|
Author
|
Topic: Granville Sewell and the Second Law of Thermodynamics
|
2ndclass
Member
Member # 1979
|
posted 12. March 2007 12:22
As I thought, LE, you have avoided my very simple requests:
1) Show how my example on March 9 runs afoul of Shannon or algorithmic information theory.
2) Provide a reference that explains your notion of information.
Here's a yes or no question: Are you capable of fulfilling these requests? You can answer this with two or three keystrokes, but I suspect that even this is asking too much of you. We'll see if I'm right.
[ 12. March 2007, 12:56: Message edited by: 2ndclass ]
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 12. March 2007 13:06
I am not an expert in information theory. However, it appears to me that 2ndclass has asked LifeEngineer a valid question with regard to the relationship between Shannon information and random data.
quote:
When it comes to Shannon (which is the kind of information that you're talking about, see your post from March 8) or AIT, the capacity to generate random data is the same as the capacity to generate information.
Can you please tell me how my example from March 9 runs afoul of these, or any, texts on Shannon or algorithmic information theory?
So LifeEngineer, I would like to agree with you when you say:
quote:
The creation of information question as it relates to scientific analysis of evolution is not about random number generators.
But, unless you explain how random data is not Shannon information I remain reluctant.
2ndclass, your position is a little clearer on this. However, Shannon aside, can you really see no difference between generating a string of random numbers and generating the information in a DNA molecule?
One more small question while I have your attention, I think I remember way back when I took physics that nobody, try as they might, has ever observed a closed physical system (thermodynamically isolated). Do I remember correctly (or is someone making a lot of money off of one that I am not aware of) and if so then why would anyone assume there would be such a thing as a closed information system?
-Mel
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 12. March 2007 13:42
Melvin,
Information is not a universal concept but rather a concept defined by those using it. If, which seems highly doubtful, shannon information theory defines random numbers of information generated by a closed system then the definition is useless with respect to the study of intelligence or evolution. I am not aware of any type of math that views the generation of data, random or otherwise, as equating to the generation of information. I answer to 2Cs question, whether random numbers are or are not considered new information by shannon information theory then that definition of information is irrelevant to the discussion here.
I can't imagine a true real world closed system, but the concept apparently is useful in abstract analysis. It is useful, in the abstract, to create or imagine systems that are not influenced by millions of outside variables. In information processing, it is useful to think about or visualize systems that are only influenced by defined input and by processing algorithms. These would be idealized computers or logic machine.
To get back to the issue actually being discussed, it is generally recognized that in a close deterministic system (logic machine) with defined input and defined processing algorithms the output is determined or predicted by the input and the processing algorithm. The real question or real issue here is whether we can perform scientific analysis of intelligent behavior, including evolution, using this closed system deterministic model, or whether we need to assume that there is some type of closed system that produces or creates information.
I am saying we can perform scientific analysis of intelligent processes without hypothesizing any closed system information generators.
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 12. March 2007 16:14
Melvin: quote:
However, Shannon aside, can you really see no difference between generating a string of random numbers and generating the information in a DNA molecule?
Melvin, you're right, there's definitely a difference, and it can be expressed in terms of algorithmic information theory.
Presumably (and I'm certainly open to correction here since I'm not a biologist) the set of DNA sequences that yield viable organisms is a comparatively small subset of all possible sequences. If we imagine a Universal Turing Machine program that tests all possible sequences and outputs only those that are viable, then each viable sequence is uniquely describable as "the nth output of UTM U". Since the length of the sequences grows faster than the bit length of our description (which grows only logarithmically with n), these sequences are compressible.
Deterministic or partially deterministic processes are guaranteed to generate compressible output, while completely random processes are virtually guaranteed to generate incompressible output. So we can safely conclude that DNA sequences of viable organisms are not the result of purely random processes.
So compressibility distinguishes DNA from random sequences, but what about from boring sequences like 111111111111111....? These sequences are even more compressible than DNA. This is where the idea of complexity as the crossover point between order (simple & boring) and chaos (random & boring) comes into play. Biocomplexity is at a happy medium between the two boring extremes, which is where interesting properties emerge. The Santa Fe Institute has done some fascinating research in this area.
quote:
One more small question while I have your attention, I think I remember way back when I took physics that nobody, try as they might, has ever observed a closed physical system (thermodynamically isolated). Do I remember correctly (or is someone making a lot of money off of one that I am not aware of) and if so then why would anyone assume there would be such a thing as a closed information system?
I imagine that constructing a 100% closed system is beyond our capability in practice, but, as LE said, we can model some real-world systems as closed if the inputs and outputs are negligible compared to what's going on inside.
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 13. March 2007 08:10
Apparently 2C is back on the same page with the rest of us. If you have a system that is closed to all inputs other than 1) defined input and 2) defined processing algorithms, then, as far as is known, such a system is incapable of creating new, useful meaningful information (information other than information that can be explained, determined or predicted by 1 and 2). It is recognized that such a system might produce noise that reduces the information content of output, but such a system can not increase information.
I am not interested in any silly word games challenging the above statement or conclusion. I am not enough of an expert in abstract math to know if this conclusion is provable in either the real world or in the world of abstract math, but it appears to be a reasonable conclusion or assumption. To my knowledge there is no hard scientific or mathematical evidence supporting any useful conclusion other than the above. Anyone wishing to use some assumption other than the above needs to provide evidence in support of their alternative position.
Does anyone disagree?
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 13. March 2007 10:50
2ndClass: quote:
Here's a yes or no question: Are you capable of fulfilling these requests? You can answer this with two or three keystrokes, but I suspect that even this is asking too much of you. We'll see if I'm right.
