“Life’s Conservation Law: Why Darwinian Evolution Cannot Create Biological Information”

_{William Dembski

May 1, 2009

Evolution, Informatics, Intelligent Design}

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

Here’s our newest paper: “Life’s Conservation Law: Why Darwinian Evolution Cannot Create Biological Information,” by William A. Dembski and Robert J. Marks II, forthcoming chapter in Bruce L. Gordon and William A. Dembski, eds., The Nature of Nature: Examining the Role of Naturalism in Science (Wilmington, Del.: ISI Books, 2009).

Click here for pdf of paper.

1 The Creation of Information
2 Biology’s Information Problem
3 The Darwinian Solution
4 Computational vs. Biological Evolution
5 Active Information
6 Three Conservation of Information Theorems
7 The Law of Conservation of Information
8 Applying LCI to Biology
9 Conclusion: “A Plan for Experimental Verification”

ABSTRACT: Laws of nature are universal in scope, hold with unfailing regularity, and receive support from a wide array of facts and observations. The Law of Conservation of Information (LCI) is such a law. LCI characterizes the information costs that searches incur in outperforming blind search. Searches that operate by Darwinian selection, for instance, often significantly outperform blind search. But when they do, it is because they exploit information supplied by a fitness function—information that is unavailable to blind search. Searches that have a greater probability of success than blind search do not just magically materialize. They form by some process. According to LCI, any such search-forming process must build into the search at least as much information as the search displays in raising the probability of success. More formally, LCI states that raising the probability of success of a search by a factor of q/p (> 1) incurs an information cost of at least log(q/p). LCI shows that information is a commodity that, like money, obeys strict accounting principles. This paper proves three conservation of information theorems: a function-theoretic, a measure-theoretic, and a fitness-theoretic version. These are representative of conservation of information theorems in general. Such theorems provide the theoretical underpinnings for the Law of Conservation of Information. Though not denying Darwinian evolution or even limiting its role in the history of life, the Law of Conservation of Information shows that Darwinian evolution is inherently teleological. Moreover, it shows that this teleology can be measured in precise information-theoretic terms.

Comments

Once again, the "meaning" of the transferred sequence depends entirely upon the context within which it is expressed.Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
07:35 AM
7
07
35
AM
PDT}

tragicmishap in #74: You make an interesting point. However, I'm not certain that there is necessarily a loss of meaningful information when the gene for human insulin is transplanted into a bacterium. If one defines the "meaning" of the transferred sequence as the protein for which it codes, this "meaning" has not been lost as the result of the transfer, as the protein is still made. However, if one defines the "meaning" of the transferred sequence as the biological function which the protein mediates, then indeed that "meaning" has been lost, because that function does not occur in the bacterium.Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
07:33 AM
7
07
33
AM
PDT}

joseph: Your comment #73 added no content whatsoever to this discussion, but rather was the equivalent of s**ting in the punchbowl. Every comment here, including both the positive and negative ones, has included substantive content except yours, which simply functioned as a rock tossed through the window. Whatever the moderators decide to do about your seemingly cumpulsive tendency to make ad hominem comments, it is clear to me that further interaction with you is pointless and counterproductive. From now on, regardless of whether they actually appear here, your comments are "invisible". Talk about "meaningless" information...Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
07:29 AM
7
07
29
AM
PDT}

Sorry I meant "Allen" not "Alan". The law of conservation of energy states that energy can neither be created nor destroyed, however it can be changed into an unusable form, which is detailed by the second law. These laws state that no work can be done without increasing entropy, that is losing energy to an unusable form. Perhaps the law of conservation of information will turn out to work the same way. For instance, Alan Fox's example of transferring the human insulin gene to a bacterium would cause no loss in information, but it would cause a loss of meaningful information.tragic mishap_{May 3, 2009
May
05
May
3
03
2009
06:59 AM
6
06
59
AM
PDT}

Is the ONLY counter-"argument" to act obtuse over the words "active information"? You guys must be really proud of yourselves.Joseph_{May 3, 2009
May
05
May
3
03
2009
06:51 AM
6
06
51
AM
PDT}

