" is a more appropriate adjective than "incomprehensible." Either way, Dembski's objective appears to be to render his arguments unintelligible to the layman, in an attempt to force an 'appeal to authority'-style acceptance. Hrafn42 05:03, 9 June 2007 (UTC)

Yeah ... and it seems very unlikely that Dembski ("Ballerina Bill", whose ability to spin around is actually impressive in a way) will ever allow himself to be pinned down on identifying who he is writing his materials for. He will either have to admit that it is the audience of professionals who are capable of understanding it -- and who have rejected it -- or for laymen -- who he knows perfectly well have no clue as to what he's really saying. MrG 4.225.208.6 13:47, 9 June 2007 (UTC)

Removal of Nowak quote

I have removed the following from the article:

"Martin Nowak, a Harvard professor of mathematics and evolutionary biology asserts that ID "cannot calculate the probability that an eye came about. We don't have the information to make the calculation"^[1]."

...

"Other critics are more direct. As noted in the introduction. Martin Nowak, a professor of mathematics and evolutionary biology at Harvard asserts we do not have the information to calculate the probabilities that Dembski relies on for his estimates,"

#^ Martin Nowak (2005) Time Magazine, 15 August 2005, page 32

It's a very poor quote. At worst, with regard to specified complexity , it's a straw man argument. Dembski has not claimed the eye as an example of specified complexity. It is Behe who has implied that the underlying biochemistry of vision is an example of irreducible complexity. Perhaps unintentionally, the quote confusingly conflates the two concepts and, by using "the eye" as an example, it misleadingly implies that ID is nothing more than Paley's old watchmaker argument which, of course, it isn't. (Indeed, the entire Time article does not accurately present the core ideas of ID.)--Johnstone 11:35, 18 August 2005 (UTC)

"It is a straw man argument". Sorry, you don't get to make that call. wikipedia reports the VIEWS of both sides of an argument. Nowak qualifies as an expert witness based on his credentials, and his view of specified complexity is reported in his quote. You are not in a position to decide that his view is invalid. The only question is whether he qualifies as an expert and whether he is expressing a his view on specified complexity. wikipedia does not report on the topic of "specified complexity" based solely on the views of what it's proponents view it to be. This is what an expert critic thinks it to be, so it is relevant. FuelWagon 13:06, 18 August 2005 (UTC)

You are misrepresenting my position.

I did not say "It is a straw man argument". I said "At worst, with regard to specified complexity, it's a straw man argument." It may not actually be one, since it may very well be due to ignorance, and thus completely unintentional.
I don't think (nor is it true) that Wikipedia reports on the topic of "specified complexity" based solely on the views of what it's proponents view it to be. But Wikipedia doesn't generally strive to include bad examples of criticism.

Because of its historical association with Paley, Nowak's decision to use "the eye" as an example is egregious. Besides making the "expert" critic of ID look uninformed or duplicitous, it will potentially confuse to the article's readers. I'm sure there are other, better quotes out there that would do neither.--Johnstone 11:45, 19 August 2005 (UTC)

The point of Nowak's quote is that these grandiose calculations of probabilities are total nonsense. What could these numbers possibly mean? It obviously can't mean probability in a frequentist sense, since there aren't that many trials possible according to Dembski. A priori probability? Well probability on such a fantastic scale is only understandable from a basic physical theory (e.g. many-worlds interpretation); However, Dembski regards these as "inflatons" of which we should say more later in the article. Indeed etting inflatons is significant new motive for attacks by anti-evolutionists (such as Cardinal Christoph Schonborn) on physical theory. Though calling it a argument from ignorance is not exactly the phrase I would use, I think it is fair to note that (a) It is close enough to what I just said about the probabilities and (b) Critics make that argument. --CSTAR 14:22, 18 August 2005 (UTC)

Of course that's the point he meant to make. He also unintentionally(?) made another.--Johnstone 11:45, 19 August 2005 (UTC)

references

On a completely unrelated topic, is it wikipedia policy to put references at the bottom of the article, rather than embedding the source right where it is quoted? I would much rather see

John stated "blah" (URL)

rather than

John stated "blah" (ref_john)

blah rest of articl blah

ref_john: (URL)

Does anyone else see the problem in having to edit two locations when you insert one quote? It makes it bloody difficult to move quotes to other articles, to make sure that URL's stay with the quote, and even to make sure the URL goes if the quote goes. Just curious. Did wikipedia make an announcement that footnotes are the way to go? FuelWagon 16:59, 18 August 2005 (UTC)

I agree with your point, but I was only following the convention already established for this particular article. This is a pain in the neck. Please change this if you want to redo all the references.--CSTAR 17:02, 18 August 2005 (UTC)

before I go and muck up the whole article, does wikipedia lean one way or the other? do guidelines say use references or embedded links? does it say "whatever the article uses"? FuelWagon 17:43, 18 August 2005 (UTC)

In other articles, I have consistently used embedded references (for external links) and some mechanism such as

Woody Woodpecker (2005), Optimal Strategies for Locating Worms in Pine Forests in Montana

for papers.--CSTAR 17:52, 18 August 2005 (UTC)

Well, no volunteers to do redo the references? --CSTAR 16:56, 19 August 2005 (UTC)

Several glaring absences in the intro and article

SC in relation to ID is not found in the intro, and less knowledgeable readers are going to be left with only part of the picture. Also neither is CSI is mentioned, nor its relation to SC, nor is how they all relate to Dembski's "explanatory filter." I'll be working these into the article. FeloniousMonk 18:29, 23 August 2005 (UTC)

Nice URL

Short, simple, to the point. FuelWagon 19:09, 23 August 2005 (UTC)

writing exercise

CSTAR, I would like to make the suggestion that a sourced quote from an expert should be able to stand on its own without a lot of explanation before or after. While I appreciate your work on this article, it is becoming overbearing in size and inpenatrable in word density. I offer to you a writing exercise. Find a way to communicate the content of this article with half the words. The article will be better for it if you succeed. FuelWagon 01:13, 24 August 2005 (UTC)

General POV

[The following unsigned comment by IP 220.236.179.40 was inserted at the beginning of the talk page. I moved it to the present chronologically justified location to].--CSTAR 16:12, 1 October 2005 (UTC) in The general spirt of this articile is very POVish, a little more defending Dembski would be nice, and some of the criticisms could be phrased more in "he said, she said" style. Consider: When Dembski's mathematical claims on specific complexity are interpreted to make them meaningful and conform to minimal standards of mathematical usage, they usually turn out to be false. I think that consitutes one of the best examples of POV I have ever seen. Why can't writers seem to understand that wikipedia is meant to be NEUTRAL, for gods sake, it's not that diffcult a concept.

The problem is that logic is not neutral. Dembski's definition quite litterally is nonsense. If specified means "compressible" then it actually means "does not contain much total information." If complex means "contains a lot of information" then putting the two together into "specified complexity" implies that Dembski is basing his claims of design on the observation of things which "contain a lot of information, but don't contain a lot of information." Given this, it's no wonder Dembski is so coy about actually mathematically defining anything. And when the article presents a definition of SC which is absurdly self-contradictory in this way, SOME mention has to be made in the criticism section, don't you think? If an article on math started out "Suppose that (x = 1) AND (x != 1)...." don't you think that would be a little bizarre? That's basically exactly what's going on here Plunge 17:32, 5 August 2006 (UTC)

"In Dembski's terminology, a specified pattern is one that admits short descriptions, whereas a complex pattern is one that is unlikely to occur by chance." So a "specified" pattern is one that is of low complexity (in the standard, Kolmogorov sense) and a "complex" pattern is one that is improbable given natural selection. "Specified complexity" then becomes simplicity that is improbable under natural selection. You're misled by his nonstandard terminology, as what he calls "complex" has little to do with complexity (except that simple sequences are usually very improbable, consider e.g. the chance of getting a long sequence of fair coin flips that is simple to describe; but evolution is not even remotely close to fair coin flips so it's not related here, unless you make a rather contrived relation).

