|
Author
|
Topic: Evolution of complexity
|
warren_bergerson
Member
Member # 262
|
posted 26. October 2002 05:46
A number of points in the discussion here are worth noting.
First, there is a suggestion that some type of process or system(in mathematical space) interacting with an external environment(also a part of the abstract mathematical space) can generate or extract information. This, if it is verifiable, is a mathematical/information theory concept. It would seem that this class of information generating system or subsystem would be useful for modeling biological systems.
Second, the mathematical processes that have been studied which generate or appear to generate information( such as EA’s), appear to generate very small amounts of information (and at a very high cost to the information generating system or subsystem.) In order to be useful in modeling biological systems, it would first appear to be necessary to identify mathematical systems capable of generating or extracting large volumes of information with a very high degree of efficiency. Finding ‘efficient information/knowledge extractors’ is a mathematical problem.
Third, if biological systems can be modeled as information generators or information extractors, then the questions which need to be addressed are "what type of information is extracted?", "How much information is extracted and how fast?", "How and where is it stored?" and "How do we go about measuring this information?".
Fourth, as Janitor points out, it is not appropriate to claim that systems or processes which generates a few bits of information per year (an EA system) provides explanations for biological systems which generate vast quantities of information every second.
Fifth, Quote Janitor: What kind of information is contained in the genome and is it subject to meaningful statistical analysis?
I thought most of your discussion of information/complexity was clear and to the point. I do not, however, understand this comment. Why would you want to subject geneomic information to statistical analysis? Wouldn’t you want to analyze ‘What the information is, what it does and how it does it?". We can design measure the volume of information but, in most instances, the interest would be with the content of the information not its statistical characteristics.
Six, the analysis of information/design/complexity in biological system raises serious questions about the adequacy of current evolutionary theory. It seems, at least to me, very obvious that ‘information generating capacity’ of current theories is completely out of line with the information content of biological systems. Why isn’t this ‘problem’ recognized by people who claim "Evolutionary theory is widely accepted by the scientific community?".
IP: Logged
|
|
Janitor@MIT
Member
Member # 125
|
posted 26. October 2002 12:54
“If one takes your remarks seriously, they imply that every single scientific experiment on any phenomenon whatsoever is part of ‘an ID research program.’”—RBH
Aw, c’mon now, RBH, you really didn’t take my statements very seriously (and in any case they should taken only for what their worth, LOL), because I really didn’t say anything like that at all.
“Seems to me you've got some category confusions going on here.”—RBH
RBH also, “But then, the use of GAs to model biological processes, like another endeavour I could name, is heavily dependent on analogies from human engineering and design.”
Seems like we share some confused categories here! In fact, I do not make any hard categorical distinction between design and evolution. But I recognize your authority here (and I’m not being facetious or flattering either): Please explain to me the categorical distinction to be made.
“Another problem I'm having is trying to tell the difference between the occasions when Janitor is referring to local and temporary 'violations' of TSL, and when he is talking about global (universal?) violations… What am I missing?”—RBH
This is a good question because it addresses some fundamental misunderstandings of the SLT. My usage is consistent with common usage amongst physicists. I believe the confusion emerges from stating the SLT something like: The SLT is a law describing the (inevitable) approach of an isolated system to a state of equilibrium. Physicists are however aware that this is an idealization and not a statement of fact (!), as there really is no such thing as an “isolated” system. Further confusing the matter is that scientists generally do not use “law” in any consistent way. As in this case a “law” can describe something that no one actually believes exists! (Poincare showed, via an ergodic recurrence theorem, that the “law” doesn’t even apply to isolated systems! In the process the “statistical” character of the law was emphasized.)
The SLT does however introduce the terms and concepts of “entropy” and “equilibrium” which have far broader scope and application, but not in terms of a “law” as is colloquially understood. Instead we revert to terms which physicists don’t use much anymore because of their “fuzziness:” tendency, propensity, or disposition. In its extended (real world) application the “law” is statistical, so there really is no question of its “violation,” “universality,” “locality,” or “temporality.” The SLT really applies statistically only to “open” systems.
