Uncommon Descent Serving The Intelligent Design Community

Human Evil, Music, Logic, and Himalayan Dung Heaps

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

When I was in college I studied classical piano with Istvan Nadas who was a Hungarian concert pianist and a student of Bela Bartok. Istvan was a miraculous survivor of one of Hitler’s death camps. The stories he told me still haunt me to this day.

The commandant of the death camp liked to play Bach over the loudspeaker system while he had random inmates shot or hung, just for fun and entertainment. Nadas told me about the horror of listening to Bach while he watched his fellow inmates being machine-gunned to death in front of him. Nadas told me, “I knew every note of that music and could play it on the piano, but I also knew that if they discovered I was a concert pianist they would break all my fingers so I could never play the piano again.”

Nadas’s death camp was eventually “liberated” by the Russians. Istvan was one of only 150 survivors from a camp of thousands. He weighed 90 pounds and was suffering from dysentery and other diseases. While the Russians were transporting him on a train to what he knew would be a Russian internment camp he managed to jump out of the train as it slowed in the mountains. Under machine-gun fire he fled into the trees, was helped my local residents, and was eventually smuggled by an African American GI under a tarp in the back of a jeep through Check Point Charlie.

Nadas eventually discovered that every member of his extended family had either been gassed or otherwise tortured and exterminated by the Nazis, or shot by the Russians, with one exception: his mother, whom he eventually tracked down in Italy after the war.

One evening, after a concert at the university while I was studying piano with Nadas, which was conducted by a guest “contemporary composer” — it was just a bunch of random cacophony, very painful to listen to, but sold as legitimate music — I asked Nadas what he thought.

“It is a Himalayan dung heap,” he replied. (Nadas spoke six languages fluently, and had a way with words.) This phrase stuck in my mind, and it’s the perfect description of something so obviously stupid that it represents a pile of crap of Himalayan proportions.

The students and faculty applauded the Himalayan-dung-heap “music” because no one had the courage to point out the obvious, except for Nadas.

This is a perfect metaphor for Darwinism. Very few people in academia have the courage to point out the cacophony and illogic of Darwinian speculation.

It takes the courage of someone like Nadas, who was willing to jump off a train in the mountains under machine-gun fire, to tell the obvious truth.

Comments
Look I realises this is frustrating for you, UPD, but you must understand that the frustration is mutual! Your 1) and 2) are NOT operational definitions!!!!!!
So what? You're off in la la land, the land of virtual reality, and UPB is asking you to come back to reality and saying that if you're going to develop an operational definition develop one that reflects reality. As your own sources point out, reality is the place to start!
Background for research consists of everything a researcher knows about a topic: 1) How well grounded the question is in the current knowledge base (the problem must have a basis in theory, research, or practice (we need to know what is already known so that we can judge how much it can add to the knowledge base; gives us an anchor)
What do we already know about information and where did we obtain this knowledge? Why do you, Elizabeth, believe there is information in the genome? What do we already know about how information is communicated and where did we obtain this knowledge? How does the information in the genome get communicated?Mung
July 26, 2011
July
07
Jul
26
26
2011
12:08 PM
12
12
08
PM
PDT
I’m sure it sounds crazy, but one might expect that at some point your response would include some of the language used in the description of the reality being operationalized...
Elizabeth is not attempting to operationalize anything that exists in reality. This would be hilarious if it weren't so utterly pathetic. First Lizzie complains that the concepts are not abstract though, they need to be more abstract so they can be applied to other instances. Then she argues that the concepts are too abstract to serve as an operational definition, that operational definitions cannot be general, they must be specific to one single instance of a measurement. We cannot develop an operational definition of a peanut butter sandwich that can identify multiple instances of a peanut butter sandwich. The operational definition must be specific to a particular peanut butter sandwich. So if we have wheat bread, instead of white bread, that would require a different operational definition for a peanut butter sandwich. Totally bizarre. So let's take a closer look at Elizabeth's own sources:
Operationalization occurs when we take a hypothesis, e.g. aggression causes further aggression, and develop a procedure, or operation, for identifying instances of the critical terms, here, aggression. http://www.newfoundations.com/EGR/Oper.html
Instances, is plural. We should be able to identify aggression, for example, in not just one subject, and not just in one scenario, but in multiple instances.
Operationalization occurs when we take a hypothesis, e.g. violence causes further violence, and develop a procedure, or operation, for identifying instances of the critical terms, here, violence. http://web.jjay.cuny.edu/~pzapf/classes/CRJ70000/Formulating%20the%20research%20question.htm
Now this is too funny. It looks like one of Lizzie's sources has plagiarized the other! Or perhaps they both plagiarized a common source. Design detection in action! Instances, is plural. We should be able to identify violence, for example, in not just one subject, and not just in one scenario, but in multiple instances. So Upright BiPed's point is completely valid, as is mine. Elizabeth, could you please develop an operational definition that is based on reality?Mung
July 26, 2011
July
07
Jul
26
26
2011
11:54 AM
11
11
54
AM
PDT
Mung:
Elizabeth, what is the purpose of having an operational definition?
