|
Author
|
Topic: Is 2nd Law a special case of 4th Law?
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 15:52
quote:
But on my understanding of CSI I do not see how it is possible to have an object (event) that is both K-simple and Specified Complex (i.e. having CSI). You clearly do because you say (other thread):
Hi Chris,
I do appreciate your concern, and I know my statements can be disconcerting. However, if one can set aside one's biases about the definition of complexity, and realize what "Complexity" in CSI refers to, it is seen that K-simple (compressible structures) can qualify as CSI.
This is actually a very important point. From Dembski's No Free Lunch
quote:
page 144:
It is CSI that within the Chaitin-Kolmogorov-Solomonoff theory of algorithmic information identifies the highly compressible, nonrandom strings of digits (see section 2.4).
and if we turn to section 2.4, here is an example of such a non-random string on page 59
quote:
1111111111111111111111111....
So we see that "complexity" in Dembski's formulation does not relate to K-complexity, but to probabilistic complexity.
This is disconcerting at first, but one eventually realizes that it is the broadest definition of CSI possible. Recall, I suggested that if the word "complexity" is hard to digest for the term CSI, think of ISI (Improbable Specified Information) to get an alternative perspective.
Does that help?
Salvador
[ 24. May 2005, 15:54: Message edited by: Salvador T. Cordova ]
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 20:14
Chris,
Bill Dembski actually uses what I just suggested to you:
quote:
Hence, within my scheme, "specified complexity" or "specified improbability" becomes the key to identifying intelligence.
Reply to Henry Morris
IP: Logged
|
|
andyg
Member
Member # 415
|
posted 24. May 2005 20:24
Sal,
I'll butt in here as the conversation has overlapped with the Nature thread.
I am still confused over your invocation of Probabilistic Complexity. In a string of binary digits chosen randomly, why is
1111111111111111
more probabilistically complex than
1001010110101001 ?
They have an equal chance of occurring - 2^16, correct?
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 20:54
quote:
Sal,
I'll butt in here as the conversation has overlapped with the Nature thread.
I am still confused over your invocation of Probabilistic Complexity. In a string of binary digits chosen randomly, why is
1111111111111111
more probabilistically complex than
1001010110101001 ?
They have an equal chance of occurring - 2^16, correct?
Hi Andy,
Sorry to be splitting time on you with Chris. The string have equal probilistic complexity. The reason I liked Bill Dembski's "specified improbability" over "specified complexity" is that it alleviates the confusion over the word "complexity".
But from Bill's literature, the two above strings have equal probabilistic complexity, and the string with all "1" is identified as K-simple (highly compressible under Chaitin-Kolmogorov-Solomonoff algorithmic information theory).
K-simple strings provide an independent specification, combined with low probability (assuming that 1111111111111111 is derived from a space of possible physical configurations), then it qualifies as CSI if one's detection criteria searches for k-simple CSI.
Thus if one walks into a room and sees a sea of coins on the table all turned heads, one reasonably (not exhaustively) presumes design somewhere in the pipeline.
[ 24. May 2005, 20:54: Message edited by: Salvador T. Cordova ]
IP: Logged
|
|
andyg
Member
Member # 415
|
posted 24. May 2005 21:28
quote:
K-simple strings provide an independent specification, combined with low probability (assuming that 1111111111111111 is derived from a space of possible physical configurations), then it qualifies as CSI if one's detection criteria searches for k-simple CSI.
Thus if one walks into a room and sees a sea of coins on the table all turned heads, one reasonably (not exhaustively) presumes design somewhere in the pipeline.
I agree that one would infer design from a sea of coins all showing heads. But this has to do with what you referred to in the other thread as the coincidence of physical and conceptual information. Our prior knowledge of coins suggests that they have two different faces and the chance of either face showing is 50%. This is our conceptual information.
If, however, you walked into a similar room in a trick coin-making factory, finding a roomful of coins all showing heads would not be very surprising and their arrangement (heads versus, well....heads) would not require design.
What I think you are saying is that it is possible to infer design when your prior experience tells you that a particular outcome is out of the ordinary. To paraphrase Paul Nelson, we talk about such improbabilities all the time when we say things to our kids like "Those cookies didn't walk away by themselves". But we say that precisely because we know a great deal about cookies and the proclivities of small children.