Although we sometimes insist that the universe is isolated, that doesn’t even make much sense, because if the universe is isolated, what exactly is it isolated from? Its not really a statement of any physical significance, if you take my meaning, instead it is stated so for the purposes of simplifying causal explanations. I.e., if the universe is open causally than all our causal explanation must be conceded to be possibly incomplete. But for all practical purposes, with any sufficiently complex phenomena, we are willing to concede that anyway. But then that leads full circle back to your statement I quoted at the top of this post. Thoughtexperiment: Two scientists performing identical experiments with identical results. One scientist perusing the results shouts, “Eureka! Thank you National Science Foundation!” The other scientist peruses the same results and shouts, “Hallelujah! Praise Jehovah!” Which one is not a scientist?
“I'd sure like to see some explication of "poorly conceived." In what respects are the experiments poorly conceived and how would you correct them? Generic and allusive (and elusive!) critical remarks are one thing; constructive criticism with concrete suggestions for improvement is another (and harder) thing to do. How about taking a shot at the latter?”--RBH
Adami & Cerf (arXiv: adap-org/9605002 v3 17 Apr 1999, which is the preprint of Ref. 4 in the PNAS article) seek to define a “physical information” by supplying what is missing from the equation statistical or physical entropy + Kolomogorov-Chaiten theory. They say that what is missing is an “environment.” But I’ll state it even more fundamentally and also more usefully and generally: What is missing is background or prior information. It can be anything that supplies the “missing factor” here and doesn’t have to be the “environment.” Cerf & Adami’s own words indicate this, as they say that the axioms of mathematics are implicit only in Kolmogorov-Chaiten. Well, I can just as easily make them explicit and substitute the axioms of mathematics for the environmental factor Adami & Cerf supply.
But actually, following precedent in design theory, I will substitute what is not exactly either the axioms of mathematics or the environment. I will instead substitute the “axioms of design,” “imperative logic of design”, or the “algorithms of design.” Under the “assumption,” which I believe is reasonably well warranted, that intelligent designers design intelligent designs, I don’t make the assumption necessary (and otherwise trivial) that life forms are correlated with their environments. (At least not in the way that Adami et al envision.) Why? Because designers don’t consider it a particularly intelligent design strategy. It assumes far too much. It assumes far too much both about the environment, the designs, and the designers.
Instead, the intelligent designer designs for adaptation of the designs. He provisions (pre-adapts) the intelligent design not to exhaustive pre-specified sets of environmental conditions, but in so far as possible, with the capability (which we strongly identify with “intelligence”) to adapt to conditions and especially to adapt to changing conditions. (This is what is commonly referred to in engineering as "intelligent design.")
This kind of information incorporated into the design (the background and prior information) does not take the form of statistically correlated bits of data. It is an algorithm(s). It is what makes possible Adami et als search for environmental correlations. They are there. But what they are missing is what makes it even possible for such correlations to exist and what they aren’t even looking for: adaptive algorithms. These algorithms are what Adami et al call a “Maxwell demon-like mechanism.” (Per the theoretical conventions of Darwin they posit this "algorithm" in the environment and not in the genome. But see the Kargupta article I cited.)
Frances is understandably confused (partly my fault) about what a “kind” of information is. Shannon information is both a “kind” of measurement of information (statistical) and a kind of information also. But I’m talking about a kind of information, and the possibility of its measurement, which has otherwise escaped all information-theoretic analysis: algorithms. An information-theoretic analysis of algorithms is necessary to extend information theory and to apply it to biology, and also necessary for its own sake, because the lacuna in the theory (recognized by Shannon) is the operational definition of information. Until the analysis proceeds I believe I am well warranted in stating that the information is contained solely and wholly in the operations and not in the instances and sets.
Does that help, Frances, RBH? Or does any of it even make sense? I’m sure it doesn’t to a “Darwinist,” so I offer it only, for whatever its worth, for anyone willing to think “outside the box.”