The purpose of an operational definition is to allow a hypothesis to be tested in a manner that objective observers can agree on whether or not the hypothesis is supported by the results.
How broadly applicable should an operational definition be?
It is usually very narrow. Generalisability is the price of precision.
Can your operational definition tell us whether there is information present on, say, a computer hard disk?
No. It would have to be operationalised for that hypothesis. The reason it is called an "operational definition" is that it is specific to an operation within a hypothesis testing protocol.
Can it detect whether there is information in the genome of a cell?
Only if operationalised within a hypothesis aimed to demonstrate this.
Or is your operational definition applicable only to your own little virtual world?
The operational definition is applicable only to the hypothesis-testing operation for which it is defined. However, all operational definitions should be consistent with the conceptual definition at issue. So if, for example, we take the Merriam-Webster definition of information, that can be operationalised to any specific instantiation. And because the Merriam-Webster definition is so clear, it would be pretty easy to operationalise for any given hypothesis: Within a genome: Does the genome contain arrangements of nucleotides that correlate with any measures of cell behaviour? Within a computer: Does the computer contain binary strings that correlate with the output of the computer? Or something. But for present purposes, the operationalisation has to be a bit more complicated, because my entities are not a given (I start with no DNA, no computer, no cells), and therefore more specific (or the thing would be explosively complicated). Also, UPD wants this bit about intermediary objects in there. Which is fair enough, I guess.Elizabeth Liddle
July 26, 2011
July
07
Jul
26
26
2011
02:30 AM
2
02
30
AM
PDT
Upright BiPed:
Dr Liddle, In response to the question I asked, your comment was a rather tragic disappointment. I asked for concrete examples of what variables you intended to measure, and you gave me sidesteps in return. I’m sure it sounds crazy, but one might expect that at some point your response would include some of the language used in the description of the reality being operationalized – but no, apparently that won’t be useful. I think back on my career and I try to remember a time when the critical issues described in discovery, are left absent from the study. Research cost money, so we typically don’t do business like that. Perhaps that’s just another one of those British/American thingies you were admiring earlier.
No Upright BiPed, I did not give you sidesteps, and I told you exactly what I intended to measure, and how I intended to measure them. I'll tell you again, and try to make it even clearer: Here is the conceptual definition of information I was using (from Merriam-Webster):
Information: the attribute inherent in and communicated by one of two or more alternative sequences or arrangements of something that produce specific effects
My "arrangments of something" will be strings of "virtual polymers". I will categorise them by their arrangements, as they will be categorical variables. I will measure their "specific effects" either as longevity (in terms of iterations) or reproductive fidelity (by comparison to "organisms" with different parentage. If the arrangement is a statistically significant predictor of "specific [functional] effects" I will consider my claim demonstrated.
If I had only known. After weeks of describing the various objects/dynamics involved in the existence of information, my conversation partner has chosen to rollover and play dead. Forever gone are those critical details which separate the existence of information from the formation of rust. I would have saved both of us the time, and just given a definition that ignores the details.
No, I have not UPD, as I think must be pretty clear to anyone reading this conversation.
Let’s look at your pitch:
Conceptual hypothesis: That chance and law alone can give rise to information, where “information” is defined conceptually, according to Merriam-Webster, as “the attribute inherent in and communicated by one of two or more alternative sequences or arrangements of something that produce specific effects.”
There is a problem with this.
Right. So you do have a problem with Merriam-Webster. OK. Well, that's a shame because it seemed to me to capture the essence of your definition, without the problematic terms.
It doesn’t cover anything of our observations; which is generally acknowledged as the “first mistake you made” in any study where you do it. The required physical objects, and their dynamic relationships to one another, are simply ignored. I know you’re a stickler for precision Dr Liddle, but how does one object “communicate” with another object?
And it's because I'm a stickler for precision, UPD that I needed a definition that did not use such terms. That is the beauty of the Merriam-Webster definition - it requires that an "arrangement of something" - which may be objects - "produces specific effects". We have agreed, I think, that if it was merely a physical object that caused specific effects (hydrogen burning in oxygen, for instance) we would not call that "information". What makes it information, I think we agreed, is when it is not the objects themselves,but a arrangement of those objects that produces the specific effects. As, for example, when a codon produces an amino acid, not because the nucleotides themselvesproduce this effect, but because the arrangement of the nucleotides produce this effect. Right?
Operational hypothesis: That, starting only with non-self-replicating entities with a physics-and-chemistry plus random kinetics, self-replicating “virtual organisms” can emerge that contain patterns of “virtual matter” whose arrangement determines the fidelity of its self replication (measured in terms of similarity to its “parent” as compared with a randomly substituted pattern).
There is a problem here as well. This supposed operationalization is based upon the definition above which ignores the required physical objects, and their dynamic relationships to one another. It also includes a mis-statement of what the final output must be.