I still don't see how this helps in the cases of poor or non-existent knowledge. Take the following string
2024562461
That string would mean very different things to Isaac Newton, a resident of Maryland, and George Bush. Newton would see it as any old ten digit number, a Maryland resident might also see it is a DC-area phone number, and George Bush might recognize it as the White House Fax number.
In the "Nature" thread, I asked Jerry to work out whether some strings of letters contained CSI. You replied that the question as posed did not contain enough information to answer conclusively, and I think you were referring to conceptual information. What more information would be required?
Finally, how do you think these problems may be overcome in the future with respect to biological objects, where our conceptual information is non-existent? You have stated that you believe the problem to be insoluble at present, but think it may become tractable in the future. How?
Thanks for sticking with this.
![[Smile]](smile.gif)
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 22:12
quote:
Our prior knowledge of coins suggests that they have two different faces and the chance of either face showing is 50%. This is our conceptual information.
I'm afraid the bolded portion is incorrect. Again, please bear with me that what we are still discussing at this point are definitions, we're not yet addressing whether ID claims are correct. And I must advise you, I refuted Wesley Elsberry on one of his writings. He does not represent Dembski's work accurately, and I would request you try to drop the misconceptions Elsberry has promoted.
Part of the reason you may not be seeing the examples clearly is Elsberry representations are not what Dembski's work really says. The reason I asked what books you have of Dembski's, is that you would be able to more clearly see the definitions with his literature in hand. That said, I'm happy to try to walk you through this material, but you have to be willing to shed the misconceptions widely promoted on the net about Dembski's work.
I'm giving you the definitions from Dembski's writings. In a sense, we are making an arbitrary choice of what we call what. "conceptual information" is a phrase that is carefully defined by Dembski.
The coins being heads or tails and the number of coins involved give us the space, Omega, of possibilities. That is a string of 5 coins spans the following space of possibilities:
T T T T T
T T T T H
T T T H T
T T T H H
T T H T T
T T H T H
T T H H T
T T H H H
T H T T T
T H T T H
T H T H T
T H T H H
T H H T T
T H H T H
T H H H T
T H H H H
H T T T T
H T T T H
H T T H T
H T T H H
H T H T T
H T H T H
H T H H T
H T H H H
H H T T T
H H T T H
H H T H T
H H T H H
H H H T T
H H H T H
H H H H T
H H H H H
Conceptual information corresponds to a set of possible patterns selected from the possible physical outcomes. A specification is constituted with conceptual information. This may be shocking, but ANY specification will work provided:
1. it is not post-dictive
2. it is improbable
The a hypothetical outcomes such as:
code:
{ H T H H T }
constitutes a specification. Even a set of outcomes like:
code:
{ { H T H H T }, { H H T T T } }
constitutes a specification. The specification are reasonably probable, so ordinarily they would not constitute adequate specifications.
I hope that clarifies what conceptual information is.
quote:
Take the following string
2024562461
That string would mean very different things to Isaac Newton, a resident of Maryland, and George Bush. Newton would see it as any old ten digit number, a Maryland resident might also see it is a DC-area phone number, and George Bush might recognize it as the White House Fax number.
Absolutely!!!!!! It's questions like these that take time to address, and I'm happy to walk you through them.
The response is, that at some point, when the space of possibilities is large, like say a 150 digit password, it becomes, short of design, impossible for someone to have the same conceptual information in their head as you, unless of course you think similarly! That is exactly why passwords that are long and creative are hard to break, no else is likely to happen upon it in geological time if the password is long enough!
[ 24. May 2005, 22:13: Message edited by: Salvador T. Cordova ]
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 22:19
quote:
If, however, you walked into a similar room in a trick coin-making factory, finding a roomful of coins all showing heads would not be very surprising and their arrangement (heads versus, well....heads) would not require design.
That is a reasonable objection, but bear in mind, the space Omega of possibilties has only one element, namely all heads. There is only one possible Event, E, that can happen. Information content of E is
The information content =
I ( E ) = - log2( P(E)) = - log2( 1 ) = 0 bits
where
P(E) = probability of an event happening, in this case the P(E) has a probability of 100%.
In otherwords, with two headed coins, there is 0 bits of information, and therefore no conceptual information.
IP: Logged
|
|
andyg
Member
Member # 415
|
posted 24. May 2005 22:23
I agree that any pattern could be specified, and that it is obviously not appropriate to specify it after the fact (as in Dembski's example of drawing the target around the arrow after the shot).