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 26. October 2002 15:33
Dear Janitor,
I appreciate your attempts to clarify your position but lets step back for a moment. I showed how researchers had argued that a 'Maxwell demon' like could be how information is transformed from its environment to the organism's genome. The term Maxwell Demon seems to have generated some confusion when it was assumed that the Demon had to violate the 2LOT in some manner and that Schneider had shown this to be impossible. As I showed, a more careful analysis of Schneider showed that in fact the Demon was quite costly in that it required 5500+ deaths to flip one bit in the genome. So now we are back to how the information increases in the genome. Information or complexity is measured in terms of Shannon information/entropy here.
How does information increase and how does it involve the correlation with the environment?
H(Max) is the entropy of the system assuming equiprobable distribution. For the genome such an assumption seems quite acceptable since the bases have no affinity to their neighbors.
The difference between this equiprobably distribution and what is actually found is refered to as information and is described by the following formula
Information stored by population A about environment E:
H(max) - H(A|E) = L - H[p_i]
where p_i is the actual probability of finding the genome given the environment.
A short course in information theory
An interesting paper which I will revisit in a different thread.
When Adami et al argue that the environment imposes the information, Janitor suggests that But I’ll state it even more fundamentally and also more usefully and generally: What is missing is background or prior information. It can be anything that supplies the “missing factor” here and doesn’t have to be the “environment.”
In fact the missing factor could be an intelligent designer who organizes the genome in exactly the correct order to correlate strongly with a preset environment. Thus leading to the question of adaptive algorithms.
I fail to see how this makes Adami et all poorly conceived, their adaptive algorithm is RM&NS. Whatever the terminology, what is shown is how 'adaptive algorithms' can be applied by either intelligence or nature to generate information in the genome.
Thus the idea that only intelligence can generate CSI in the genome seems contradicted by these observations.
I wonder though why Janitor needs to fire a parting shot towards 'darwinists'
quote:
Or does any of it even make sense? I’m sure it doesn’t to a “Darwinist,” so I offer it only, for whatever its worth, for anyone willing to think “outside the box.”
I would say that your comments make very good sense, especially to a Darwinist.
And Shannon entropy has become quite a foundation for the development of cryptographic algorithms. Whatever the algorithm underlying the process, it's the shared entropy between the organisms and its environment that is the relevant factor.
quote:
More specifically, information is defined as the correlated entropy between two ensembles [2], and the ambiguous concept of entropy [3] represents a fundamental link between thermodynamics and information theory [4].
Source
Warren raises some good questions. The fact that EA's can generate information does not mean that they can generate enough information. That of course is a reasonable argument for additional research but at the moment I am interested in a more basic concept, can EA's generate information? The answer seems to be yes.
[ 26. October 2002, 17:43: Message edited by: Frances ]
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 27. October 2002 08:55
Frances,
Quote: Warren raises some good questions. The fact that EA's can generate information does not mean that they can generate enough information. That of course is a reasonable argument for additional research but at the moment I am interested in a more basic concept, can EA's generate information? The answer seems to be yes.
The basic issue is whether any mathematical system or subsystem can generate and/or extract information. One of the basic mathematical principles or concepts of information theory is that a closed system can not generate information or increase the volume of information. I am assuming this is a valid mathematical for a closed system.
The first mathematical question raised here is whether there are operations or processes which can increase the information in an open or interactive subsystem which exists within a larger closed system or universe. It seems intuitively obvious, that if information can be transferred from one subsystem to another that it must logically be possible to increase the information content of a subsystem. It would therefore appear that information theory needs to be modified or expanded to include the analysis of processes by which subsystems in a larger system increase or extract information.
If as seems likely, subsystems can increase information content, then it seems appropriate to analyze and model life forms as systems which increase information as a result of interactions with the environment. Note, the question of whether biological systems are capable of extracting information from the environment is irrelevant. The question is whether life forms can be usefully modeled as systems which increase information.
If it is possible to model biological systems as systems which generate information or extract information from the environment then the next logical question is what type or class of mathematical system should be used to model the information generating capacity of biological systems. The choices available are:
1. A ‘Demon’ which quote "As I showed, a more careful analysis of Schneider showed that in fact the Demon was quite costly in that it required 5500+ deaths to flip one bit in the genome".