No, it doesn't. I've actually specified "virtual matter" (we agree we are talking about virtual reality here, right, not an actual wet lab experiment?), and I've specified the dynamic relationship - the arrangement of the "virtual matter" must determine, at a future time (no backward causality!), i.e. dynamically, specific effects, namely either longevity or fidelity.
Here, let’s try this. Let us spill the last vesicles of illusion.
Starting only with non-self-replicating entities in a physics-and-chemistry (plus random kinetics) environment, self-replicating “virtual organisms” can emerge that contain dissociated representations embedded in the arrangement(s) of matter. These arrangements represent the system that created them, and will determine the output of that system by means of an intermediary “virtual object”. Without becoming incorporated, this object may interact with either the representation or output, or both, but where the two remain whole and physically separated. 1) Dissociated= having no physical relationship to that which it represents 2) Intermediary= serves the dynamic purpose of allowing the input representations to determine the output while they each remain discrete, a facilitator
Aaarrrghhhh! Look I realises this is frustrating for you, UPD, but you must understand that the frustration is mutual! Your 1) and 2) are NOT operational definitions!!!!!! "Representations" is NOT operationalised! Which seems really odd, because the Merriam-Webster definition manages to convey the essence of meaning without even using the word "representation", which would be circular (you can't just provide a near-synonym in a dictionary definition). Look UPD: forget (if you can bear it) for a moment your own conceptual definition for a moment and look at the Merriam-Webster one. It seems to me to encapsulate yours, but is considerably simpler, and does not use terms that themselves depend on the thing being defined (It's pretty hard to define "representation" or "meaning" without referencing "information"). It's a good dictionary. That dictionary definition, which Meyer, for example, approves, and so do I, captures the both immaterial nature of information ("arrangement of something" not "something"), the causal dynamic ("produces ...effects"), and the content ("specific effects"). I've made it even more rigorous by specifying "specific functional effects". It seems to fit your concept, without the burden of introducing potentially circular terms. It only lacks your requirement of a "virtual intermediary". Which I will try to add.
Which one of these observed requirements would you like to have removed? If there are none, then I wish you the very best of luck. If you succeed, things are likely to start moving real fast, so remember to mention me in your acceptance speech.
None, but they need clarification. Let me try:
Starting only with non-self-replicating entities in a physics-and-chemistry (plus random kinetics) environment, self-replicating “virtual organisms” can emerge that contain dissociated representations embedded in the arrangement(s) of virtual matter represented as strings that These arrangements represent the system that created them cause the virtual organism to self-replicate with fidelity, and willthus determine the output of that system namely a copy of that system.
So far I don't think I have done too much damage to yours, and we aren't a million miles from Merriam-Webster. However, you include an additional requirement:
The arrangment must produce its output by means of an intermediary “virtual object”. Without becoming incorporated, this object may interact with either the representation or output, or both, but where the two remain whole and physically separated.. This "virtual object" must take the form of a second arrangement of "virtual matter" that may interact with the strings and with some other "virtual object" that affect the fidelity of the self-replication of the "virtual organisms" without either permanently altering, or being altered by, the interaction.
I'm less confident I can pull off this second requirement, but I'm happy to give it a go. ********************************************************** What about the rest of your long post to me, to which I responded in detail? Do you have any response to mine? Because I spent quite a bit of time thinking about it, and wondering what you meant. Did you find my responses satisfactory? In which case do you agree that categorical variables are variables? It seems important, because my demonstration will involve categorical variables (the arrangements in my strings).Elizabeth Liddle
July 26, 2011
July
07
Jul
26
26
2011
02:09 AM
2
02
09
AM
PDT
The Digital Code of DNA http://www.nature.com/nature/journal/v421/n6921/full/nature01410.htmljunkdnaforlife
July 26, 2011
July
07
Jul
26
26
2011
01:40 AM
1
01
40
AM
PDT
Elizabeth, what is the purpose of having an operational definition? How broadly applicable should an operational definition be? Can your operational definition tell us whether there is information present on, say, a computer hard disk? Can it detect whether there is information in the genome of a cell? Or is your operational definition applicable only to your own little virtual world?Mung
July 26, 2011
July
07
Jul
26
26
2011
01:35 AM
1
01
35
AM
PDT
I know you’re a stickler for precision Dr Liddle, but how does one object “communicate” with another object?