As an aside, his claim of "the inability of evolutionary algorithms to select or generate configurations of high specified complexity" seems tautologous to me, as this says in other words that evolutionary algorithms are unable to select simple configurations that are highly improbable under natural selection. Sure, and intelligent designers are unable to design simple configurations that are highly improbable for intelligent designers to make. Both are equally true and vacuous. It may be that he means something else, but in that case it is not clear from the article. Why not state the claims in standard terms and then make a clear exposition of the math? I thought that's what mathematicians do, and the article should do him justice as a mathematician. Coffee2theorems 20:20, 25 August 2006 (UTC)

The example you cite would indeed constitute POV if the assertion it makes were untrue or not found outside of these pages.

But since fellow mathematicians have issues with Dembski's methods and have stated as much publicly (which the article points to), having the article state this is well within the bounds of WP:NPOV. FeloniousMonk 21:25, 1 October 2005 (UTC)

However, it still doesn't make Dembski's model untrue. Basically, his critics use all sorts of deconstruction and shrinking in order to make his thesis appear less valid, or so they think. Because the majority of these so-called "critics" are materialists, and find the idea of ID absolutely out of question, they will do their outmost in order to slander and ignore anything that points to this direction. However, in their obsessive criticism, and wich Wikipedia very much reflects (unfortunately), they can't offer any proofs of their own. Criticism of Dembski's model cannot be proved; it's only a matter of intellectual rhetoric and hypocricy. If every work of an evolutionist would have been as thorough reviewed and criticized as Dembski's, I guess we would have far more progress on the matter.

There are two different issues:

A model
The mathematical content of a model. This includes theorems and definitions.

Dembski could make any model he wanted; the criticism you allude to does not touch on that. However, much of the mathematics is wrong or meaningless. For example, he consistently confuses various measures of "information" such as Kolmogorov complexity and Shannon entropy. Not only are these different, but they apply to different things. It would be as though one confused speed with temperature. Moreover, his mathematical bungling is amply documented by the references.

--CSTAR 15:14, 23 October 2005 (UTC)

The second last paragraph made an interesting comment "if every work of an evolutionist would have been as thorough reviewed and criticized as Demski's". What makes you think the work of evolutionists is reviewed any less thoroughly??? MvH Jan 17, 2005.

Teleological argument?

WHere does Benapgar get the idea that specified complexity comes from teleological argument? Dembski goes to great pains to explain that his argument is entirely based on models of probability and computation plus some facts about physics.--CSTAR 14:58, 7 November 2005 (UTC)

Sorry, my objection was badly formulated. I should rephrase my objection: Dembski claims that existence of specified complexity rules out randomness as a hypothesis. This "ruling out randomness" he claims is an application of a statistical test, as he tries to explain at great lengths. Even if we recognize it is a variant of an old argument, Dembski doesn't seem to think that it is. --CSTAR 17:38, 7 November 2005 (UTC)

It looks like more original research from Ben. I've read Dembski's Design Inference, Design Revolution and Intelligent Design: The Bridge between Science and Theology, and no where does Dembski appeal to the teleological argument in connection to his claim that his explanatory filter, specified complexity, is a reliable marker for indicating design by an intelligent agent. Without a direct quote from Dembksi connecting specified complexity to the teleological argument, it's original research and needs to be removed. FeloniousMonk 18:33, 7 November 2005 (UTC)

Criticisms of specified complexity

A more appropriate title for this article would be "Criticisms of specified complexity". The article should describe the idea, and then the "Criticisms" section should criticise. As is common for wikipedia articles on "pseudoscience" the whole aritcle is a small description of the subject at hand and riddled with criticisms. I was reading the article, which is filled with discreditations of specified complexity and everything Dembski has to say about it, and suddenly I see a section called "criticisms". I think this is unfair. The article should be cleaned up. --ChadThomson 09:49, 30 November 2005 (UTC)

Unless you intend to fix or point out specific points, I'm going to remove the boiler. I count one paragraph of criticism in the intro, and a few lines of criticism outside of the criticism section. Perhaps the wording could be better, though. Please show specifically why you placed this banner, as opposed to trying to fix some things. It appears you have not made a single substantial edit to this article, yet you claim to be fully aware of specific problems. Sofixit. -- Ec5618 10:53, 30 November 2005 (UTC)

Addition by Anon user User:24.35.104.249 moved to bottom of page

I've added the importance of "independence" to the concept of Specified Complexity. No understanding can be complete without it. Brings page in-line with the ISCID defintion available here: http://www.iscid.org/encyclopedia/Specified_Complexity. More specifically...The idea behind specificity is that not only must an event be unlikely (complex), it must also conform to an independently given, detachable pattern. Where "detachable pattern" refers to one that is independent (or not enslaved) to the local environmet.

I have yet to see a reasonable definition of "indepenent" in this context.Flarity 06:08, 1 July 2006 (UTC)

Independence of Prespecifications

Recent edits of anonymous user keep inserting "independence of prespicifications"; my justiftication for removing them is the following quote from the addendum of Specification: The Pattern That Signifies Intelligence August 2005.

"The first question a reader familiar with my past treatment of these concepts is likely to ask is whatever happened to such key notions as detachability, conditional independence, tractability, and universal probability bounds that in the past had characterized my account of specification. They are still here, but they no longer need to be made explicit because of a fundamental simplification to the account of specification that appears in this paper. The simplification results from not demanding that the theory of specification do double-duty as also a theory of prespecification. The problem, as I point out in section 5 of this paper, is that a theory of prespecification can do an excellent job of ruling out the reoccurrence of a chance events, but it cannot rule out whether prespecifications themselves resulted from chance — recall the difference between the sequence (R) and the Champernowne sequence (ψR) in sections 5 through 7. The upshot of these considerations was that a theory of prespecifications can rightly tell you that lightning won’t strike twice in the same place, but it cannot tell you whether lightning that’s struck only once in a given place struck there by chance."

The edits of this anonymous user also include, what are as best I can tell, original phrases such as "processes enslaved to the environment". These are not relevant to the article. --CSTAR 03:30, 8 March 2006 (UTC)

Regarding removal of Independence and Detachability

Recent edits have been removing the repair to include the concept of independent & detachability from the introductory paragraphs for this topic. Specified Complexity is a dual-pronged criterion. A complete definition of Specified Complexity is available at http://www.iscid.org/encyclopedia/Specified_Complexity. ISCID has Mr. Dembski as a fellow, and he is a frequent contributor to the discussion area as well. This link contains the conscise and complete description of Specified Complexity.

Please note the addendum to the article originally cited with a relevant portion already cited on this page. "...detachability, conditional independence, tractability, and universal probability bounds...They are still here, but they no longer need to be made explicit because of a fundamental simplification to the account of specification that appears in this paper. Previous edits are confusing Dembski's specific treatment in regards to a single-prong focus outside of prespecification. This Wiki entry makes no such distinctions, and in the end, the Aug 2005 paper by Dembski is a treatment of only one-half of the entire concept. A complete description of Specified Complexity requires both prongs to be addressed. An important distinction in the early introductory portions of this treatment.

While Dembski does not use the terms "enslaved to the environment," he does describe the notion of detachability. The use of the phrase "enslaved to the environment," is to thematically fit Dembski's notion of detachability within the current structure of this article without the need for a major re-write.