2. An EA process which probably has similar efficiency characteristics. Or
3. Forms of mathematical systems which can generate vast quantities of information with little or no cost in lives.
In the absurd world of academia, we have physicists arguing we must analyze information in biological systems in terms of thermodynamics and evolutionary biologists arguing we must analyze information in biological systems in terms of EA processes. It should, but apparently is not, obvious that neither demons nor EA processes have anything even vaguely approaching the capacity to explain the information generating/extracting capabilities of biological systems.
There are perfectly sound alternatives to Demons and EA processes. If scientists are really serious about analyzing life forms in terms of information theory, then they need to start looking a mathematical processes and systems that bear some resemblance to the information generating/extracting capacity of real world biological systems.
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 27. October 2002 19:34
Warren states: One of the basic mathematical principles or concepts of information theory is that a closed system can not generate information or increase the volume of information.
An interesting question. Since information seems to be correlation with respect to 'something' one may wonder if information can arise in closed systems (although one may want to define the meaning of closed system). If the relationship between information and thermodynamical entropy holds then that may be a valid conclusion.
Concepts of self-organization and 'autocatalytic' processes may have something to say about this.
But I digress...
Warren claims without much supporting evidence that quote:
It should, but apparently is not, obvious that neither demons nor EA processes have anything even vaguely approaching the capacity to explain the information generating/extracting capabilities of biological systems.
This seems hardly self evident nor would I consider this assertion to be supported by any appeal to fact. Perhaps Warren may have something more to say about his claims. Arguments like 'it should be self evident/obvious" are perhaps useful as rethorical tricks but in a scientific discussion they should be avoided lest they can be supported.
Nor is it self evident that the choices are exclusive, in fact I would not be surprised if there are other mechanisms than natural selection/mutation to deal with the flow of information in biological systems.
Finally I wish to ask Warren if he considers himself to be a scientist, serious about analyzing life forms in terms of information theory? If so I am looking forward to mathematical foundations which can find support in the available evidence.
I would argue that there are quite a few serious scientists out there who surely have proposed a very powerful foundation to information and the genome. Perhaps its time to show that there are additional approaches which can similarly contribute?
But in the end Warren seems to support my thesis that in principle one cannot argue that there are no natural pathways to increasing information in the genome. We may argue about the relevance of the various processes but one cannot use the presence of (complex/specified) information in the genome as evidence against natural processes, let alone as evidence for intelligent design.
So now the question that would be of interest to this board is: Can we expect ID to provide us with hypotheses that can be tested? Can we restrain ID so that it can be used as a scientific paradigm or is ID 'doomed' to remain an inference based on our paucity of knowledge? I fully believe that hypotheses of intelligent design are possible, we need not look further but at SETI or archaeology or forensics.
In an attempt to provide with additional mechanisms to increase information which do NOT require the death of countless organisms, I would like to point to a new paradigm in evolution _adaptive mutations_. But before I digress I would like to point out that ID or alternative mechanisms cannot do better in increasing information, that is ID needs to 'waste' as much information/energy.
"adaptive mutations", a bit confusing in that it suggests that contrary to Darwinian principles the mutations are preferentially beneficial, is now quite well understood. Non-somatic cells under stress enter a state of hypermutation. And while in many organisms, hypermutations are limited to somatic cells and thus are not passed along. In bacteria however, the sharing of plasmids can help spread such mutations quickly.
as such hypermutations may very well increase the information in the organism but not the genome since for instance acquired immunity through such mutations is not passed on.
In case of such mutations the 'cost' is in replication being suspended until the source of the stress has been addressed (or the bacteria die).
A fascinating hypothesis would be to combine gene duplication with hypermutation as an opportunity to evolve new information.
[ 27. October 2002, 20:57: Message edited by: Frances ]
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 28. October 2002 02:08
Janitor - I'm completely swamped, and it'll be several days (at least) before I'll have time to return to your remarks. I'll try this week, but no guarantees!
RBH
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 28. October 2002 14:02
Frances,
In answer to your question, I consider myself an applied scientist. One of the key differences between applied scientists and academic scientists is that applied scientists require that pieces of knowledge fit together in a consistent logical structure. A fair portion of academic scientists appear to be totally unconcerned with whether or not different pieces of knowledge are compatible.