I know! I know!Mung
July 26, 2011
July
07
Jul
26
26
2011
01:03 AM
1
01
03
AM
PDT
kf:
Dr Liddle: Pardon. Digital simply means discrete as opposed to continuous state. The alphabet is digital. GEM of TKI
And if that is all that is meant by it, I have no problem with it. But then any molecule is digital by that definition! Atoms are discrete. But I take your point.Elizabeth Liddle
July 26, 2011
July
07
Jul
26
26
2011
12:45 AM
12
12
45
AM
PDT
Dr Liddle, In response to the question I asked, your comment was a rather tragic disappointment. I asked for concrete examples of what variables you intended to measure, and you gave me sidesteps in return. I’m sure it sounds crazy, but one might expect that at some point your response would include some of the language used in the description of the reality being operationalized - but no, apparently that won’t be useful. I think back on my career and I try to remember a time when the critical issues described in discovery, are left absent from the study. Research cost money, so we typically don’t do business like that. Perhaps that’s just another one of those British/American thingies you were admiring earlier. If I had only known. After weeks of describing the various objects/dynamics involved in the existence of information, my conversation partner has chosen to rollover and play dead. Forever gone are those critical details which separate the existence of information from the formation of rust. I would have saved both of us the time, and just given a definition that ignores the details. Let’s look at your pitch:
Conceptual hypothesis: That chance and law alone can give rise to information, where “information” is defined conceptually, according to Merriam-Webster, as “the attribute inherent in and communicated by one of two or more alternative sequences or arrangements of something that produce specific effects.”
There is a problem with this. It doesn’t cover anything of our observations; which is generally acknowledged as the “first mistake you made” in any study where you do it. The required physical objects, and their dynamic relationships to one another, are simply ignored. I know you're a stickler for precision Dr Liddle, but how does one object “communicate” with another object?
Operational hypothesis: That, starting only with non-self-replicating entities with a physics-and-chemistry plus random kinetics, self-replicating “virtual organisms” can emerge that contain patterns of “virtual matter” whose arrangement determines the fidelity of its self replication (measured in terms of similarity to its “parent” as compared with a randomly substituted pattern).
There is a problem here as well. This supposed operationalization is based upon the definition above which ignores the required physical objects, and their dynamic relationships to one another. It also includes a mis-statement of what the final output must be. Here, let’s try this. Let us spill the last vesicles of illusion.
Starting only with non-self-replicating entities in a physics-and-chemistry (plus random kinetics) environment, self-replicating “virtual organisms” can emerge that contain dissociated representations embedded in the arrangement(s) of matter. These arrangements represent the system that created them, and will determine the output of that system by means of an intermediary “virtual object”. Without becoming incorporated, this object may interact with either the representation or output, or both, but where the two remain whole and physically separated. 1) Dissociated= having no physical relationship to that which it represents 2) Intermediary= serves the dynamic purpose of allowing the input representations to determine the output while they each remain discrete, a facilitator
Which one of these observed requirements would you like to have removed? If there are none, then I wish you the very best of luck. If you succeed, things are likely to start moving real fast, so remember to mention me in your acceptance speech.Upright BiPed
July 26, 2011
July
07
Jul
26
26
2011
12:22 AM
12
12
22
AM
PDT
Actually, Ilion, I'm not tossing any. I'm asking for them. And I'm happy with Merriam Webster, as long as you all are.Elizabeth Liddle
July 26, 2011
July
07
Jul
26
26
2011
12:15 AM
12
12
15
AM
PDT
Mung: "How many different definitions of information do you plan to toss our way?" As many as it takes until one of them to stick to the wall? As many as it takes until you to decide that the "argument" isn't worth your time?Ilion
July 25, 2011
July
07
Jul
25
25
2011
05:06 PM
5
05
06
PM
PDT
Mung:
Hi Lizzie, ok, so let’s see if we can clarify a couple points. Please read the entire post and think on it before responding. You, personally, believe that Chance + Necessity can create meaningless information. Is that correct?
Yes. Using Shannon's definition, which allows us something called "information" which can be meaningless.
Do you also think believe that Chance + Necessity can create meaningful information?
Yes. And that is what I intend to try to demonstrate.
I don’t understand the following:
…no ID proponent is proposing that Chance and Necessity can’t create meaningless information.
Is that what you meant to say, or did you mean to say something different? Because it really makes no sense, to me. I know that I personally deny that Chance + Necessity can create meaningless information. Perhaps you meant to say that no ID proponent accepts that information can be meaningless? [But is that true?]
You can put it either way. If information is defined in such a way as to include meaningless information (as in Shannon Entropy) then I would expect that ID proponents would agree that Chance and Necessity could produce it. Meyer and Dembski, for instance, quantify information in Shannon terms, and agree that it can be produced by Chance (and Necessity, sometimes, for Dembski). What they dispute is that Specified Information can be produced by Chance and Necessity. Other ID proponents, I now understand, more sensibly IMO, contend that meaningless information is an oxymoron, and take no stand on whether it can be created by Chance and Necessity alone or not. What they (in the person of Upright BiPed, inasmuch I understand him) contends is that meaningful information cannot be created by Chance and Necessity. It is that claim that I hope to falsify.
Or, stated another way, ID proponents believe information must have meaning. [Again, questionable.]
Well, I don't suppose either of us can speak for all ID proponents.
Or that no ID proponent accepts that Chance can generate information. [iirc, Debmski wrote that it can.]
Yes, but not Complex Specified Information, and AFAICT, he defines plain ole information in Shannon terms.
Or finally, that ID proponents themselves are confused about information and some believe that there are different kinds of information, including information that can have no meaning and produce no effect, but that’s not the kind of information that needs to be demonstrated in order to falsify ID.