Looks like I was right. Not only is LE unable to provide a reference for his notion of information, he's also incapable of admitting that he can't provide one.
quote:
If you have a system that is closed to all inputs other than 1) defined input and 2) defined processing algorithms, then, as far as is known, such a system is incapable of creating new, useful meaningful information (information other than information that can be explained, determined or predicted by 1 and 2).
Earlier you were talking about Shannon information, which has nothing to do with utility or meaning. By adding the "useful" and "meaningful" qualifiers, you're raising the goalposts kinda late in the game, don't you think?
And what do they mean? Useful to whom? Meaningful to whom? So much for your "hard science".
quote:
Does anyone disagree?
I suspect that if you ever provide operational definitions of useful and meaningful, it will be trivial to show that such information can emerge from a closed system. But I can't say for sure, since you haven't provide such definitions.
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 14. March 2007 09:44
It is rather amusing to watch the efforts of people like 2C to disrupt attempts at serious discussion of ID related issues. The ability of closed systems, or systems with limited inputs is a critical issue in the formulation of both AI and evolutionary theories.
There are actually two key issues. First, is there any evidence that real world systems with constricted input can generate information? Second, is there any evidence that abstract mathematical systems with limited inputs can create information?
We know or recognize that apparently 'closed' real world systems can create noise which reduces information. We also know that closed systems can create various types of data (including random numbers), but such data has not been shown to represent new information.
If there is no evidence for either real or mathematical creation of information, the next logical question is "Can scientific analysis of intelligent behavior be performed without assuming the ability of closed systems to create information?" The closely associated question is "Why do posters like 2C who clearly have nothing to contribute to such a discussion insist on attempting to disrupt serious discussion of the issue?"
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 14. March 2007 11:31
This is your idea of a "serious" discussion? Information is a mathematical concept, and serious math discussions involve doing the math, or at least pointing to sources where the math is done, neither of which I have ever seen you do. It seems that such discussions are way beyond your abilities, LE.
I suspect that you're a bot. And a poorly designed one, as you're incapable of responding to specific questions, and can do nothing more than endlessly repeat vague and unsupported assertions.
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 14. March 2007 13:38
It seems all agree there exists a happy medium between simple [high compressibility and low complexity] and chaos [low compressibility and high complexity]. The following is a serious question from a non-expert. Is this information – happy medium – always useful and or meaningful?
-Mel
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 14. March 2007 14:09
Melvin, again, we would have to pin down operational definitions of useful and meaningful. From what I've seen, interesting properties need not, and usually do not, emerge from "edge of chaos" systems. Likewise, most mathematical recursions don't result in something as fascinating as the Mandelbrot set. Interesting emergent phenomena are the exception, not the rule, but they're not too rare to be studied.
IP: Logged
|
|
John A. Davison
Member
Member # 1425
|
posted 14. March 2007 14:25
Sorry, self-named 2ndclass, whoever that is, but information is NOT a mathematical concept and to attempt to define it that way is both scandalous and cowardly. This message to you is information and it is NOT a mathematical concept. Who do you think you are kidding?
"A past evolution is undeniable, a present evolution undemonstrable."
John A. Davison
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 14. March 2007 15:05
quote:
Sorry, self-named 2ndclass, whoever that is, but information is NOT a mathematical concept and to attempt to define it that way is both scandalous and cowardly.
Well, then, everyone who teaches information theory is a scandalous coward, since it is always defined mathematically. The first sentence in the wikipedia entry for information theory is, "Information theory is a discipline in applied mathematics...". Wolfram's MathWorld defines information theory as branch of mathematics. Shannon, the father of information theory, titled his groundbreaking paper, "A Mathematical Theory of Communication".
Information theory is often taught in engineering school, but that's also true for other mathematical concepts like phasors, Laplace transforms, and Fourier transforms.
IP: Logged
|
|
John A. Davison
Member
Member # 1425
|
posted 14. March 2007 15:14
Perhaps everyone who teaches information simply as a mathematical concept IS a scandalous coward. I wouldn't know. I know only that those who hide their identities are cowards at least in my book.
"A past evolution is undeniable, a present evolution undemonstrable"
John A. Davison
The above is information.
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 14. March 2007 15:58
John: quote:
I am sorry I invaded this thread. I should have known better. It won't happen again.
Don't tease us, John. Some readers might have actually believed you, only to be disappointed when you reappeared.
quote:
Perhaps everyone who teaches information simply as a mathematical concept IS a scandalous coward. I wouldn't know.
Of course you wouldn't, but as usual, that didn't stop you from pontificating as if you DID know.
quote:
I know only that those who hide their identities are cowards at least in my book.
Guilty as charged. But don't worry, I'm sure that cowards such as Martin and myself will reveal our identities as soon as they become pertinent to the issues being discussed.
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 14. March 2007 19:00
Let the record show that I do not regard 2ndclass as a scandalous coward even though I am a long way from accepting information as a mathematical concept and I am a little baffled by the internet custom of screen names. I find 2ndclass to be honest and logically engaging.
We can’t ignore the fact that information has been studied by a great many minds within the framework of mathematical construct. But 2ndclass would agree I hope that these important studies don’t make information a mathematical concept any more than fractal models of clouds make those physical structures mathematical concepts.
I am trying to pin down an operational definition of the terms “useful” and “meaningful”. 2ndclass, would you agree that the information stored in DNA is useful and meaningful? If so, what makes it useful and why is it meaningful? And, it is useful and meaningful to whom or what? Is the mRNA a well designed “bot”, as you might call it, programmed to respond to specific stimuli or is it more like a pin ball slammed about by forces it can’t recognize nor respond to while endlessly repeating vague parabolic curves?
-Mel
IP: Logged
|
|
|
Printer-friendly view of this topic