Nakashima #69: You're right about the apples and oranges, although the problem is easily fixed by normalizing both searches to a single-query search on an m-fold Cartesian product space where m is the maximal query count. In any case, the active information associated associated with single-hill fitness functions remains unchanged, so given that p increases by going to the Cartesian product means that the information problem only gets worse.William Dembski_{May 3, 2009
May
05
May
3
03
2009
06:45 AM
6
06
45
AM
PDT}

Alan: Far from stopping science, ID would then be a search for a context that makes information meaningful. Thus we could see someplace where there is information, defined as the exclusion of real possibilities, and then search for a context within which that information is meaningful. For instance, it is puzzling why the vertebrate eye is backwards when it could have been forwards. That would then inspire a search for a context which would make that information meaningful.tragic mishap_{May 3, 2009
May
05
May
3
03
2009
06:43 AM
6
06
43
AM
PDT}

I think that q represents the probability of finding the specific entity given the new search methods. For example, q is much higher than p once we know that the treasure is on Bora Bora. By the way Bora Bora is in my opinion the prettiest island in the world. I have certainly not searched for them all but a more efficient search was driven by articles on it, pictures of it, stories about it, actually visiting it, photographs I took of it and then comparing it with information on other islands etc. All of which have an information cost. And also my personal experiences as to what makes something pretty which may be analogous to whether the protein is functional or not. Or maybe it isn't analogous. But that hasn't nothing to do with the current paper.jerry_{May 3, 2009
May
05
May
3
03
2009
06:31 AM
6
06
31
AM
PDT}

Mr MacNeill, While you are pursuing interesting questions on meaningful information, I think you should not be so quick to identify that term with active information as used in the book chapter presented here. For one thing, the actual definition of information and how to measure it is a "princess in another castle" - note 48 refers to forthcoming articles. The authors do attempt a simpler exposition of what the calculation of active information is at that point in the paper. I must say that I find it confusing. In the WEASEL, for instance, Dawkins starts with a blind search whose probability of success in one query is roughly 1 in 1040. This is p. He then implements an alternative search (his evolutionary algorithm) whose probability of success in a few dozen queries is close to 1. This is q. And then proceed to calculate log(q/p). But this is comparing apples and oranges. In the above, p is expected success of a single query, while q is the expected success of a set of queries. I can think of two ways to fix this situation. In the first, we compare expected number of trials. 'q' is several dozen, p is half the size of the search space. But in this formulation log(q/p) will be the inverse of the previous formula. The second option is to compare the probability of success of a single query. Not a problem for p - all the queries have the same probability of success. But for q, what do we do? The queries of the first generation of the evolutionary algorithm are also random, they have the same probability of success as the first algorithm. The queries of the last generation have a far higher expected probability of success. You can't just take the average, that has other problems. I am really at a loss to understand why the authors perform a calculation using p and q when the definitions of each are so different. Perhaps the forthcoming articles will clarify this.Nakashima_{May 3, 2009
May
05
May
3
03
2009
05:58 AM
5
05
58
AM
PDT}

Dembski uses the phrase "meaningful information" once in the context of language. We can certainly have a thread of 300+ comments by everybody under the sun on the metaphysical significance of "meaningful information" but it may not have meaning with this paper. Maybe if we are confused we should concentrate on "active information" and try to get that into layman's English so we can try to understand the paper. Certainly we can also try to understand the Law of Conservation of Information and what that means in terms of layman's English. Does it mean that if a book is burned that we lost information, a potentially natural act, or does it mean something else. I am not sure any of us know so it may be fruitful to try and ascertain what it implies as opposed to sneer from our ignorance. If it is a vapid concept it will come out eventually. I believe that the idea behind the paper is to show that using an alternative search criteria to find something desired, for example, buried treasure, requires as much information to produce the search criteria as the blind search does. So we have a more efficient search but the information to enable this more efficient search has a price. To use Alan Fox's example of buying a map in a bookstore would require information to prepare the map, build the bookstore, find a bookstore that has the map, and then find the right map in the store.jerry_{May 3, 2009
May
05
May
3
03
2009
05:51 AM
5
05
51
AM
PDT}