Well, they need not be confused. Each one might be perfectly clear. But they do appear to differ. As I said, I regard this as healthy.
lol. [Laughing at the situation, not you.]
Well, that's something :)
Look, In some respects I don’t blame you. There is a lack of clarity in some authors and that can be confusing. But what I do doubt is that in your conversations with Upright BiPed, myself, and kf that any of us has engaged in that sort of lack of clarity. I think we’ve been clear and consistent from the beginning of this subject.
I'm not actually accusing any of you of lack of clarity, except in the sense that the definitions offered have not proved easy to operationalise. Which I guess does mean that things weren't clear. Certainly the essential thing in an operational definition is to remove ambiguity. That, as those links I gave pointed out, often means that the operational definition lacks generalisability - it's the price one pays for lack of ambiguity. But it buys something very worthwhile IMO, which is the ability to actually subject claims to rigorous testing. If I succeed, I will not have demonstrated that life wasn't designed (lack of generalisability). I will, however, have demonstrated that a fairly fundamental ID claim is not always true.
I once again think back to your coin tossing example, way back when, when Upright BiPed asked you what that 100 bits of Shannon information that you thought you had generated by tossing a coin was about. So what I think you are saying is that we are requesting that in your demonstration any information you generate must have meaning or be meaningful. And so you are attempting to come up with a definition of information that incorporates the concept of meaning.
Yes.
How am I doing so far?
A perfect 10. Well, 9.99.
Shannon information is not an appropriate measure of information for this purpose because it doesn’t incorporate the concept of meaning
Now assuming we are on the same page so far, and what a minor miracle that would be, how is it that you think that Shannon information cannot measure the information in a message that has meaning?
I think it cannot distinguish between meaningful information and non-meaningful information. Which is the point, essentially, made by both Dembski and Meyer, who do use Shannon information, but have to add this extra criterion of Specification. I think Upright BiPed's approach is better, although it wasn't the approach I had in mind when I made my original claim. But then my original claim was intended to cover any definition of information (unless it's a Humpty Dumpty one).
Surely that cannot be true. And yet you seem to think that Shannon information can only measure information in a message in which there is no meaningful information. How odd.
No. See above. Use your lawyer skills. Or just think in terms of Specificity and Sensitivity (in the Signal Detection Theory sense). The Shannon criterion is has Sensitivity (few false negatives) but not Specificity (lots of false positives). What I hope we can do using the Webster Merriam definition is get a definition that has both Sensitivity (no false negatives, which is obviously of importance to me as I want to demonstrate my claim) and Sensitivity (no false positives, which is obviously of importance to you, as you want to make sure my claim is sound).
Just because Shannon information does not concern itself with the meaning in a message, how is it thereby invalidated as a measure?
See above. More specifically (pun un-intended but apt) we need criteria that allow us to say that meaning is present, i.e. specificity for meaningful information. I think that Merriam-Webster covers that, but for good measure, I have added an additional criterion, which is that the "arrangement of [virtual] matter" has not only to have "specific effects" but "specific functional effects", which I operationalise as producing greater reproductive fidelity and/or greater longevity of my virtual critters (if they form in the first place, which they may not :))
If you in fact demonstrate that your messages have meaning, why can’t Shannon information measure that information?
It could. But it wouldn't be terribly relevant, because may claim is only that Chance and Necessity can create [meaningful] information,not a specific quantity of it. But sure, quantifying the information in the meaningful strings (if there were any) might a cool extra.
Shannon information is not an appropriate measure of information for this purpose because it doesn’t incorporate the concept of meaning
All this would indicate is that we need to have some other way to determine meaning. Correct?
Yes. Right now, I'd really like to go with Webster, especially as it has Meyer's imprimatur.
e.g., It [information] changes what someone believes to be the case. (How’s that for an operational definition!)
Yup :) But operational definitions are operation-specific, and that isn't the operation I intend to perform. However, as an epiphenomenon it would be very welcome :)
It does not mean we cannot use Shannon information to measure the information in the message. Right?
Right. It's just that we cannot use Shannon information to determine whether the information is meaningful. And if we can determine that my critters contain meaningful information, we won't need to quantify it because my job will already be done. But if you want to pay me per bit, I won't refuse. Cheers Mung. I needed a constructive conversation today :) Hope that was one. LizzieElizabeth Liddle
July 25, 2011
July
07
Jul
25
25
2011
03:49 PM
3
03
49
PM
PDT
KF, RE 213, Was it you that referenced a proposed functional target ratio of 1 in 10^74 (or something similar) for proteins in another thread? I can't find the reference.material.infantacy
July 25, 2011
July
07
Jul
25
25
2011
03:40 PM
3
03
40
PM
PDT
Dr Liddle: Pardon. Digital simply means discrete as opposed to continuous state. The alphabet is digital. GEM of TKIkairosfocus
July 25, 2011
July
07
Jul
25
25
2011
03:21 PM
3
03
21
PM
PDT
Hi Lizzie, ok, so let's see if we can clarify a couple points. Please read the entire post and think on it before responding. You, personally, believe that Chance + Necessity can create meaningless information. Is that correct? Do you also think believe that Chance + Necessity can create meaningful information? I don't understand the following:
...no ID proponent is proposing that Chance and Necessity can’t create meaningless information.