The concept of design includes the concept of information. It therefore makes sense that information theory would be an appropriate form of mathematics for analyzing genomic design and change in genomic information. You suggest that demon’s might be a useful starting point for analyzing genomic information and RBH suggested EA’s might be applicable.
When I point out that the ‘information producing capacity of demon’s and EA’s is not of the same magnitude as the magnitude of genomic information generated by living systems you respond:
Quote WB: It should, but apparently is not, obvious that neither demons nor EA processes have anything even vaguely approaching the capacity to explain the information generating/extracting capabilities of biological systems.
Quote F: This seems hardly self evident nor would I consider this assertion to be supported by any appeal to fact. Perhaps Warren may have something more to say about his claims. Arguments like 'it should be self evident/obvious" are perhaps useful as rethorical tricks but in a scientific discussion they should be avoided lest they can be supported.
Your comments imply that scientists have no techniques for measuring the information generated by either or both 1)biological systems or 2)demons and EA’s. Only if they lacked one of these two pieces of information could they be unaware of the vast disparity between the volume of information generated/extracted by biological systems and the volumes of information generated by EA’s or demons. You are suggesting that scientists are discussing a topic and publishing papers on a subject with out even the most elementary knowledge of the subject. It would clearly appear that you are the one playing rhetorical games. If you wanted to know how I measure genomic information you could have asked instead of stating your opinion as a personal attack.
To begin, there are undoubtedly many different possible ways of measuring volume of genomic information. In the approach I use, information at a point in time is defined in terms of a form Dxt, which is a member of a set of possible forms D and also a member of Dft the subset of functional forms at time t. The volume of information at a point in time is measured by the size of D divided by the size of Dft. The amount of information a system generates is defined by D/Dft times the average rate at which Dft changes so as to require a change in Dxt.
The approach used here recognizes that genomic information is generated at the level of individual cell on an ongoing basis. A unit or bit of genomic information is defined by the combination of a gene or allele and active/inactive status of the gene at a point in time (there obviously may be more than just active and inactive status). At a point in time each unit thus has two possible forms ( D has size 2) and only one form is functional at a point in time (Dft has a size 1). If an organism has 1 trillion cells, 10,000 genes, and each cell has an average of 1000 changes in gene status per day, then the organisms would generate ‘1000 trillion’ bits of information per day. A volume vastly greater than 1 bit every 550 lives. Similar measures of information generation can be calculated for nervous system operation.
Biological systems generate/extract vast quantities of purposeful or teleological information. Understanding the materialistic processes by which biological systems generate this information is key to understanding life systems. Information theory, particularly information theory that recognizes the information generating capacities of certain types of logic machines, is key to understanding life forms. Maxwell demons and EA systems are simply too primitive to provide any meaningful knowledge of how biological systems generate information and designs.
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 28. October 2002 14:32
Warren: I am not sure if your observation about 'a fair number of scientists appear to be totally unconcerned with whether or not different pieces of knowledge are compatible."
But I agree that often the approach involves the comparisson of one or more theories/hypotheses against the available data. Unification of ideas first requires the collection of data and the formulation of many hypotheses before one may attempt to unify a concept.
You point out or actually more correctly, you assert that there is a difference in magnitude between the information producting capacity of the demon(s) and the the genomic information generated by living organisms. Your response to my request for additional data to support what you consider to be obvious is countered by you arguing a strawman namely that I am implying that scientists do not have techniques for measuring the information of 1) biological systems 2) demons and EA's. On the contrary I am arguing that I have yet to see that such an analysis show that there is an obvious discrepancy between the two.
I am arguing not that scientists are discussing these topics without any knowledge in this area but rather that your argument seems to be lacking in supporting evidence. It's a challenge on my part to find out if your comments have been carefully thought and by looking for supporting and falsifying evidence or that they are merely 'obvious' in the sense of intuition. The use of terms such as 'obvious and self evident' have been used by you in the past and I have found, empirically that they often remained unsupported. Thus my reluctance to accept 'obviousness'. Your 'calculation' shows that organisms seem to have enough information generating capacity. But you seem to forget one important issue, a flipped bit is not necessarily information. Information indeed requires some correlation between the object and its environment. If most of the mutations can be shown to be neutral or even detrimental then a lot of your 'potential' information evaporates quickly.