And even more fundamentally, why is it called the “Law of Conservation of Information” when it doesn’t rule out the loss of information? As I understand it, the paper says there is no information loss.tribune7_{May 3, 2009
May
05
May
3
03
2009
05:20 AM
5
05
20
AM
PDT}

then the nucleotide sequence that specifies the amino acid sequence for human insulin is “active information” in humans, but not so in bacteria, even though it is the exact same sequence. Allen, as I understand the paper the active information in the sequence would be the difference between the probability of it occurring in a blind search vs. the probability of it occurring in the (successful) "alternative search". So, whether the sequence is used or not the information content would be the same. I comparison might be having a "men at work" sign at a work site or having it in a storage room.tribune7_{May 3, 2009
May
05
May
3
03
2009
05:19 AM
5
05
19
AM
PDT}

This thread has a very strange "feature". If I come into it first (that is, by entering UD from somewhere else) and then try to register to post comments to it, I get a message that says "Comments are closed". However, if I first go to some other thread and register, the comment box opens up there. If I then return here, the comment box is open here, too. Another weird side effect of "meaningful" information, perhaps? My attempt register to comment on this thread "means" one thing here, but something completely different at some other thread, even thought the information I attempt to enter is exactly the same. Context is everything...Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
05:04 AM
5
05
04
AM
PDT}

"Damn, damn, DAMN"
Relax Allen. From the standpoint of reason, the past few days here have not been so kind to you. Maybe its time to take a break. :-)Upright BiPed_{May 3, 2009
May
05
May
3
03
2009
04:18 AM
4
04
18
AM
PDT}

The birds are finally singing again; time to do something (anything!) else...Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:54 AM
3
03
54
AM
PDT}

...but the trail of "meaningless" bit strings is infinitely longer.Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:53 AM
3
03
53
AM
PDT}

Damn, damn, DAMN: left out an "it" in the mangling of Shakespeare in the previous quote...oh, the Bard (and my old director) would be so PISSED (indeed, he often was when I did this during performances, thereby altering the "meaning" of a line in ways the Bard of Avon never intended). To the trail of meaningful "words" there is no end...Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:53 AM
3
03
53
AM
PDT}

To all: Does the foregoing imply that "meaningful" information can, indeed, be created out of "airy nothing" (which seems to give a local habitation and a name), and therefore is there perhaps something not quite right about the concept of "active information", at least insofar as it is asserted that it must be conserved? Just curious...Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:49 AM
3
03
49
AM
PDT}

Sorry, I somehow dropped the phrase "appears to me that" following the word "it" in the first sentence of the next-to-last paragraph above, rendering its meaning less "meaningful". Perhaps my subconscious is playing little tricks on me, eh? Or maybe it's because I had a mug of cocoa/coffee with my lover's chocolate cake last night, and have been awake intermittently ever since. Interesting: the caffeine in the coffee and the phenylethylamine in the chocolate have a different "meaning" in the plants that make them. In the coffee and cacao plants, they function as herbivore repellents, but in me, they function as sleep repellents...Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:46 AM
3
03
46
AM
PDT}