Is that what you meant to say, or did you mean to say something different? Because it really makes no sense, to me. I know that I personally deny that Chance + Necessity can create meaningless information. Perhaps you meant to say that no ID proponent accepts that information can be meaningless? [But is that true?] Or, stated another way, ID proponents believe information must have meaning. [Again, questionable.] Or that no ID proponent accepts that Chance can generate information. [iirc, Debmski wrote that it can.] Or finally, that ID proponents themselves are confused about information and some believe that there are different kinds of information, including information that can have no meaning and produce no effect, but that's not the kind of information that needs to be demonstrated in order to falsify ID. lol. [Laughing at the situation, not you.] Look, In some respects I don't blame you. There is a lack of clarity in some authors and that can be confusing. But what I do doubt is that in your conversations with Upright BiPed, myself, and kf that any of us has engaged in that sort of lack of clarity. I think we've been clear and consistent from the beginning of this subject. I once again think back to your coin tossing example, way back when, when Upright BiPed asked you what that 100 bits of Shannon information that you thought you had generated by tossing a coin was about. So what I think you are saying is that we are requesting that in your demonstration any information you generate must have meaning or be meaningful. And so you are attempting to come up with a definition of information that incorporates the concept of meaning. How am I doing so far?
Shannon information is not an appropriate measure of information for this purpose because it doesn’t incorporate the concept of meaning
Now assuming we are on the same page so far, and what a minor miracle that would be, how is it that you think that Shannon information cannot measure the information in a message that has meaning? Surely that cannot be true. And yet you seem to think that Shannon information can only measure information in a message in which there is no meaningful information. How odd. Just because Shannon information does not concern itself with the meaning in a message, how is it thereby invalidated as a measure? If you in fact demonstrate that your messages have meaning, why can't Shannon information measure that information?
Shannon information is not an appropriate measure of information for this purpose because it doesn’t incorporate the concept of meaning
All this would indicate is that we need to have some other way to determine meaning. Correct? e.g., It [information] changes what someone believes to be the case. (How's that for an operational definition!) It does not mean we cannot use Shannon information to measure the information in the message. Right?Mung
July 25, 2011
July
07
Jul
25
25
2011
03:08 PM
3
03
08
PM
PDT
UPD: Thanks. I'm going off to bed shortly, but will check in tomorrow.Elizabeth Liddle
July 25, 2011
July
07
Jul
25
25
2011
02:48 PM
2
02
48
PM
PDT
Mung:
Elizabeth Liddle:
If we define information as an arrangement of matter that has specific effects…
How does tossing a coin generate information?
It doesn't, by that definition.
How many different definitions of information do you plan to toss our way?
I'm hoping we can agree on one.Elizabeth Liddle
July 25, 2011
July
07
Jul
25
25
2011
02:44 PM
2
02
44
PM
PDT
Dr Liddle at 214 I see your response now. I will return this evening to comment on it.Upright BiPed
July 25, 2011
July
07
Jul
25
25
2011
02:30 PM
2
02
30
PM
PDT
Elizabeth Liddle:
If we define information as an arrangement of matter that has specific effects...
How does tossing a coin generate information? How many different definitions of information do you plan to toss our way?Mung
July 25, 2011
July
07
Jul
25
25
2011
02:20 PM
2
02
20
PM
PDT
Mung #209
Mung: Please describe the message and the measurement and explain how the measurement demonstrates that there is 0 bits of information in the message. Elizabeth Liddle: OK, let’s take a channel that is dominated by noise at a frequency of 50 hertz. We listen to this channel. We fail to detect any departure from the 50 Hz tone. We therefore receive 0 bits of information. In one sentence the 50Hz is noise. In the next sentence the 50Hz is a tone.
Yes, noise can be a tone. It’s why we use notch filters.
So if all that is present is a 50Hz tone, it’s the noise that’s being dominated, not the tone. But where’s the message? Well, you say we cannot detect the message, because we have not detected any departure from the 50 Hz tone. Yet you appear to argue that from this line of reasoning we can conclude that we’ve measured the message and found no information in it. We therefore receive 0 bits of information. We haven’t received any message. So we haven’t measured a message that contains 0 bits of information. Try again, please.
A message with 0 information isn’t a message, is it? But no, I won’t try again. It is utterly irrelevant. Mung #210
I suppose that I just imagined that someone suggested that we could use Shannon information because it [Shannon information] can tell us whether information is present or not. Elizabeth Liddle @168: An operational definition can also give criteria for establishing whether some categorical variable is present or absent. As I said. One that allows you to measure the variable will of course also allow you to say whether it is present or absent. Shannon’s definition does that – if there are 0 bits in the message, there is no information present, and if there are >0 bits in the message, there is information present.