That biological systems generate teological information is irrelevant to the discussion of genomic information. Unless you argue that our learning and actions are translated into genetic information (Lamarckian?).
Perhaps it is important to differentiate between "Maxwell demonst and EA systems" are simply to primitive to provide any meaningful knowledge of how biological systems generate information and design" verus "can maxwell demons and EAs' explain the _GENOMIC_ information? You surely realize that you were arguing a strawman argument by switching from the specific argument to a more general case. In fact iirc, I actually differentiated between to two topics in my response.
So before it seems to me that obviousness was used in the meaning of 'intuitive' rather than quantifiable and seems to suggest that the idea of information is a tricky concept in that not all genetic change is an increase in genomic information. Perhaps it may be helpful to look at Shannon's original work which shows how information is quantifiably defined in a manner inconsistent with your intuitive notions of information being just flipping bits.
Thus my argument remains, can you show in a quantifiable manner that does justice to the concept of information, that there is an obvious discrepancy between information in the genome and the potential of EA's and Maxwell Demons?
I will be glad to provide you with some useful links about the mathematical theory of communication papers by Shannon.
The Information Content of Binding Sites on Nucleotide Sequences
Schneider gives some insight in the (un)likelihood of mutation explaining the human genome information content (~ 4x10^9 bits which assumes 1 bit per base (an overestimate according to Schneider), the human genome could evolve in a billion years without adding any additional processes
[ 28. October 2002, 14:52: Message edited by: Frances ]
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 29. October 2002 11:19
Frances,
Effective, successful, productive scientific analysis of biological complexity/design, IMO, requires solutions to two fundamental types of problems. First, effective analysis depends on solving the numerous and complex ‘how to’ technical challenges. Second, and equally as essential, successful analysis of biological complexity requires avoiding and/or circumventing the counter-productive academic restrictions on analysis.
In the same way that ‘too many cooks spoil the broth’, too many academic experts or pseudo-experts impose restrictions and unnecessary and unproductive requirements that make solving a problem impossible. In what amounts to little more than ‘turf wars’, academic authorities impose restrictions and requirements that make it extremely difficult to implement new, multiple discipline solutions to problems. The subject of information generating systems is particularly interesting because it offers the possibility of solving both the technical and the political challenges associated with analyzing biological complexity and design.
Of concern here is a set or class of mathematical logic machines which generate or extract information by interacting with an external environment. Of particular interest here is the subset of information generating machines that generate information using selection processes not limited to Natural Selection and variation generating mechanisms not limited to random or chance variation. I will be glad to discuss in detail this class of logic machines either here or off-line.
Non-Darwinian information generating logic machines offer ‘interesting’ solutions to the ‘political’ challenges surrounding the scientific analysis of design.
To begin, non-Darwinian logic machines are ‘outside the established scope of expertise’ of any of the existing fields of science, math, or philosophy. Evolutionary biology and physics have not recognized even the possibility that such machines may exist and may have manifestations in the real world. Individuals in these fields can not legitimately claim expertise on a subject whose existence has not yet been recognized. Current forms of information theory don’t recognize the existence of such machines. Current concepts and principles of philosophy of science don’t recognize the possibility of scientific theories constructed from such complex logic machines. Non-Darwinian information generating machines are, at least in part, a new concept in science, mathematics, and philosophy.
The fact that this class of information generating machine is new doesn’t mean that existing scientists, mathematicians and philosophers are not qualified to perform analysis. it simply means that blank assertions of the type ‘this must be analyzed, addressed, viewed or approached from this approach’ are not valid. It also means that assertions of the form ‘this must be demonstrated from the existing literature’ or ‘this must be acceptable to current peer review processes’ are not valid.
Second, the mathematics and analytical techniques associated with these machines are engineering mathematics and techniques. This means these logic machines can be evaluated in terms of objective, quantifiable performance measures and standards.
Third, information generating logic machines can be applied to the analysis of a wide range of biological design processes besides evolution. In particular, these mathematical machines can be applied to the analysis of development processes, ongoing adaptive processes, nervous system functioning and human intelligence.