But.......
quote:
The response is, that at some point, when the space of possibilities is large, like say a 150 digit password, it becomes, short of design, impossible for someone to have the same conceptual information in their head as you, unless of course you think similarly! That is exactly why passwords that are long and creative are hard to break, no else is likely to happen upon it in geological time if the password is long enough!
Exactly. This is my problem. How does a naive observer access the conceptual information? Without the conceptual information, the seemingly random 150 digit number is just, well, random. Where is the conceptual information in a flagellum, or a chromosome, or a mitochondrion?
![[Confused]](confused.gif)
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 22:37
quote:
Exactly. This is my problem. How does a naive observer access the conceptual information? Without the conceptual information, the seemingly random 150 digit number is just, well, random. Where is the conceptual information in a flagellum, or a chromosome, or a mitochondrion?
The way this happens is IF the designer is willing to have his work discovered as designed. Heavily encrypted designs will elude detection.
I cannot emphasize that enough, "designs that are meant to be detected" have the best chance of being detected.
This is echoed by Bill Dembski in Mere Creation
quote:
Intelligent causes can do things that unintelligent causes cannot and can make their actions evident. When for whatever reason an intelligent cause fails to make its actions evident, we may miss it. But when an intelligent cause succeeds in making its actions evident, we take notice. This is why false negatives do not invalidate the explanatory filter. The explanatory filter is fully capable of detecting intelligent causes intent on making their presence evident. Masters of stealth intent on concealing their actions may successfully evade the explanatory filter. But masters of self-promotion intent on making sure their intellectual property gets properly attributed find in the explanatory filter a ready friend.
It is a misconception widely promoted on the net the Explanatory Filter (EF) was intended to detect stealthy designs. No, it was intended to detect self-promoting designs.
So let's start with 150 coins. You wish to be a master of self promotion. Your goal is to take those coins and some how figure out what sort of patterns, I the observer, would recognize as designed. What would you do? Most likely you'll arrange them in a k-simple structure.
Think about it. What conceptual patterns (specifications, conceptual information) would you use as a blue print to arrange the coins such that they conveyed physical information which would match the conceptual patterns I would recognize????
If you can answer that, at least intuitively by giving examples, we can then move on to biology.
Salvador
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 22:48
quote:
Exactly. This is my problem. How does a naive observer access the conceptual information? Without the conceptual information, the seemingly random 150 digit number is just, well, random. Where is the conceptual information in a flagellum, or a chromosome, or a mitochondrion?
What if you found a 5495 digit number inside a cell that match a number independently arrived at elsewhere, would one begin to suspect design?
Take a look at this number:
Turing Number
We have good evidence this number is actualized within cells. There is no reason from physics that it should exist in the cell (unless one invokes miracles of chance).
[ 24. May 2005, 22:50: Message edited by: Salvador T. Cordova ]
IP: Logged
|
|
Jerry D. Bauer
Member
Member # 756
|
posted 24. May 2005 22:53
May I just throw this out? Andy seems to think there is only one kind of specificity: that as explicated by Dembski. This could not be further from the truth.
Dembski did not coin the terms "specified" or even "complex specified" as these terms predate him.
A recognizable pattern IS one valid form of specificity as Dembski accurately noted. But another is the type that Behe uses which has nothing to do with patterns but hinges on the definition that a part does a "specific" role in a given system. IOW, no other part in the system will function as that part does, so that part is highly specified.
Finally, there is mathematical specification which goes back at least to Brillioun I am aware of and maybe much further. This specification is calculated mathematically (as I did in another thread) and is used to quantify polypeptides and the like.
Here are at least three different specificities.
Ok. Carry on. ![[Wink]](wink.gif)
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 24. May 2005 23:06
OK Andy,
A little more about biology.
Under certain circumstances, compressible structures (symbollically speaking) convey design.
A salt-crystal is a symbolically compressible structure. Na-Cl-Na-Cl ........
But the structure of salt-crystals is not seen as designed system because the structure has almost 100% probability of being acutalized from a space of reasonable probabilities (constrained by chemistry).
However, biotic materials are not constrained by such requirments. Thus, the homo-chiral pattern in DNA's an proteins are highly suggestive of design, since an a mostly random mix of left and right monomers are expected. When one finds them to be all the same hand, design becomes a possiblity.
Granted, it is speculated a catalytic reaction may yield one-handedness, but this happens only under highly improbable chemical conditions.