If, indeed, "meaningful" information (as demonstrated earlier in this thread) is synonymous with "active" information, as defined by Dembski and Marks, then we have the very interesting situation described by Alan Fox in comment #52:
"What if that protein turned out to have a completely novel and beneficial function in humans, such as dissolving the plaques that cause Alzheimer’s? Would it then have more information?"
If we assume that such a thing is possible (and I can see no biological reason that it cannot be), then indeed a bit string that is "meaningless" in bacteria (i.e. one that has been completely artificially constructed by the experimenter, as described in Alan's comment) has come to be imbued with biologically "meaningful" information, simply by having it expressed in a different context (i.e. not by changing anything intrinsic in the bit string itself, and therefore not by changing its Kolmogorov or Shannon information). Let me return to Dr. Dembski and Dr. Marks' definition of "active information". If "active information" is indeed synonymous with "meaningful information" as demonstrated by the gedankenexperimenten performed above, then the "activeness" of "active information" is entirely a function of context. To me, this implies that genetic information (i.e. the information contained in the nucleotide sequences of DNA and RNA and the information contained in the amino acid sequences of the proteins for which they code) is rendered "meaningful" (and therefore "active") completely by its context within a living cell/multicellular organism. And, by the same argument, the "meaningful" information that makes up the organism is rendered "meaningful" via the interactions between that organism and its ecological environment. Which, in turn, seems to imply (at least to me) that for any kind of information to be "meaningful" in an evolutionary sense (i.e. for it to have evolutionary consequences), it must be expressed in an ecological context (in the same way that, to paraphrase G. Evelyn Hutchinson, an evolutionary "play" only becomes "meaningful" when performed on an ecological "stage". Ergo, it ultimately the ecological context of any form of biological information that renders that information "meaningful" (and therefore "active") in a biological sense. Which strongly implies (at least to me) that the "missing information" that Dembski and Marks refer to in their paper (i.e. the information that is "smuggled in" and that makes "active information" active) is just what evolutionary biologists say it is: the totality of the information contained in any given ecological environment which renders the biological information contained in the phenotypes (i.e. "bodies") of the organisms that live in it "meaningful". If this be error and upon me prov'd, I never writ, nor no information Were ever made "meaningful" By being expressed in context... ...maybe? (sorry, Bill)Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:41 AM
3
03
41
AM
PDT}

In #52 Alan Fox wrote:
"So I have a genetically engineered bacterium that contains the appropriate sequence for producing human insulin. I can use a broth of these bacteria to produce insulin commercially."
A very interesting point, because the human insulin made by the bacteria, which is OC exactly the same insulin as we find in humans, has no biological "meaning" for the bacteria. This is one of the reasons we we produce the human insulin in bacteria: they don't use it, and therefore don't modify it, rendering it of questionable use in humans. Note that only a specific sequence of nucleotides in humans produces a "meaningful" version of human insulin, whereas the exact same sequence of nucleotides in bacteria produces a completely "meaningless" version of exactly the same polypeptide. If I am reading Dembski and Marks' paper correctly (and it is OC quite possible I am not), then the nucleotide sequence that specifies the amino acid sequence for human insulin is "active information" in humans, but not so in bacteria, even though it is the exact same sequence. As Nakashima-san points out, context seems to be everything when it comes to "meaningful" information. In the same comment Alan Fox also wrote:
"I experiment by substituting a different sequence of the same length. I produce my new protein and it does not have any biological function that I can find. Which bacterium contains more information?"
I believe that if one is speaking of either Kolmogorov or Shannon information, the answer is that both the functional and the non-functional nucleotide sequences have exactly the same amount of information, but this is manifestly not the case if one is speaking of "meaningful" information. Ergo, we have the very peculiar (and to me, very interesting) situation in which the exact same sequence of nucleotides (i.e. theoretically the exact same string of bits/information) contains absolutely crucial "meaningful" information when produced in its full biological context (i.e. if a human can't "read" the information in this bit string, s/he is dead), but in a different biological context, the exact same string has no "meaningful" information at all. Indeed, if the accumulation of human insulin inside the bacterium has any negative effect on the bacterium (say, by reducing the ability of the bacterium to make its own "meaningful" proteins), then the exact same string of information can even have negatively "meaningful" information in the bacterium. Let me emphasize once again that the "meaningfulness" of the bit string is clearly not an intrinsic characteristic of the bit string itself. Rather, all of the "meaningfulness" of the bit string can only become manifest when it is "expressed" in a particular biological context. Once again I ask the question, does the foregoing interpretation of "meaningful" information match Dembski and Marks' definition of "active information"? I think perhaps it does (at least in some ways), but I'm not certain.Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:24 AM
3
03
24
AM
PDT}

