|
Author
|
Topic: The Origin of Biological Information
|
Doug Wedel
Member
Member # 1901
|
posted 05. March 2006 19:54
As everyone here knows, a debate is raging about the origins of biological information. Since this audience is so well versed on these matters, I will not provide any additional background to the topic except to say that the extreme positions on the question seem to be nicely bookended by two quotes:
From Manfred Eigen:
quote:
Our task is to find an algorithm, a natural law that leads to the origin of information.
From William Dembski:
quote:
Algorithms and natural laws are utterly incapable of producing information.
With your permission I would like to try to carefully develop a proof that Eigen is right and Dembski is wrong. I propose to do this by first describing a rather difficult problem in nonlinear dynamics which requires that one determine the values of 10 interacting variables. After stating the problem as clearly as I can, I then propose to ask participants in here if they have any idea how to solve this problem. It is my hypothesis that no one will be able to solve it. Then I will demonstrate how the problem can be solved algorithmically through random number generation combined with artificial selection.
Before I do this, though, I would like to engage in a bit of preliminary dialogue to clarify what is precisely meant by Dembski's comment above, or more specifically, what is meant by the No Free Lunch theorem on which I believe he bases his assertion. Could anyone state for me as explicitly as possible without getting too technical _exactly_ what is meant by the assertion that "algorithms and natural laws are utterly incapable of producing information"?
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 05. March 2006 20:49
Because by defintion, Information is independent of the natural environment?
IP: Logged
|
|
Doug Wedel
Member
Member # 1901
|
posted 05. March 2006 21:28
I'm not sure what you mean by "independent of the environment." What does information reference if not "the environment," aka the real world? And also, doesn't information have to be embedded somehow in physical switches of some type in order to have real-world functionality?
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 05. March 2006 22:16
Perhaps we first need to define what is meant by Information?
The blueprints of an aircraft on paper is Information about the design of the aircraft; however, aircraft-design is "independent" of the physical properties of flattened wood-pulp, and ink.
A Pulsar emits a consistent, regular pattern (at least in relation to a fixed observer). While the pattern provides information "about" the Pulsar, the pattern isn't really considered "information." At least in a No Free Lunch sense. However, if the Pulsar emitted the decimal notation of Pi (i.e. 3.1415926535...), then the emission would "contain" information. The physical properties of the emission being merely "transport" in an OSI stack, communication protocol sense... The emission contains information, because the information is Independent of the physical environment of it's transmission.
IP: Logged
|
|
Doug Wedel
Member
Member # 1901
|
posted 05. March 2006 22:53
When I Googled OSI I saw that you are bringing to the table here a rather extensive conceptual framework with which I am _totally_ unfamiliar. Nevertheless, let's try to narrow down your idea a little further. You say:
quote:
A Pulsar emits a consistent, regular pattern (at least in relation to a fixed observer). While the pattern provides information "about" the Pulsar, the pattern isn't really considered "information." At least in a No Free Lunch sense. However, if the Pulsar emitted the decimal notation of Pi (i.e. 3.1415926535...), then the emission would "contain" information.
I am trying eliminate ambguity, so let me ask you: if the emission from the pulsar contained a stream of digits that _never_ included an '8' would that also mean that the pulsar was emitting contained "information" in your view?
IP: Logged
|
|
IF
Member
Member # 1904
|
posted 05. March 2006 23:09
Irving,
Doesn't your definition of information presuppose consciousness? There can be no information without an informer and an informed?
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 06. March 2006 07:25
quote:
I am trying eliminate ambguity, so let me ask you: if the emission from the pulsar contained a stream of digits that _never_ included an '8' would that also mean that the pulsar was emitting contained "information" in your view?
No
IP: Logged
|
|
Irving
Member
Member # 535
|
posted 06. March 2006 07:27
quote:
Doesn't your definition of information presuppose consciousness? There can be no information without an informer and an informed?
For the most part, Yes. Though not always...there's always mimicry. Recognize I'm taking this definition off of Doug's reference to "No Free Lunch."
IP: Logged
|
|
Doug Wedel
Member
Member # 1901
|
posted 06. March 2006 08:30
Lots and lots of ambiguities! Now consciousness rears its head -- if information is hard to define, and intelligence is even harder, then consciousness is impossible.
Anyway, let's try to make progress. You say the pulsar that emits pi is emitting information but the pulsar that emits a stream of digits that does not contain the digit "8" is not emitting information. Why not? How do you define "information" such that pi is it and a non-normal number doesn't count? Please give me your definition of information if you would.
IP: Logged
|
|
Doug Wedel
Member
Member # 1901
|
posted 06. March 2006 09:09
Rather than getting bogged down here, let me try to broaden the discussion a little.