There is more, but this is a start. Questins about flagellums, etc. are still a long way off. But first the basics.....
[ 24. May 2005, 23:06: Message edited by: Salvador T. Cordova ]
IP: Logged
|
|
Christopher D. Beling
Member
Member # 723
|
posted 25. May 2005 12:44
Hi Sal: I want to say that my understanding is improving through this brainstorm with you and others. I now see that the definition of CSI offered by Dembski (with out any direct recourse to K-complexity quantification) is best. I am now persuaded that the event (object) B* is to be classified as CSI because
(i) It is improbable below the UPB
(ii) It has a pattern.
So why bring in K-complexity at all? Well it is useful in detecting the pattern - right - and knowing there is a simple algorithm behind the sequence implies inteligence.
I would, however, ask you to address my point about the danger of sequence B* being due to regularity (hence no design) - this to me is dangerous - as the last thing we want for the complexity-specification criterion to accept a false positive. [A crystal of III-V semiconductor, i.e. gallium arsenide incidentally exhibits the 353535 ... regularity where the number corresponds to the valency of the atom traversed in one direction!]
I offer an attempted solution, hoping you will agree with me that there is a problem. One needs to place a question mark against events (objects)like B* that are TOO K-simple. [The K-simplicity can be quantified according to the symbol delta on p61 NFL that gives the length of the algorithm required to produce the CSI. i.e. delta must be less than something (dont know what?)]. So we are still left saying B* is CSI, but because its delta is too low - we cannot unambigously say it results from design. MOREOVER, such low-delta events such as B* should be common-sense checked to see if they possess simple periodicity (symmetry) in which case we must let the event fall through the C-S filter for fear of getting a false positive.
PROBLEM: The buckminsterfullerene also self assembles through natural forces so it appears that we may need more than just a simple threshold criterion on delta. We somehow need a basic symmetry check.
[ 25. May 2005, 12:57: Message edited by: Christopher D. Beling ]
IP: Logged
|
|
Christopher D. Beling
Member
Member # 723
|
posted 25. May 2005 13:13
Hi: Jerry,
I appreciate your pointing out that the 'complexity' being referred to in the term 'irreducible complexity'(IC) is of a quite different nature than in 'specified complexity'. In IC the complexity is referring to the connectivity and compatibility of the different system components. In SC (as Sal has pointed out) the complexity refers to extreme improbability (i.e over 500 bits of event information). The buckminsterfullerene exhibits IC in the connectivity and compatibility sense (take a away a few atoms and I believe it falls to pieces) but its structure can I guess be expressed in an algorithm of rather short length (i.e. it is K-simple) and thus it is not a SC object. Sal: this in fact is a problem for your Venn diagram - because you have K-simple IC systems placed in the CS domain.
Hi Andy,
I want to make a comment about your concern over not being able to identify conceptual information. Strictly speaking I guess you are right - it is impossible to know the information content of another mind. However, as Sal has pointed out, the string 11111111111... (i.e. all coins heads up on the floor) does strike a note of amazement in anyone. Why? I believe this is because the human mind sees here something REMARKABLE. The concept of everything being the same is separable by the human mind as something SPECIAL (remarkable). Personally, I think we must also take a common-sense approach to detecting REMARKABLE patterns when it comes to biology. Perhaps I wont be so popular in saying this, but I do think that life is REMARKABLE - and every child can see this - and most students want to explain it - it is our common-sense that tells us so. This REMARKABILITY stems from rather obvious things - like life reproduces itself (by incredibly complex mechanisms) - animal life moves around (by incredibly complex little motors called flagella - or in the case of humans - legs which perhaps are just as IC and SC in their detailed mechanism). Life metabolizes and dispenses energy also in a remarkable way (by many biochemical molecular machines). Life is both IC and SC - the patterns I have mentioned (reproduction, motility, metabolism) are patterns but not in some simple way as in the types of humanly conceived and generated strings (SI and CSI)that we have been recently considering. These are indeed the REMARKABLE patterns of nature - that demand from us an explanation. Moreover I think we can argue well that these patterns are not just anthropomorphisms (humanly subjective). In my understanding it is only with regard to these patterns of life that the CSI as mentioned in the 4th law is relevant? Yes, human minds can generate CSI that does functional things like ring telephones, open complex locks etc. Yes, human minds can also generate (CSI+IC) that makes functional/structural things like airplanes - but these are surely not to be equated/confused with the real antecedant entity - the self sustaining CSI of Life. Thus I would argue that in dealing with the 4th law we make a very clear distinction between humanly generated CSI (which does not sustain very long without human effort) and nature embedded (note I did not say generated!) CSI that does sustain on its own over geological timescales [which is to be found on Sal's Venn diagram in the IC+CSI region]. Can you agree?