Information generating logic machines will not be of much interest to those who believe scientific analysis consists of 1)a credential, 2)generating a list of references from the literature, and 3)asserting your subjective opinion. But for those interested in real scientific analysis, these machines offer an interesting opportunity for analyzing and understanding life forms.
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 29. October 2002 23:15
Warren,
Due to my new imposed limitations on the number of allowed postings on ISCID/Brainstorms I had to determine which posting I would reply to first. Given that I am still working on a more focused response to Dembski, I have chosen this thread as the 'lucky winner'
I am somewhat confused where you are going with your arguments. Lets refocus rather than delving into arguments of culinary methods (...spoil the broth...).
You made an assertion and I asked you to provide some supporting evidence.
quote:
the information producing capacity of demon’s and EA’s is not of the same magnitude as the magnitude of genomic information generated by living systems
I asked you to provide for supporting data. You suggested that the organism would generate '1000 trillions' bits of information. I pointed out to you that flipping bits is not the same as generating information. Information in the Shannon sense for example is quite well defined and does not seem to be related much to your concept of information. Thus perhaps you could explain to us what measure of information you have chosen and how you want to compare this to the 1 bit/550 lives of Shannon information generated in scenarios of evolution? As I have shown Schneider and others have shown how simple mechanisms seem to be quite able to explain the amount of Shannon information in the genome. So not only does it not seem to be obvious but also erroneous IMHO.
Now onwards to some of your other claims:
May I ask you how how claims of 'counter-productive academic restrictions on analysis' have any relevance to our conversation or even reality? I am neither academic, nor am I placing many restrictions on analysis. In fact I am open to you explaining to me how you reached your conclusions. As a scientist I am also quite curious but that may be a character flaw of mine.
I am also not asking you to show peer reviewed research or anything else other than to support a claim which involved the term obvious.
I am also quite well aware that there are countless information generating mechanisms active outside the genomic information present in our bodies and I am very aware that these amounts of information may (or may not) be quite impressive. But they do not have much relationship to the issue at hand.
So back to obvious:
quote:
Obvious
Easily discovered, seen, or understood; readily perceived by the eye or the intellect; plain; evident; apparent; as, an obvious meaning; an obvious remark.
Now it may be my intellect that is the problem in perceiving this obviousness in your claim thus please allow me to delve into these issues deeper by asking you to help me see the obviousness of your claims. Friends are known to have me called thick-headed occasionally so I am totally open to the possibility that this is all due to me.
Let me finish by stating that your parting comments
quote:
Information generating logic machines will not be of much interest to those who believe scientific analysis consists of 1)a credential, 2)generating a list of references from the literature, and 3)asserting your subjective opinion. But for those interested in real scientific analysis, these machines offer an interesting opportunity for analyzing and understanding life forms.
make total sense to me. I could not have said it better I am sure. I also do not beleive that arguments from authority or arguments by assertion are very useful that's why I am asking you to provide some non-subjective support for your opinions. Unless the word obvious is totally subjective in its interpretation. That of course would be another explanation for why I may not see it the same as you do.
I also totally agree that science has shown indeed a real interest in pursuing the research of complexity and information and have done an excellent job in presenting their ideas, hypotheses, and theories.
All I am asking you is to address the obviousness of your claim. Is your argument that the organism can generate far more information than EA/Maxwell demons can generate in these simulations? Is your argument that one would expect EA/Maxwell demons to explain this form of information? Is your argument that the amount of information that one expects to have to be explained by EA/Maxwell demons is much larger than their capacity?
You seemed to be making the claim wrt evolutionary theory so correct me if I am wrong.
"the analysis of information/design/complexity in biological system raises serious questions about the adequacy of current evolutionary theory. It seems, at least to me, very obvious that ‘information generating capacity’ of current theories is completely out of line with the information content of biological systems."
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 30. October 2002 11:54
Frances,
Quote: You suggested that the organism would generate '1000 trillions' bits of information. I pointed out to you that flipping bits is not the same as generating information.