Dembski's 2005 paper is an attempt to respond to earlier criticisms of detachability and independence. As was pointed out by his critics, Dembski was trying to use independence of prespecifications in a way which was never quite satisfactory, apparently not even to himself: there was no natural underlying distribution, though he made various attempts to see if one such probability could be produced. For instance, there was a paper of his on "uniform like" probability measure on metric spaces. Though his current "theory" has problems of its own, I see no reason to confuse further his presentation. If you would like to add the independence condition as part of his earlier characterization, that would be fine with me. It seems to me, you are only confusing Dembski's position by your edits. --CSTAR 22:50, 8 March 2006 (UTC)

Re: "This link contains the conscise and complete description of Specified Complexity." I don't think it does. The main point in Dembski's current published thinking is the use of Fisherian model of statistical hypothesis testing.--CSTAR 22:56, 8 March 2006 (UTC)

I would suggest that what Dembski thinks should take precedence. The concise defintion is taken from the ISCID definition database. An organization for which Dembski was a founder and is still a frequent contributor. If the definition there were wrong, Dembski would have changed it.

The discussion of independence is mentioned later in the article. These edits are to the Introductory section which currently fails to introduce both prongs of the criterion...even though independence is briefly mentioned later.

Dembski's motivations in publishing a more specific examination of the pre-spec vs. spec issue is a subjective interpretation, while the introductory portions of this article should remian true to Dembski's own defintion objectively available at ISCID.ORG None of this is an attempt to advocate for or against Dembski's claims, but merely an attempt to assist in explaining his concepts correctly.

I've changed the wording from "enslaved" to "attached" to bring it more precisely in-line with Demsbki's "detachable" terminology.

Re: I would suggest that what Dembski thinks should take precedence.

Are you Dembski?

Re: Dembski's motivations in publishing a more specific examination of the pre-spec vs. spec issue is a subjective interpretation.

There is little doubt in anything that Dembski writes to suggest that Dembski's quantitative definitions were not intended to supersede the preceding ones

The concise defintion is taken from the ISCID definition database. An organization for which Dembski was a founder and is still a frequent contributor. If the definition there were wrong, Dembski would have changed it. Specified complexity as a numerical concept is not defined there. --CSTAR 05:21, 9 March 2006 (UTC)

Re: since un-guided processes are attached to their local environment.

Sounds like your OR addition on Dembski. What does this mean anyway? --CSTAR 05:24, 9 March 2006 (UTC)

Are you Dembski?

Specified complexity as a "numerical concept?" That appears to be a new qualifier and if the basis for the continued censorship of a critical Specified Complexity concept...then one of too narrow a point of view. There is certainly little doubt as to Dembski's intention, as he himself spells it out.

The first question a reader familiar with my past treatment of these concepts is likely to ask is whatever happened to such key notions as detachability, conditional independence, tractability, and universal probability bounds that in the past had characterized my account of specification. They are still here, but they no longer need to be made explicit because of a fundamental simplification to the account of specification that appears in this paper.

The phase...They are still here... is unambiguous, and clearly stated. And that they no longer need to be made explicit [in this treatment] because of the account of specification in this paper.

Dembski is the Executive Director of ISCID. The definition of Specified Complexity provided there is the definitive definition. Regardless of any "numerical concept" qualifications. Specified Complexity is what it is, and that includes the concepts of Independence and Detachability. The introductory portions of the article presented here mis-represent the full components of Specified Complexity. The concepts of Independence and Detachability must be presented for any serious presentation of the material...especially in the introduction!.

Given the clear objective case of the defintion of Specified Complexity present in the ISCID database of which Dembski is the executive director...it is encumbant upon others to provide sufficient cites and references for the wholesale removal of a foundational element of the Specified Complexity concept.

un-guided processes enact change through interaction with their environment. In other words, they are "attached" to that environment...and any results produced by such processes are therefore attached as well. Detachability refers to specifications that are detached from the environment.

Reply

Re: "The phase...They are still here... is unambiguous, and clearly stated."

It is also clear that the phrase, which you chose not to cite, but they no longer need to be made explicit because of a fundamental simplification to the account of specification that appears in this paper is just as forceful and unambiguous. Your citing the above assertion out of context proves nothing.

Re: Dembski is the Executive Director of ISCID. Is there an identifiable casual chain from that fact (which I haven't checked, but I certainly believe you) and your assertion that the definition provided in that webpage is the "definitive" one (again, I'm not sure that "definitive" here means very much, since that site is not meant to be technical.)

--CSTAR 02:46, 10 March 2006 (UTC)

I have not ignored the entire context of the addendum, as I've specifically pointed out that the treatment in the paper cited is specific to the treatment in that paper. Dembski is specific on that point. Dembski is the executive director of ISCID which contains an ISCID Encyclopedia of Science and Philosophy with a specific defintion of Specified Complexity. A board and forum in which Dembski has been an active participant as recently as last Monday. It escapes me why there is such a need to censure one of the foundational concepts of Specified Complexity. Is there pride of authorship here?

Cerainly one is free to editorialize regarding Dembski's motivations, and perform deeper explorations of Specified Complexity as a "numerical concept" further in the article. However, any such article in Wiki should feel obligated to present the Specified Complexity concept fully as a basis in the introductory section. There is no justification for the removal of such critical components as independence and detachability...even if such elements need not be made "explicit" in "single-duty" explorations of the concept. But the denial that Specified Complexity in it's full incarnation includes Independence is completely un-justified.

On Dembski's own blog, Uncommon Descent, was posted a primer on ID as recently as Februrary 21st 2006. Available here:

http://www.uncommondescent.com/index.php/archives/847#more-847

Containing the paragraph:

Next, the pattern has specification. The pattern conforms to an independently given specification. In this case siblings from the same family is the indendently given specification.

Reply

You added:

Re :"a specified pattern is one conforms to independent requirements,"

Independent of what? The sentence is extremely confusing since "independent" requires a complement (or at least the complement should be understood from the context: as in The US is an independent nation). Is this the case here?

Re : "Cerainly one is free to editorialize regarding Dembski's motivations,"

Is this editorializing done in the article?

Re: "It escapes me why there is such a need to censure one of the foundational concepts of Specified Complexity. Is there pride of authorship here?"

Who's trying to censor? The article goes to great efforts to fairly present Dembski's thinking based on what was his most recent and coherent paper. It seems to me that by your insistence on including the disputed phrase (or some variant), you're actually making the concept appear much less coherent and making the Wikipedia entry less useful. Do you really believe that I'm trying to expunge that sentence to make Dembski look bad or undermine the concept?--CSTAR 23:38, 10 March 2006 (UTC)

Re: "Independent of what?"

This has got to be the sticking point. Independence of specification is one of the primary concepts of Specified Complexity and the theory of ID. The point being that if one cannot understand what is meant by independence, then how can that concept be properly reflected in any discussion of Specified Complexity?

Independent of the local environment. This is the key differentiation of Specified Complexity, than something like Shannon's treatment of information & complexity. As Dembski explains, the rocks of Stonehendge align with Astonomical reference points. The alignment of astronomical bodies of the type to which the rocks of Stonehendge are specified...is "independent" of the forces of gravity, weather, & glacial movement that might otherwise account for a complex pattern of inter-related rocks. Dembski proposes Specified Complexity as an indicator of Intelligence, not because of what he knows about a Designer...but because of what he knows about Nature. Specifically, that nature builds based upon natural selection due to environmental pressure--local environmental pressure. A specification is Complex Specified Information (CSI) if it has meaning that is independent of the physical environment within which one finds the specification. i.e. it must be "detached" from that environment.

This concept may or may not have any validity. It may be total hogwash, but a specification being an independent, detachable pattern, has been, and still is a central component to Dembski's concept of Specified Complexity.

You perhaps aren't censoring the concept of independence to undermine Dembski (as there are plenty of critiques than are around to try that), but haven't grasped why the independence concept is so critical to the concept of Specified Complexity and further exposition in ID exploration.