That would be Shannon information duh. Which we have agreed we are not interested in. So let’s move on. Mung #211
Lizzie has constructed a box which information must fit into regardless of what information actually is and regardless of whether information can actually fit into the box she has created. This is the the problem, in a nutshell.
Um, no, it isn’t, and saying so doesn’t make it so. Mung #212
And Her name is Elizabeth Liddle (Lizzie). kairosfocus: Is there a serious objection to this, if so, what is it, why? The objection arises because someone thinks that because this measure cannot tell us the meaning of a message, that it follows by force of deductive logic that information can exist which is completely devoid of meaning. Therefore this person thinks she can generate 100 bits of information simply by flipping a two-sided coin 100 times. It follows that if she were to write a coin-flipping program consisting of necessity (two-sided coin, must come up either heads or tails) and chance (some probability that either heads or tails will appear) then she will have demonstrated that chance + necessity can generate information and will have falsified ID. So why has she not flat-out suggested that she can falsify ID by just such a computer simulation?
Um, I did, ages back, to demonstrate that we needed a definition of information that ID proponents actually called information. And that Shannon information wasn’t it. Obviously. For the precisely the reasons you give above. So we aren’t using Shannon information, right?
Probably because she knows it’s absurd and she would be laughed at and possibly even mocked. Now this same person will put forth a definition of information according to which information brings about a reduction in uncertainty. And then, when asked asked whence this uncertainty arises, and what is it uncertainty about she will say that she must have an operational definition of about. And yet we persist in attempts to carry on a rational conversation with this person.
I’m not at all sure why I persist in trying to carry on a conversation with someone who insists that I am saying the very opposite of the thing I have been trying to say since the beginning of the conversation. But let me try one more time: Shannon information is not an appropriate measure of information for this purpose because it doesn’t incorporate the concept of meaning and we all agree that no ID proponent is proposing that Chance and Necessity can’t create meaningless information. We agreed it about three months ago. So let’s stop talking about it right? We agree. There is no argument. It is a ex-subject. It is pushing up daisies. Nail it the perch right now. The Webster-Merriam definition, as quoted by Meyer seems good to me.Elizabeth Liddle
July 25, 2011
July
07
Jul
25
25
2011
02:15 PM
2
02
15
PM
PDT
Upright BiPed:
That is what I need you to tell me, in sufficiently unambiguous terms that I can derive a set of measurable variables from it.
/ Dr Liddle, before this goes any further, would you please clarify your thought for me. In statistical research, a variable is often a measurement of the relationship between at least two different characteristics. For instance, a researcher might want to know the relationship between the number of persons in a population who are mothers of children under the age of five years, and those who are also over the age of 40. The researcher might then take that relationship and compare it over different cultures to gain an insight into those various cultures. That is an example of one such variable in statistical research. Another example would be a variable such as the asking for the amount of physical exercise within a given population. The variable comes into play in the form of defining what constitutes exercise: is it working out in a gym or is it also riding a bike to the grocery store. Such variables have to be established in order to either include or exclude them from a measurement.
Indeed.
Even so, these examples can be categorically different than asking for the presence of a particular characteristic within a population – where measuring the presence of the characteristic constitutes no more than simply finding that particular distinction. In that sense, there is no variable except the distinction between ‘yes’ and ‘no’.
Yes. But that does not absolve one from providing criteria. With some categories it is pretty easy (sex,for instance, although not gender), others easy-ish (handedness - we usually use a questionnaire with a cut-off score), and others much more arbitary (diagnosis, for instance, again, using a cut-off score on a screening instrument, or a decision from a "consensus diagnostic conference"). But sure. I am familiar with categorical variables.
An example of this would be to ask someone if their age is over 50, or if maximum circumference of their head is less than 60 centimeters, or if they currently possess a valid pilot’s license. These are three such examples of queries that have no variables.
Well, I guess this might be a US/Brit thing (but in which case it's also a US/Canadian thing) but here we call categorical variables "variables" even if they only vary between 0 and 1 (i.e. absent or present). And, as I said, the fact that you code a variable as a categorical variable doesn't absolve one from operationally defining it.
There are also others which have virtually no variable – meaning the variable is so small as to be statistically meaningless – such as ‘are you a male’ or “have you ever broken your tibia’. Even though these types of queries have no variation, or no statistically-viable variation, they are each completely valid queries within a sample. They are measured in research all the time, just as they are.
I'm sorry, you've lost me. By "variable" do you mean "variance"? Assuming you do, and that you are saying "the variance is so small as to be statistically meaningless" - it is quite untrue that the variable has "no statistically-viable variation" if by that you mean no quantifiable variance. Obviously if every subject in your sample had broken their tibia (or none), you would have no variance, and you would exclude the variable from your analysis, or, if very few had broken their tibia, you might find yourself with cell sizes so small that your results would be unreliable, but that isn't the same as having "no variance". In a typical population sample, half your sample will be male and half female. If you dummy code male as 0 and female as 1, the standard deviation of your variable will be .5. If the variable is a predictor,that's all you need to do - you get your regression coefficients, and then your predicted values will include the beta for the variable if your dummy is 1 and not if the dummy is 0 (ie. beta * 0 = 0). If the variable is a dependent variable, then you'd run it as a logistic regression, and your betas will sum in the exponent of the denominator to give you a probability, given your predictors, rather than a predicted value. If you don't mean this, can you explain what you do mean?