It is difficult to take any part of your argument seriously when you start by suggesting that on-off switches and changes in on-off switches are not relevant to the discussion of information and the generation of information/complexity/design. If you look at changes in genetic switches or changes in regulatory genes you clearly see that biological systems generate/extract vast quantities of information. To suggest that a process with the ability to generate 1 bit of information per 550 lives provides a basis for explaining a systems that generates ‘1000’s of trillions’ of bits of information in each individual each day, is at the very least a questionable suggestion.
One of the central problems with the discussion of design/complexity is the attempt by certain groups, particularly it appears, physicists and evolutionary biologists, to restrict the discussion of complexity/design to a set of phenomena that typically change very slowly or not at all. This manipulation of the definition of complexity/information is used to justify the suggestion that biological design can be produced/explained by very, very weak EA processes.
If you use even semi realistic definitions of design/complexity, then the whole evolutionary model of generating biological complexity falls apart. If you use realistic definitions of biological complexity/design, much of the ID argument for ‘evidence of design by an external designer’ also falls apart.
Quote: I am also quite well aware that there are countless information generating mechanisms active outside the genomic information present in our bodies and I am very aware that these amounts of information may (or may not) be quite impressive.
This is an interesting statement. You claim to be aware of information generating mechanisms while the information theory you claim to understand suggests, I understand, that closed systems cann’t generate information. If you are aware of mathematical processes which can generate vast quantifies of information, then you must similarly be aware that there a vast number of alternatives to neo-Darwinian theory and that some of these alternatives provide a much better fit to evolutionary data that neo-Darwinian models.
Quote: Now it may be my intellect that is the problem in perceiving this obviousness in your claim thus please allow me to delve into these issues deeper by asking you to help me see the obviousness of your claims. Friends are known to have me called thick-headed occasionally so I am totally open to the possibility that this is all due to me.
In addressing certain types of problems it is sometimes said that solutions are either 1)obvious or 2)obviously flawed. If you understand the perspective from which a problem is being addressed and if you understand the techniques being used to address the problem then a proposed solution is either obvious (follows logically and inevitably from perspective and techniques used) or is obviously flawed (the solution does not work if you recognize and adjust for an identifiable flaw in the proposed solution).
This obvious/flawed dichotomy assumes that the individuals evaluating the validity of a proposed solution have the technical knowledge/ability and integrity to evaluate the proposed solution in the manner it is presented. If the group of individuals understand the issues involved, if an objective answer exists, and if the individuals involved are ‘honest’, then they should be able to reach agreement that a proposed solution or position is either obvious or obviously false. The process of reaching agreement may at times be long and acrimonious, but if the appropriate conditions exist then agreement can be reached. IMO, differentiating between obvious and obviously flawed is what science is or at least should be about.
All too often both here and in academia, we find individuals expressing authoritative opinions on subjects without understanding either the approach being used or the technical concepts involved. These subjective unqualified opinions then become the ‘literature’, the ‘logical precedent’ on which other subjective opinions are based. At some point in this authoritative process, individuals not only express opinions without the technical qualifications, but they start arbitrarily accepting and rejecting facts to support their arguments, and the begin arbitrarily accepting or rejecting both facts and forms of technical analysis. [ Your comment that "I pointed out to you that flipping bits is not the same as generating information" is a prime example of such arbitrary and ingenuous manipulation of facts. On-off switches, as you are well aware, is the basic unit of information in computers. ]
My concern is not so much with flawed evaluation practices in academia/peer reviewed literature, but with finding some way to avoid the flaws. The subject of ‘information generating machines’ (shall we label them IGM’s) seems to be an excellent opportunity to circumvent existing peer review processes. IGM’s are clearly outside the realm of recognized knowledge therefore evaluation of IGM’s need not and should not be based on the literature. IGM’s are clearly complex mathematical phenomena whose characteristics can be analyzed by mathematicians using quantifiable engineering techniques. Although it may not be apparent to individuals who have not worked with IGM’s, they are complex deterministic causal relationships which are capable of generating and modifying other causal relationships. As such they provide a potential basis for formulating scientific theories of the operation of life forms.
I am more than happy to present the details of IGM’s and to discuss/debate whether the conclusions reached from the use of these mathematical entities is obvious or obviously false.
IP: Logged
|
|
|
Printer-friendly view of this topic