What do you mean by "since un-guided processes are attached to their local environment"? Guettarda 03:19, 11 March 2006 (UTC)

This definition makes no sense. It would mean that "agsevbfthdfcx" is specified, as the "patttern" is independent of local conditions.Flarity 06:10, 1 July 2006 (UTC)

"un-guided processes", not my usage. Existed before my edits. Could probably use a change to un-intelligent processes. Additional cite for Specified Complexity Definition from Dembski's Expert Witness Report available at:

http://www.designinference.com/documents/2005.09.Expert_Report_Dembski.pdf

What is specified complexity? Recall the novel Contact by Carl Sagan.16 In that novel, radio astronomers discover a long sequence of prime numbers from outer space. Because the sequence is long, it is complex. Moreover, because the sequence is mathematically significant, it can be characterized independently of the physical processes that bring it about. As a consequence, it is also specified.

You are now discussing whether specified complexity is a menaingful mathematical concept. This is not the point of a discussion page, which should only focus on what should be in the article.

BTW, given that complexity is defined only up to a constant, comparisons of complexity are meaningful only as an asymptotic property of ensembles.

Please read Kolmogorov complexity. You might also want to look at Infinite monkey theorem--CSTAR 05:20, 11 March 2006 (UTC)

Specified Complexity is a concept of Information, the cites are justifications for the complete definition of the terminology to include the concepts of independence and detachability. It is more than just a "mathematical concept." Guettarda asked about "un-guided," a term of usage pre-dating my current edits. Guettarda is correct in his questioning of that usage. Whomever began with that usage, meant well, but was slightly off the point.

This is a discussion on what should be in the article, and I've provided numerous cites for what Dembski's definition of Specified Complexity is. If folks don't understadnd, Dembski's concept of independent, detachability, that not in itself justification for the censorship of the concept from the explaination in this article. As I pointed out, as recently as two weeks ago, a primer was posted on Dembski's own blog discussing the concept of Independence as a primary component of Specified Complexity. You might want to read my most recent cite, and understand what Dembski is saying with:

Moreover, because the sequence is mathematically significant, it can be characterized independently of the physical processes that bring it about.

Independence is critical to a full understanding of Specified Complexity as a concept. There is absolutely no valid reason for it's continued censorship from this article.

References to Kolmogorov complexity and the Infinite monkey theorem, are fine for further discussion about whether Specified Complexity is a valid concept; But the point of these edits is to present Demsbki's concept accurately at the start, and this continued attempt to censor what has been stated tome & again as a critical concept within Specified Complexity, I find baffling. No justification has been provided as to why Independence should not be properly introduced. Especially given, that "independence" is mentioned later in this very article. A mention that pre-dates any of my edits!

Reply
The following is from Dembski's article Aug 2005. Boldface is mine
"Moreover, we define the logarithm to the base 2 of M·N·φS(T)·P(T|H) as the context dependent specified complexity of T given H, the context being S’s context of inquiry: χ~ = –log2[M·N·φS(T)·P(T|H)]. Note that the tilde above the Greek letter chi indicates χ~’s dependence on the replicational resources within S’s context of inquiry. In a moment, we’ll consider a form of specified complexity that is independent of the replicational resources associated with S’s context of inquiry and thus, in effect, independent of S’s context of inquiry period (thereby strengthening the elimination of chance and the inference to design). For most purposes, however, χ~ is adequate for assessing whether T happened by chance. The crucial cut-off, here, is M·N·φS(T)·P(T|H) < 1/2: in this case, the probability of T happening according to H given that all relevant probabilistic resources are factored is strictly less than 1/2, which is equivalent to χ~ = –log2[M·N·φS(T)·P(T|H)] being strictly greater than 1. Thus, if χ~ > 1, it is less likely than not that an event of T’s descriptive complexity and improbability would happen according to H even if as many probabilistic resources as are relevant to T’s occurrence are factored in."
"As defined, χ~ is context sensitive, tied to the background knowledge of a semiotic agent S and to the context of inquiry within which S operates. Even so, it is possible to define specified complexity so that it is not context sensitive in this way. Theoretical computer scientist Seth Lloyd has shown that 10¹²⁰ constitutes the maximal number of bit operations that the known, observable universe could have performed throughout its entire multi-billion year history.31 This number sets an upper limit on the number of agents that can be embodied in the universe and the number of events that, in principle, they can observe. Accordingly, for any context of inquiry in which S might be endeavoring to determine whether an event that conforms to a pattern T happened by chance, M·N will be bounded above by 10¹²⁰ . We thus define the specified complexity of T given H (minus the tilde and context sensitivity) as χ = –log2[ 10¹²⁰ ·φ S(T)·P(T|H)]. "
He couldn't be clearer as to what his intended meaning is.--CSTAR 17:34, 11 March 2006 (UTC)

I notice you didn't highlight:
In a moment, we’ll consider a form of specified complexity that is independent of the replicational resources associated with S’s context of inquiry and thus, in effect, independent of S’s context of inquiry period (thereby strengthening the elimination of chance and the inference to design).
I suggest you become more widely read than a single paper.
I have cited where in the addendum to that paper where Dembski says:
...whatever happened to such key notions as detachability, conditional independence, tractability, and universal probability bounds that in the past had characterized my account of specification. They are still here...
I have cited the definition of Specified Complexity from ISCID where Dembski is the Executive Director:
The idea behind specificity is that not only must an event be unlikely (complex), it must also conform to an independently given, detachable pattern.
I have cited the definition from Dembski's Expert Witness Testimony for the Dover Trial:
Moreover, because the sequence is mathematically significant, it can be characterized independently of the physical processes that bring it about. As a consequence, it is also specified.
I have cited an ID primer posting on Dembsi's own blog from February of this year:
Next, the pattern has specification. The pattern conforms to an independently given specification. In this case siblings from the same family is the indendently given specification.
I've even cited within this very article, the definition:
Dembski asserts that specified complexity is present in a configuration when it can be described by a pattern that displays a large amount of independently specified information and is also complex...
The fact that you do not understand Dembski's concept of independence, as you mentioned earlier, is not justificstion in itself for the removal of that material from this article.
These are edits to the Introductory portion (paragraph 2), to keep it consistent with the concept of Independence that appears later in the article (paragraph 7).
Another editor has highlighted the poor word-choice of "unguided," to which you've reverted. I suggest there may be pride of authorship here rather than objective consideration.
I don't find that any of this makes much difference to the article, which is accurate as it stands and obviously enjoys broad consensus. Demsbki's assertions and math used to support them are widely considered to be spurious in the scientific community. How about contributing to the project constructively instead of ignoring consensus and repeatedly disrupting this article with daily reverts. FeloniousMonk 16:53, 12 March 2006 (UTC)
Reply
It is true that detachment and independence were characterizations of specification that Dembski used in earlier versions of his work, but which he subsequently dropped as characterizations. I have also agreed with you that according to Dembski "detachment", and "independence" are still there, as consequences of his definition of specified complexity, but the way you are inserting that phrase into the text is meaningless. In any case the correct formulation is already in the article below.
Re: Your comment
I notice you didn't highlight:
In a moment, we’ll consider a form of specified complexity that is independent of the replicational resources associated with S’s context of inquiry and thus, in effect, independent of S’s context of inquiry period (thereby strengthening the elimination of chance and the inference to design).
That form of specified complexity is simply the one in which the number of replicational resources is multiplied by 10¹²⁰. The portion of the text from Dembksi's article where it is introduced is highlighted.
--CSTAR 17:45, 12 March 2006 (UTC)
Response