[ 25. May 2005, 21:05: Message edited by: Christopher D. Beling ]
IP: Logged
|
|
andyg
Member
Member # 415
|
posted 25. May 2005 21:06
Sal,
I’ll deal with your posts all at once.
quote:
So let's start with 150 coins. You wish to be a master of self promotion. Your goal is to take those coins and some how figure out what sort of patterns, I the observer, would recognize as designed. What would you do? Most likely you'll arrange them in a k-simple structure.
Think about it. What conceptual patterns (specifications, conceptual information) would you use as a blue print to arrange the coins such that they conveyed physical information which would match the conceptual patterns I would recognize????
If you can answer that, at least intuitively by giving examples, we can then move on to biology.
My answer is that I would arrange them in some pattern that you would recognize as being either simply out of the ordinary (such as a k-simple pattern), or as part of a code (say Morse code for “Hi Sal, how’s it hanging?”, or just using the coins to make the shape of words).
However, there is a disconnect between the everyday examples of coins and dice and human knowledge of humans and biology. I will return to this point at the end of this message. Why don’t you give me a biological example of how the Designer has made its actions evident?
With respect to the Turing number, you’ll have to give me more details, I’m afraid. What is the number, and what is the evidence for it being actualized in cells?
quote:
However, biotic materials are not constrained by such requirements. Thus, the homo-chiral pattern in DNA and proteins are highly suggestive of design, since an a mostly random mix of left and right monomers are expected. When one finds them to be all the same hand, design becomes a possibility.
Granted, it is speculated a catalytic reaction may yield one-handedness, but this happens only under highly improbable chemical conditions.
It is quite a while since I completed my biochemistry degree in Great Britain, but even then I was aware of a variety of explanations for the homochirality of amino acids and sugars. I am sure that the literature has grown significantly since then. I question your assumption that a racemic mixture of amino acids or sugars would be "expected", since a) neither you or I are experts in chiral chemistry and b) no one has anything resembling full knowledge of the conditions in which the first pre-biotic molecules arose.
This speaks to the difference between "remarkable" things that we have a great deal of information about, and "remarkable" things we know very little about. Tying in a quote from Christopher:
quote:
However, as Sal has pointed out, the string 11111111111... (i.e. all coins heads up on the floor) does stike a note of amazement in anyone. Why? I believe this is because the human mind sees here something REMARKABLE. The concept of everything being the same is separable by the human mind as something SPECIAL (remarkable). Personally, I think we must take a common-sense approach to detecting RERMARKABLE patterns when it comes to biology. Perhaps I wont be so popular in saying this, but I do think that life is REMARKABLE - and every child can see this - and most students want to explain it.
As I mentioned to Sal earlier, coins are a simple system about which we know a great deal. We know that for most coins tossed in a fair way, the chances of getting a head or a tail are 1/2. Throwing 150 heads in a row would indeed be remarkable.
Another example I gave earlier was the mother scolding her cookie-snarfing toddler safe in the knowledge that "cookies don't walk away by themselves". The mother knows a great deal about the ambulatory capabilities of cookies, and the proclivities of naughty toddlers. Catching the cookies as they marched out the door would indeed be remarkable.
We know a great deal about the things that humans design. Finding a metal cog in the Burgess Shale, or finding the symbols "Hello Mom!!!" etched in frozen methane on Titan would be remarkable.
The existence of life in all its glory is "remarkable" in one sense, because it is so striking. But we know far less about the probability of life emerging than we do about the probability of cookies walking away of their own accord. In fact, I would suggest that calculating the probability of, say, the bacterial flagellum evolving is simply impossible. I have no idea whether it is 10^2, 10^150, 10^500 or 10^1,000,000. And neither does anyone else. The less one knows about the likelihood of something happening, the harder it is to label that event "remarkable" in a probabilistic way. As a result, invoking a design event by an appeal to a spurious notion of probability seems rather shaky - especially when the probability of a designer existing is not estimated at the same time!
IP: Logged
|
|
|
Printer-friendly view of this topic