In #51 Clive Hayden wrote:
"Replace “Dr. Dembski and Dr. Marks” with “Dr. Dawkins’ and the Weasel simulation” and you have a point."
OK:
”Ergo, it seems to me that none of the current theories of information, including the one presented in the post that heads this thread, address the question of “meaningful information” directly. It also seams to me that the question of where “meaningful” information comes from, and what makes it “meaningful” (and what makes other strings, such as those in non-coding DNA “meaningless”) may be tied in some way ~~to Dr. Dembski’s and Dr. Mark’s concept of “active information”~~ “Dr. Dawkins’ and the Weasel simulation”.
To me, that seems to have rendered this particular quotation meaningless, at least in the context of this discussion. Was that your point? Indeed, what exactly was your point, Clive? You gave no clue in your comment, and I must admit I'm too obtuse to get it now...Allen_MacNeill_{May 3, 2009
May
05
May
3
03
2009
03:04 AM
3
03
04
AM
PDT}

Alan Fox asks:
More generally, how can I claim that information is conserved if I can’t quantify it?
And even more fundamentally, why is it called the "Law of Conservation of Information" when it doesn't rule out the loss of information? Conservation laws are about quantities that remain constant.beelzebub_{May 3, 2009
May
05
May
3
03
2009
01:02 AM
1
01
02
AM
PDT}

PS Perhaps Paul Vitanyi should be credited with coining the phrase "meaningful information".Alan Fox_{May 3, 2009
May
05
May
3
03
2009
01:01 AM
1
01
01
AM
PDT}

The four papers mentioned above are all available on free access. Wolpert is hard going, Tom!Alan Fox_{May 3, 2009
May
05
May
3
03
2009
12:53 AM
12
12
53
AM
PDT}

It seems clear that in biology, “meaningful” information corresponds to functional proteins, nucleotides, etc. While the meaning of words in a human language is quite arbitrary, there is no such arbitrariness in biology; either the amino acid sequence is functional or not.
So I have a genetically engineered bacterium that contains the appropriate sequence for producing human insulin. I can use a broth of these bacteria to produce insulin commercially. I experiment by substituting a different sequence of the same length. I produce my new protein and it does not have any biological function that I can find. Which bacterium contains more information? I can see which bacterium is more useful; which bacterium contains more meaningful information, to take on Allen's word, but how can I know without testing the new protein. What if that protein turned out to have a completely novel and beneficial function in humans, such as dissolving the plaques that cause Alzheimer's? Would it then have more information? More generally, how can I claim that information is conserved if I can't quantify it?Alan Fox_{May 3, 2009
May
05
May
3
03
2009
12:44 AM
12
12
44
AM
PDT}

Allen, ----"Ergo, it seems to me that none of the current theories of information, including the one presented in the post that heads this thread, address the question of “meaningful information” directly. It also seams to me that the question of where “meaningful” information comes from, and what makes it “meaningful” (and what makes other strings, such as those in non-coding DNA “meaningless”) may be tied in some way to Dr. Dembski's and Dr. Mark’s concept of “active information”. Replace "Dr. Dembski and Dr. Marks" with "Dr. Dawkins' and the Weasel simulation" and you have a point.Clive Hayden_{May 2, 2009
May
05
May
2
02
2009
11:25 PM
11
11
25
PM
PDT}

Mr jlid, There is a position that the choice of chirality was arbitrary, and we use L-amino acids and D-sugars by chance. In our biology, an L-sugar is non-functional, not metabolized, which makes it attractive as a kind of sweetener. It would still "function" in that way! :) So meaning is still contextual, even in biology.Nakashima_{May 2, 2009
May
05
May
2
02
2009
10:50 PM
10
10
50
PM
PDT}

Allen, It seems clear that in biology, "meaningful" information corresponds to functional proteins, nucleotides, etc. While the meaning of words in a human language is quite arbitrary, there is no such arbitrariness in biology; either the amino acid sequence is functional or not.jlid_{May 2, 2009
May
05
May
2
02
2009
09:21 PM
9
09
21
PM
PDT}

Allen, Here are two references, one of which was discussed recently here but I think superficially The Capabilities of Chaos and Complexity by Abel http://www.mdpi.com/1422-0067/10/1/247 Information Theory in Molecular Biology by Adami http://arxiv.org/abs/q-bio.BM/0405004jerry_{May 2, 2009
May
05
May
2
02
2009
08:19 PM
8
08
19
PM
PDT}

Prev 1 … 3 4 5 6 7 Next

You must be logged in to post a comment.

Leave a Reply