I don't find that any of this makes much difference to the article, which is accurate as it stands and obviously enjoys broad consensus. Demsbki's assertions and math used to support them are widely considered to be spurious in the scientific community. How about contributing to the project constructively instead of ignoring consensus and repeatedly disrupting this article with daily reverts.
It makes a huge difference. Cetainly your not objecting to the fair and accurate presentation of Specified Complexity because it's considered spurious in the scientific community? Specified Complexity should have a full and accurate initial presentation with plenty of time for criticisms and examples of spuriousness later in the article.
My edits have been an attempt to move the introductory passages towards a fuller and more accurate presentation of the material. It appears that the objections have been a "pride of authorship" issue rather than a serious consideration of the consistency and adequacy of the initial material.
For example the primary objection has been the inclusion of independence, & detachability as a crucial part of the definition of Specified Complexity. The claim provided for removal of this material is that Dembski has "dropped" such issues in regards to Specified Complexity. This objection itself is spurious. I have provided numerous cites to back-up my assertion that it is an important aspect of the definition. The counter-claim has provided only a single cite, which itself contains the claim that "independence" and "detachability" are still there... I have provided a cite from Dembski's own blog dated as recently as Feb 2006 (after the date of the supposed single counter-cite).
But most dramatically, this very article is self-citing! The beginning of paragraph 7 says:
Dembski asserts that specified complexity is present in a configuration when it can be described by a pattern that displays a large amount of independently specified information and is also complex, which he defines as having a low probability of occurrence.
If the consensus opinion is that Dembski has "dropped" independence and detachability from the definition of Specified Complexity, then why is this material still here? Unfortunately, I must conclude it is there, not because there is a real objection to independence as a part of the definition, but that there is a reluctance to accept any change to material. It comes across as a pride of authorship issue, not a valid objection.
When it was claimed that Dembski does not use the word "Enslaved," I had no problem changing it to "attached" to more closely align it with Dembski's exact wording. When another reviewer questioned the use of the term "unguided," I concurred and provided the change to "un-intelligent." But here, is a perfect case in point.
Unguided is not my original word usage. I merely adopted and extended the current usage to create as little a disruption as possible. I'm not looking to disrupt the page, but the article's introduction needs to properly introduce Dembski's concept, and it currently does not.
Unguided is a poor word, yet there's continued reverting back to that word, even though the use of the word unguided in relation to Dembski's concept doesn't help to clarify. Dembski certainly is trying to make a distinction between intelligent activity and natural activity. His claim is that Random Mutation & Natural Selection cannot produce Specified Complexity. Dembski is well aware of the point that Natural Selection is not unguided. That Natural Selection is, in fact, guided by the environment. Environmental benefit is the natural selection pressure that guides development under RM / NS. Look at the current sentence:
Dembski argues that it is impossible for specified complexity to exist in patterns displayed by configurations formed by unguided processes.
That's not Dembski's point. Dembski's trying to draw a distinction between Intelligence guiding a process versus the natural environment guiding a process. Unguided is a poor choice of word here. Pointed out by more than just me, but we seem to always revert back the poorer word. A better phrasing is un-intelligent process.
This distinction is extremely important to understanding Dembski's conception of Specified Complexity. Natural processes "build-to" local environmental requirements. Therefore, as a marker of Intelligent activity, Intelligence alone can build-to non-local, or "external" requirements. This is Dembski's point here:
Moreover, because the sequence is mathematically significant, it can be characterized independently of the physical processes that bring it about. As a consequence, it is also specified.
Independent of the "physical processes." In other words, the environmental selection pressures that would be involved in it's development if purely under natural guidance. Mathematics is predominately used because Mathematics is an abstract concept and mathematical expressions should be recognized as independent from environmental selection pressures. However, it doesn't have to be entirely mathematical in concept. Specified Complexity is more than just probablistic calculations of bit strings...
Natural Selection developing a mechanical rover on Earth would optimize the design for the Terran environment. It takes Intelligence to develop a rover on Earth that is optimized for use on Mars. The Martian environment is an independent, detachable specification from the Terran environment. The Mars requirements are "external" requirements.
All I've done is added "independence," a recognized component if Specified Complexity into the Introductory portion. An important factor that needs to accompany any defintion of Specified Complexity. In response there's a single cite, from one paper that attempt to provide a specialized look at Specified Complexity in the absence of a prespecification. A paper that sill claims that independence & detachability are "still there" as an important consideration (though not to be made explicit in that particular treatment). The claim that it's been dropped from the defintion of Specified Complexity is contradicted by this very article...and may indicate that the reasons for removal of these edits lies somewhere else.

Re-edited intro for better clarity and consistency with rest of article. The intro also needs re-written to address Dembski's 2-stage process of Intelligent detection. As currently formulated it mixes CSI with Biology without sufficient set-up. Specified Complexity is an Information Theory. Step one, is to determine if Specified Complexity is a valid indicator of Intelligent Activity. Step two is to determine if Life exhibits CSI. Whether Life exhibits CSI is an independent consideration of the validity of CSI as an indicator of design in the larger sense.
Cite: http://www.leaderu.com/offices/dembski/docs/bd-specified.html
Does nature exhibit actual specified complexity? This is the million dollar question. Michael Behe's notion of irreducible complexity is purported to be a case of actual specified complexity and to be exhibited in real biochemical systems (cf. his book Darwin's Black Box). If such systems are, as Behe claims, highly improbable and thus genuinely complex with respect to the Darwinian mechanism of mutation and natural selection and if they are specified in virtue of their highly specific function (Behe looks to such systems as the bacterial flagellum), then a door is reopened for design in science that has been closed for well over a century. Does nature exhibit actual specified complexity? The jury is still out.
Questions about meaning
This article writes: "The concept is intended to formalize a property that singles out patterns that are both specified and complex." - what is a pattern that is "specified" vs one that is not? Also, the word "complex" is as subjective as "disorder" (i.e. entropy) if left undefined - as it is in this article. Fresheneesz 08:58, 5 May 2006 (UTC)
You got it. Ask Dembski. Although, he does attempt a definition. See further down in the article. --CSTAR 13:24, 5 May 2006 (UTC)
Information Conservation Law (Professor Stephen Hawking, conservation of information of block hole) Law of conservation of information different Dembski.
The overall flow of information into the system must be equal to the total outflow from the system, coupled with changes in the internal information systems; The information can transform, transforms from one condition other one condition; The information may create, may lose saves. That the formula used for
NQ = NW + ΔNU
Law of Information Conservation and Transformation (basic information equation) definition 1: in 2000, 2004, Deng Yu et al.
The law of conservation of information refers to the "system of information storage systems to increase equal access to information, less information to leave the system."
ΔNU = NQ-NW
Information stored in the system changes = into the system information - left the system information
= Create a new message - the loss of ( 'disappear' to leave) the information Information stored in the system changes = create a new message - keep missing.
ΔNU = Ncre-Nlos
System changes mean that information into the system to create a new information system, less to lose (to leave, keep missing, disappeared, eradication).
Information is the conversion law of conservation of the unique properties of the essential meaning: "The message can be created, you can keep missing," the original expression of the formula:
Information stored in the system changes = create a new message - keep missing information
ΔNU = Ncre-Nlos
Total flows in system's information to be equal to that total the information which flows out from the system, in addition system insider information change; be able to convert information from one state into another state; information can be created, you can keep missing.
Always flows in the system information to have to be equal to always the information which flows out from the system, in addition system interior information change. —Preceding unsigned comment added by DalaoDy (talk • contribs) 14:21, 23 October 2008 (UTC)
No chance?
In the section "Specificity" specified complexity is defined as
$\sigma =-\log _{2}[R\times \varphi (T)\times \operatorname {P} (T)],$
which isn't quite, what Dembski writes in his Specification paper.
The dependency of $\varphi (T)$ on the semiotic agent S is missing, which might be accepted; but the dependency of $\operatorname {P} (T)$ on a "chance" hypothesis H should have been there. That is, it should have been $\operatorname {P} (T|H)$ . Otherwise the probability function P is not defined.
Also notice that the first paragraph in the section "Dembski's explanation of specified complexity" ends with "P(T) (where P is the "chance" hypothesis)", which doesn't make too much sense - P is a probability measure, not a hypothesis.
Dembski's approach is Fisherian, not Bayesian. That means that Dembski "proves" design by eliminating chance, which is quite important. Design can always generate specified complexity according to Dembski, so P(T | D) = 1, where D is a design hypothesis. The dependency of specified complexity on the chance hypothesis is therefore important.
--FreezBee 09:22, 15 May 2006 (UTC)
Well yes that's right, but a hypothesis in the Fisheroian approach is a set of probability measures. In this case, it is a singleton {P} where P is what Dembski refers to as the chance hypothesis. The chance hypothesis is something like a uniform measure on the sample space -— whatever that means, at least that's what Demski suggests. In thecase of evolutionary pathways, the chance hypothesis would be something like a uniform random walk (uniform in the sense that steps in any direction are equally likely.) BTW, the original form in which I wrote this explanation, I did use the notation used by Dembski (and the one you use), which is not a great one, because it suggests a conditional expectation. However, it was suggested by others on this talk page (or some archived version of it) that it was too complicated.
Ok, I understand that the notation P(T | H) may be confusing; but unfortunately without the emphasis on the chance hypothesis, certain important points are missing - e.g., if you want to follow Dembski's discussions of the Fisherian approach vs. the Bayesian approach to design detection. Elliot Sober e.g. is Bayesian, and for those who want to figure out, why Dembski and Sober disagree, it makes a difference. But, ok. this may mostly be something for the extremely interested :-)