I am assuming that posters and readers here know that David Wolpert, one of the authors of the No Free Lunch Theorems, has _strongly_ objected to the use which Dembski makes of these theorems. To see Wolpert's argument, entitled "William Dembski's Treatment of the No Free Lunch Theorems is Written in Jello," go to
http://www.talkreason.org/articles/jello.cfm
Wolpert specifically accuses Dembski of trying to perform a "category-change trick" but Wolpert says that Dembski fails to
quote:
turn this trick because despite his [Dembski] invoking the NFL theorems, his arguments are fatally informal and imprecise.
I am trying to remove the imprecision and ambiguity from Dembski's extrapolations before attempting my proof, and I'm hoping that people here will be able to help in this task.
IP: Logged
|
|
William A. Dembski
Member
Member # 7
|
posted 06. March 2006 09:36
Doug: The information we're talking about is complex specified information or specified complexity. The texts you need to be focusing on are chapters 3 and 4 in my book NO FREE LUNCH and then my articles "Searching Large Spaces" and "Specification: The Pattern That Signifies Intelligence" (both 2005 at www. designinference.com). The latter two papers respond to Wolpert's jello objection.
--WmAD
IP: Logged
|
|
Doug Wedel
Member
Member # 1901
|
posted 06. March 2006 10:43
Thanks for the references. I have a read a great deal of your work including some of the work you cited. I find much of it extremely lucid and compelling, particularly from the perspective of remininding us of what we don't know! I will review the pieces you reference that I am _not_ familiar with before continuing this post, but in the meantime, since the work you cite is from 2005, I wonder if you have seen/responded to Wolpert and Macready latest's publication on this theme published in the Dec '05 issue of IEEE Transactions on Evolutionary Computation under the title "Coevolutionary Free Lunches"? This is particularly apposite to the proof I propose to unfurl ;-)
IP: Logged
|
|
Micah Sparacio
Member
Member # 6
|
posted 06. March 2006 10:48
Doug,
(If early indicators are correct)-> If your intention is to compel readers who tend to side with Dembski, via proof, that Eigen is correct, then I don't think that your project is going to accomplish much at all.
Quite simply, if what you say is true:
quote:
Then I will demonstrate how the problem can be solved algorithmically through random number generation combined with artificial selection.
Then you will have completely failed to address the concerns that Dembski, et al have raised regarding the origin of complex specified information via non-intelligent sources.
To put things simply: your algorithm will have to employ artificial selection. Dembski holds something like the following: that algorithms might be used as tools for generating complex specified information, but only in virtue of the intelligence that shapes the algorithm.
[ 06. March 2006, 10:54: Message edited by: Micah Sparacio ]
IP: Logged
|
|
IF
Member
Member # 1904
|
posted 06. March 2006 11:27
Irving,
Mimicry with regards to consciousness (i.e. awareness) or with regards to information, or both?
Doug,
Algorithms as in computers? If so then hopefully you are aware of:
Jon Conway's "Game Of Life" http://www.bitstorm.org/gameoflife/
Enhanced, advanced and elaborated By Stephen Wolfram http://www.wolframscience.com/
IP: Logged
|
|
Doug Wedel
Member
Member # 1901
|
posted 06. March 2006 11:45
quote:
(If early indicators are correct)-> If your intention is to compel readers who tend to side with Dembski, via proof, that Eigen is correct, then I don't think that your project is going to accomplish much at all.
I certainly am not trying to 'compel' anyone to do or think anything. Also I am fully aware that many here will be inclined by disposition to reject my conclusions. If you don't find my logic persuasive, please feel free to disregard it ;-)
quote:
To put things simply: your algorithm will have to employ artificial selection. Dembski holds something like the following: that algorithms might be used as tools for generating complex specified information, but only in virtue of the intelligence that shapes the algorithm.
I understand. I hope to specifically disprove this. I am going to reveal every single axiom and assumption that goes into the little problem I am going to propose. I will reveal the algorithm's pseudocode. Nothing will be hidden. Then BEFORE I demonstrate the solution, we will see if the human beings in here, possessed of consciousness and advanced intelligence, can perceive any hidden "intelligence" lurking in my problem that somehow prefigures the solution to the problem.
quote:
Algorithms as in computers? If so then hopefully you are aware of:
Jon Conway's "Game Of Life" http://www.bitstorm.org/gameoflife/
Enhanced, advanced and elaborated By Stephen Wolfram http://www.wolframscience.com/
I am obsessed with cellular automatas in all their forms, and have studied these two authors and many other students of the subject from Ulam to Fogel.
IP: Logged
|
|
Printer-friendly view of this topic