Dembski is never too specific about the probability measure P - apparently he has never heard about about a statistical model - but as you say, he tends to assume some kind of equiprobable random walk - without any particular proof. He just assumes that physical laws cannot produce anything new and chance can only produce random walks (which have high entropy), which is highly unlikely to produce specified complexity (which has low entropy). But as above, this may be too technical for the average reader.

--FreezBee 14:40, 15 May 2006 (UTC)
Re: Dembski's approach is Fisherian, not Bayesian. That's noted in the article.--CSTAR 12:32, 15 May 2006 (UTC)

Re: Design can always generate specified complexity according to Dembski so P(T | D) = 1. I'm not sure Dembski ever says this in symbolic form. I'm not sure he even defines a set D as you suggest.--CSTAR 12:46, 15 May 2006 (UTC)
I'm sure he doesn't. Since Dembski has design capable of everything he need not worry about design hypotheses. Let E be some observed event and D be a design hypothesis and H a chance hypothesis. Then we can talk about as well P(E | D) (the probability of E if it was caused by design) as P(E | H) (the probability of E if it was caused by chance); but also about as well P(D | E) (the likelihood that design was the cause of E) as P(H | E) (the likelihood that chance was the cause of E). Dembski doesn't want to discuss the versions with D, since that would require him to make assumptions about the capabilities of the designer - which kind of designer could make a bacterial flagellum? A mad scientist or something?

--FreezBee 14:40, 15 May 2006 (UTC)
Hmm. Interesting. Never thought of that. --CSTAR 21:38, 15 May 2006 (UTC)
Well, I have added an "Example" section, which may give some people some intuition about, what Dembski means. The example doesn't come around all corners of Dembski's theory, but it does touch upon spme of them :-)

You are free to edit or delete that section as you see fit; I just thought that it might clear up some of the confusion about, what Dembski is trying to say. It's of course only my personal interpretation of Dembski.

--FreezBee 11:11, 16 May 2006 (UTC)
I haven't read your example carefully, altough I right away see several problems:
(a) There is a fundamental conceptual problem with Dembski's rank function. Namely, the rank is an asymptotic concept, applicable to infinite ensembles. The reason for this is that Kolmogorov complexity is defined up to a constant. If you have a finite set of strings, then if you try to rank them by complexity, you are likely to get a lot of "complexity ties" which will be resolved by some other ordering, such as lexicographical ordering. So for finite sets, Dembski's ordering is useless. Note that this comment is not in the article, because I couldn't find it (explicitly at least) in some published work.

(b) If the example remains, it should be cleaned up for style and terminology.

(c) I don't think we should give the impression that WP supports his view. What I tried to do in writing that section is present his idea in the cleanest form possible (principle of charity) but clearly not endorsing it, because this idea has absolutely no acceptance in the scientific community. --CSTAR 14:52, 16 May 2006 (UTC)
(a) Dembski's rank ordering isn't extremely well defined; but involves the concept of 'semiotic cost', apparently the number of symbols used in a specification. For instance, an "outboard rotary motor-driven propeller" has a semiotic cost of 4, and a "Royal Flush" has a semiotic cost of 2. Dembski states that the rank of a pattern is the lowest semiotic cost for a description of that pattern among those descriptions known to the semiotic agent in question. It's beyond me to say, how well-defined all this is - but it's what Dembski writes :-)

This is clearly a kind of Kolmogorov complexity, which as you write is only defined up to a constant - because different Turing machines have different instruction sets. But just as you can define Kolmogorov complexity relative to some specific Turing machine, Dembski's rank-ordering is defined relative to a semiotic agent, it's φ_S(T). For instance, a person who knows about poker may identify a "Royal Flush", whereas a person who doesn't know about poker may identify "Ten to Ace of the same suite", which uses more symbols.

(b) ok - if you let it stay, I'll try to take care of that.

(c) My intention was only to illustrate what Dembski's terminology means with an example that most people should be able to relate to. I read some critiques of Dembski that miss some of his points. It's partly Dembski's fault, because he isn't the most consistent person regarding terminology, which makes trying to figure out what he does mean and what he doesn't mean quite a laborous task. I just wanted to spare my fellow human beings the trouble. In the example I only illustrate Dembski's terminology, I don't say that Dembski's Explanatory Filter is really the way we detect design, nor that design detected using the EF really is design.

--FreezBee 16:02, 16 May 2006 (UTC)
Yes it is a kind of Kolmogorov complexity. But just as you can define Kolmogorov complexity relative to some specific Turing machine, This involves an arbitrary choice of Turing machine. For the definition to be "invariant" (e.g., not involving arbitrary choices), it has to range over all Turing machines (or over all universal Turing machines), but in any case, complexity of a single object is not invariantly undefined. This implies (as I mentioned above) that complexity comparisons within finite sets are meaningless and in particular, complexity rank is meaningless.

BTW, it's not "my choice" whether your contribution stays or not! WP articles are written by consensus. I'll try to read it more carefully later and give a more reasoned opinion.--CSTAR 16:22, 16 May 2006 (UTC)
OK I'm not sure I believe the following paragraph, as per the reasons I gave above:
which in this case would be the number of individual cards (that all can be named on the form X of Ys) plus the number of special combinations (“One Pair”, “Two Pairs”, “Full House”, “Royal Flush”, and so on) with at most two symbols minus 1 for the “Four Aces” (to not count it twice). This count up to more than 52 (= the number of indivudual cards), but certainly less than 100, and since you’re in a good mood, you don’t mind to favor the chance hypothesis, so you say that φ(T) = 100.
In particular, what does "the number of special combinations" mean? --CSTAR 22:00, 16 May 2006 (UTC)
Presumably you mean those that can be named on the form X of Ys (similar to Dembski's description of the flagellum). I'm not sure I believe your upper bound.--CSTAR 22:10, 16 May 2006 (UTC)

Read p. 19 of the Specification paper. Dembski writes:
think of the game of poker and consider the following three descriptions of poker hands: “single pair,” “full house,” and “royal flush.” If we think of these poker hands as patterns denoted respectively by T₁, T₂, and T₃, then, given that they each have the same description length (i.e., two words for each), it makes sense to think of these patterns as associated with roughly equal specificational resources.
True, Dembski only says that they have "roughly equal specificational resources", not what the specificational resources are. It's a point, where he could be a lot clearer. But on p. 17 he writes that
φ_S(T) is the cardinality of {U ε patterns(Ω) | φ′_S(U) ≤ φ′_S(T)}, where patterns(Ω) is the collection of all patterns that identify events in Ω.
In other words, we are only dealing with patterns of relevance, in casu patterns that identify poker cards, maybe only poker hands, not all two word phrases.

By "the number of special combinations" I mean the number of poker hands that have a special name, that is "Royal Flush", "One Pair", "Full House", and so on. How many are there of those with at most two symbols?

I'm unsure why you don't believe my upper bound. If you like, you can make it 1,000 :-); in the case in question, it makes no difference.

--FreezBee 09:27, 17 May 2006 (UTC)
That's one point; the upper bounds are so large, they're virtually useless. Even your example concludes "no design". And how many two word phrases are there? Well, why isn't "All red" or "Some black" or "with number" an acceptable two-word description of some deck of cards? Say if we allowed 100 possible choice for each word, that would give us 10,000 choices right there (see the discussion in the previuos paragraph on the bacterial flagellum).--CSTAR 14:33, 17 May 2006 (UTC)

I have changed the description of the patterns to make the clearer (hopefully). As for your examples, is e.g. "All red" a poker hand? I'm no expert on poker; but I do believe there's no such poker hand, but if you insist, it's ok with me. The example is only supposed to give some idea about how it works. After all, Dembski isn't himself too precise regarding exactly, which reference class of possibilities too choose. Remember you don't count different descriptions as such, but events - each event represented by the simplest description at disposition.

Regarding the flagellum, Dembski uses a dictionary with 100,000 words and describes the flagellum as an "bidirectional rotary motor-driven propeller"; that is, by using four words. Using a dictionary with 100,000 = 10₅ words gives at most 10₂₀ possible 4 word concepts, and since 10₅ + 10₁₀ + 10₁₅ + 10₂₀ actually is on the order of 10₂₀, Dembski calculates the specificational resources to that value. A person with a more limited vocabulary would come to a lower number for specificational resources; but since a design interference is made by someone, this subjective element is actually part of the story. Dembski claims that his Explanatory Filter models how we actually make design inferences, so there's nothing wrong with a subjective element in that way.

I don't really understand your problem. Remember that I'm only trying to illustrate what I think is Dembski's position, not whether I myself consider the Explanatory Filter to have any merit.

--FreezBee 15:07, 17 May 2006 (UTC)
Re: Remember that I'm only trying to illustrate what I think is Dembski's position, not whether I myself consider the Explanatory Filter to have any merit. Understood.

My problem is that φ is not well-defined (or defined up to a presposterously large constant) and any upper bounds for it are going to be large.

As far as descriptions of poker hands, I of course realize that these are no card descriptions using that name, but I dn't see right off how you can exclude them from reasonable patterns. ISuppose you had a "designer" (of card hands) that was mathematically oriented and chose particular kinds of mathematical sequences in dealing out cards.

But anyway I think you should try to fix up the example and make it a little "third person" description.--CSTAR 15:18, 17 May 2006 (UTC)

I agree that φ is not well-defined; but since Dembski uses it, and it's important for understanding, what he means by specified complexity, there should be some explanation of how it's calculated. That the upper bounds are large is just the way it is, and what's wrong with it? Unless you want more designinferences, that is ;-)

--FreezBee 16:38, 17 May 2006 (UTC)

I've removed the example placed the article discussed above to here. Though an impressive piece of work I'm not sure it adds anything to the article due to it's complexity (keep in mind our readers), and unless the example given was published by Dembski, it consititutes original research. Is this example used by Dembski? FeloniousMonk 15:20, 17 May 2006 (UTC)

It constitutes an example to actually figure out, what Dembski means. Is that originally research? I don't think that Dembskiology even exists yet, so it's pre-original research, if there's such a thing. I added it, because apparently people have great troubles figuring out, what Dembski means, and since the article doesn't give any examples of how to calculate specified complexity. And I have found it frustrating to discuss with people, what specified complexity is/isn't - that "complexity" doesn't mean "complex", and that there's a subjective element in it. That was the background for my example. That the example exhibits complexity (specified?) is unavoidable, when trying to explain something that even the inventor (not of the phrase, but Dembski's version is peculiar to him) has spent around 10 years trying to explain it and changing his mind ever so often along the way.

That the example doesn't add anything to the article is because it's an example - I examplify the words used in the article, that's what an example does, isn't it? As the article is now, can you tell me how φ(T) is calculated? Is it an objective or a subjective function? What does it really measure? What is the factor R? How is it measured in given circumstance? I really think the article needs an example that shows how the factors are calculated, otherwise specified complexity will simply appear to be some kind of magic, and after all, it isn't.

--FreezBee 16:38, 17 May 2006 (UTC)
Well specified complexity is magic. It's certainly not math.--CSTAR 16:41, 17 May 2006 (UTC)
I see a clear, principled reason for excluding FreezBee's example. Specified complexity is not an established concept (far from it, in fact). This isn't like providing an example application of, say, the Pythagorean Theorem. There is widespread disagreement, even in the ID camp, on how to interpret and apply Dembski's method. As the sole developer of the specified complexity concept, only Dembski is in a position to say whether FreezBee's interpretation is faithful to his ideas.
Some objections to FreezBee's example might include:
- On what basis does FreezBee restrict his specification space to poker hands? I'm guessing that every 5-card hand has meaning to someone somewhere. For example, if I google "7J82K", I get 8 hits.
- On what basis does FreezBee restrict his null hypotheses to uniform chance? To show that other null hypotheses are irrelevant, he would have to assess their prior probabilities, which Dembski says is problematic.

--Secondclass 00:49, 31 October 2006 (UTC)
== Example == Say you are playing poker in the WikiPalace. The player next to you puts down four Aces with an Eight of Diamonds to spare, even without having to change any cards. All you have is Two and Five of Spades, Ten of Hearts, Ten of Clubs, and King of Diamonds. Does the player next to you have his lucky day, or do we have a case of specified complexity? The chance hypothesis would here be that the five cards the player had were dealt from a deck of cards shuffled sufficiently that we could consider each possible poker hand (= set of five cards) equally probable. There are C52,5 = 52*51*50*49*48/5*4*3*2*1 = 2,598,960 different poker hands, and subject to our hypothesis, each of these has the same probability, 1/2,598,960 But you don’t really care about that Eight of Diamonds, only about the four Aces, since they have a specific meaning in poker; that is, the observed event is “four Aces and Eight of Diamonds”, but the target pattern ''T'' is “Four Aces”. The four Aces can be combined with any of the other 48 cards and still fit the “Four Aces” pattern; that is, subject to our hypothesis, P(''T'') = 48/2,598,960 = 1.84689*10-5. The semiotic cost (number of symbols) of “Four Aces” is 2, while the semiotic cost of your own hand, Two and Five of Spades, Ten of Hearts, Ten of Clubs, and King of Diamonds, is something like 9 (“Two”, “Five”, “Spades”, “Ten”, “Hearts”, “Ten”, “Clubs”, “King”, and “Diamonds”). The specificational ressources, ''φ''(''T'') is the number of patterns that are at least as simple as “Four Aces” and that describe possible targets in the reference class of possibilities Ω; that is, the number of patterns with a semiotic cost of at most 2 and that describe some poker hand. Note we are here counting patterns, not how many poker hands correspond to a pattern; e.g. "Royal Flush" is ''one'' pattern, although there are four different Royal Flush hands. In our case the relevant patterns would be something like * individual cards on the form ''X'' of ''Y''s, e.g. "Jack of Spades" - 52 in all; * one pair on the form Pair of ''X''s, e.g. Pair of Tens - 13 in all; * three of a kind on the form Three ''X''s, e.g. Three Queens - 13 in all; * four of a kind on the form Four ''X''s, e.g. Four Nines - 13 an all; and * named hands - "Royal Flush", "Straight Flush", "Full House", "Flush", "Straight", and "Two Pairs" - 6 in all These patterns count up to 97, but since you’re in a good mood, you don’t mind to favor the chance hypothesis by rounding up, so you say that ''φ''(''T'') = 100. You notice that there are many other poker players in the WikiPalace and wonder if any of them have had a “Four Aces” besides the one you are dealing with. Compare it with the argument for abiogenesis, that although abiogenesis on ''one'' planet is highly improbable, there are most likely billions of planets in the universe, so it’s very probable that abiogenesis should occur on ''some'' planet. In other words, we need to multiply the probability of getting “Four Aces” in ''one'' game with the number of players and the number of games they have played. Say there are 100 players, and they have all played at most 10 games. You are still in a good mood, so you calculate the replicational ressources to ''R'' = 100*10 = 1,000 = 1*103. Now you need to multiply it all together; that is, calculate ''R'' × ''φ''(''T'') × P(''T''). You’ll follow Dembski’s advice and cut it even; that is, if ''R'' × ''φ''(''T'') × P(''T'') < ½, the verdict is “design”, and if ''R'' × ''φ''(''T'') × P(''T'') ≥ ½, the verdict is "chance". You actually find that <blockquote>''R'' × ''φ''(''T'') × P(''T'') = 1*103 × 1*102 × 1.84689*10-5 = 1.84689 ≥ ½</blockquote> so the player just has a lucky day. Since –log2(½) = log2(2) = 1, we could also have for checked, if the context dependent (of WikiPalace) specified complexity <blockquote>σ = -log2(''R'' × ''φ''(''T'') × P(''T'')) = -log2(1.84689) = -0.8851</blockquote> is larger than 1, which it isn’t.
Related links?
Weasel program seems to address the concept underlying Dembski's maths, and Hoyle's fallacy appears to be another version of the same claim, though that article needs work to be useful. It has a link to TalkOrigins' Lies, Damned Lies, Statistics, and Probability of Abiogenesis Calculations which looks like a good source for such improvements. There's not much interlinking between these articles all touching on the same basis topic – comments? ...dave souza, talk 16:57, 21 September 2006 (UTC)
"arguments for existence of god" IMO does not belong here
The "see also" points to "arguments for existence of god" (actually, "existence of god", but maybe some specific part), and I think that it has nothing to do with specified complexity, or at least not enough to deserve to be mentioned in the "see also". It looks like those associations of theism with fundamentalism/anti-evolution and vice-versa. --Extremophile 01:11, 25 November 2006 (UTC)
Specified Complexity and Information Content
I do not fully understand the arguments on either side of this controversy, but several questions came to mind while I read the article. One is:
If the receivers at SETI detect a signal, what criteria are used by researchers to determine whether the signal is noise or information? Can these same criteria apply to the CSI discussion? If not, why?
No, because SETI is looking for narrow-band radio signals, which, as far as we currently know, are not made by natural phenomenon. This isn't the same as information. I suppose an analogy would be that if we found a cell that was using fission of uranium for power, that might indicate the need for an intelligent designer, but, just as the existance of pulsars doesn't need intelligent life for an explanation, despite producing extremely ordered signals, nothing currently known in biology is beyond the capacity to be evolved, as far as can be known. See here.Adam Cuerden ^talk 22:13, 17 February 2007 (UTC)
POV?
While generally NPOV, I think we need to mention the heavy criticism in the lead paragraph, if briefly. Otherwise, undue weight is given by not making the reader aware of the flaws until after a lengthy description of the supposed merits. Adam Cuerden ^talk 15:25, 17 February 2007 (UTC)
Hmmm, I thought mention of the heavy criticism was already there. As far as the "supposed merits" there is merely a description of what Demsbski says, based the principle of charity. However, I don't think the article should try to convey that anybody other than Dembski (and his tiny band) actually claims to believe this stuff. Any specific thing you would change?--CSTAR 16:41, 17 February 2007 (UTC)
Agreed. FeloniousMonk 19:14, 17 February 2007 (UTC)
I don't think we need to add a lot, even as little as adding a quick sentence to the end of the first paragraph, along the lines of "....Dembski claims that specified complexity is a reliable marker of design by an intelligent agent, a central tenet to intelligent design and which Dembski argues for in opposition to modern evolutionary theory. However, this is widely disputed." might suffice. The rules of writing probably ask for a bit more detail, but, really, all that's actually needed is to make it clear, before the long paragraphs summarising his claims, that it's highly controversial, so that the reader is not in an unduly credulous state during them. Beyond that, it seems fine. Adam Cuerden ^talk 19:25, 17 February 2007 (UTC)
It might be good also to note that Dembski has dropped talking about SC in the last few years and focused only on irreducible complexity. If we can get a source for this that would be good to include. JoshuaZ 23:49, 17 February 2007 (UTC)
If so, then it was after August 2005. However, it wouldn't surprise me. The Isaac Newton of information theory is NOT getting much traction in the mathematical sciences.--CSTAR 03:47, 19 February 2007 (UTC)
POV problem
The current leadstill waits until paragraph four to mention the heavy criticism of specified complexity. It doesn't mention that it hasn't once actually been rigourously applied by anyone. Didn't we discuss this before? Adam Cuerden ^talk 08:56, 9 March 2007 (UTC)
Well it wasn't clear who was going to do what. Anyway a shifted a sentence around to the 1st paragraph. Is this any better?--CSTAR 13:28, 9 March 2007 (UTC)
Could we now remove the POV banner?--CSTAR 16:21, 9 March 2007 (UTC)
I'm tempted to say it's too soon to be sure. However, if you're confident that you've addressed the reasons for the banner being atatched in the first place... be bold... we can always pick up the pieces later :-) SheffieldSteel 16:59, 9 March 2007 (UTC)
It's better, though I do think this article as a whole is a little too... credulous about mathematical concepts that have never once actually been applied to anything as set out by Dembski. Is there any point even mentioning the mathematics when they've never been used? Adam Cuerden ^talk 18:33, 10 March 2007 (UTC)
That fact the article reports on Dembski's claims is hardly being credulous; I think it states the problems Dembski's "definitions" have and why these concepts fail to be mathematically well-defined. "Is there any point even mentioning the mathematics when they've never been used?" Yes. This and related material has been written about, criticically, by competent people, eg., by Jason Rosenhouse, Perakh, Shallit and others. The material is notable, obviously not because of its scientific interest, but for other reasons. --CSTAR 19:38, 10 March 2007 (UTC)
"The concept of specified complexity is widely regarded as mathematically unsound and has not been the basis for further independent work in information theory or complexity theory. [1][2][3]" The sources are 2 websites and 1 paper. And it does not say int thme that it is 'widely regarded' 68.109.234.155 23:50, 26 March 2007 (UTC)
Intro is too long; apparently getting longer
we need to make sure we don't include too much in the intro section that would be better mentioned in the main body of the article. I know it's tempting to always make additions at the top where everyone will see them straight away, but that doesn't mean it's the best way to incorporate new material. SheffieldSteel 17:02, 27 April 2007 (UTC)
Monkeys & Typewriters
Would there be any objection to describing Dembski's calculation of the probability of the bacterial flagellum as based on the "monkeys & typewriters" concept of Darwinian evolution? I have the impression that almost all such probability calculations, going back to the Reverend William A. Williams, are. MrG4.225.210.148 21:26, 11 July 2007 (UTC
An entertaining thought, but to meet out policies of verifiability and no original research all statements have to be attributed to a reliable source which directly relates them to the subject. ... dave souza, talk 21:46, 11 July 2007 (UTC)
I won't contest that, too trivial to be worth the effort, but still, when the article says: "These methods assume that all of the constituent parts of the flagellum must have been generated completely at random ..." -- that says "monkeys & typewriters" to me. MrG 00:54, 12 July 2007 (UTC)