Thus far, you have said several times that the conceptual definition we developed is satisfactory, but simply needed to be operationalized.
Well, it's satisfactory if it is operationalisable! I hope it is.
In that process we may find the need to query for characteristics that do not have variation – in other words, they either exist or they do not. And if that is the case, such a situation does not invalidate the query by anyone’s conceptualization of research.
Well, as I said, characteristics that either exist or do not still have variation! Statistically speaking, they have variance. More to the point, if you are going to code them as present or absence you need a criterion by which to decide whether they are present or absent. I have provided one such criterion, using the Merriam-Webster definition, above.
In that vein, what exactly (in concrete terms) is the variable you are expecting to measure in the presence of information?
Well, I've given it above, but I'll give it again. If we define information as an arrangement of matter that has specific effects, we need to operationalise "arrangement of matter" which shouldn't be too hard for my sim as they will be strings of virtual matter; and "specific effects", which again isn't too hard. To ensure that the effects are "functional" ie. "useful" to the virtual organisms, I've specified that they must contribute to reproductive fidelity (i.e. not just create pretty patterns). What is important however (for me), is to make sure that the effects are correlate with the arrangement, and don't just happen haphazardly. Otherwise you win :) We could even set it up as a logistic linear regression, with the arrangements coded categorically, and the effects (reproductive fidelity) coded as continuous variable, and look for a significant beta. So we can be even more specific: If I succeed, my virtual organisms will contain "arrangements of matter" (strings representing "polymers") that affect reproductive fidelity. If certain strings result in greater reproductive fidelity than others, then I will have demonstrated that the strings embody information, because the arrangement will not just have "specific effects" but "specific functional effects" in that they increase reproductive fidelity. Or we could use longevity, I guess, measured in iterations. Look forward to your response. Cheers LizzieElizabeth Liddle
July 25, 2011
July
07
Jul
25
25
2011
01:22 PM
1
01
22
PM
PDT
If information is a prearranged subset P of a larger set S consisting of all permutations of a configuration space of size N, then the only part that should really trip anyone up is the prearranged subset P. From what I gather, any permutation in S is considered potential information by those who have a problem accepting a definition of information that includes P. However the set P in a peptide chain is merely the selection of permutations that equate to functional sequences in set S. We can define set U to be all the arrangements which do not equate to functional sequences. Until there is some sort of agreement on the size of both P and thus U, I'm certain the equivocations will continue. S needs to be the same set as that formed by a union of P and U. Determining the size of P relative to S will likely continue to be controversial. KF provided a reference to an estimate of something like 10^-74 in another thread, but I can't find it at the moment.material.infantacy
July 25, 2011
July
07
Jul
25
25
2011
01:10 PM
1
01
10
PM
PDT
And Her name is Elizabeth Liddle (Lizzie). kairosfocus:
Is there a serious objection to this, if so, what is it, why?
The objection arises because someone thinks that because this measure cannot tell us the meaning of a message, that it follows by force of deductive logic that information can exist which is completely devoid of meaning. Therefore this person thinks she can generate 100 bits of information simply by flipping a two-sided coin 100 times. It follows that if she were to write a coin-flipping program consisting of necessity (two-sided coin, must come up either heads or tails) and chance (some probability that either heads or tails will appear) then she will have demonstrated that chance + necessity can generate information and will have falsified ID. So why has she not flat-out suggested that she can falsify ID by just such a computer simulation? Probably because she knows it's absurd and she would be laughed at and possibly even mocked. Now this same person will put forth a definition of information according to which information brings about a reduction in uncertainty. And then, when asked asked whence this uncertainty arises, and what is it uncertainty about she will say that she must have an operational definition of about. And yet we persist in attempts to carry on a rational conversation with this person.Mung
July 25, 2011
July
07
Jul
25
25
2011
12:44 PM
12
12
44
PM
PDT
Lizzie has constructed a box which information must fit into regardless of what information actually is and regardless of whether information can actually fit into the box she has created. This is the the problem, in a nutshell.Mung
July 25, 2011
July
07
Jul
25
25
2011
12:19 PM
12
12
19
PM
PDT
I suppose that I just imagined that someone suggested that we could use Shannon information because it [Shannon information] can tell us whether information is present or not. Elizabeth Liddle @168:
An operational definition can also give criteria for establishing whether some categorical variable is present or absent. As I said. One that allows you to measure the variable will of course also allow you to say whether it is present or absent. Shannon’s definition does that – if there are 0 bits in the message, there is no information present, and if there are >0 bits in the message, there is information present.
Mung
July 25, 2011
July
07
Jul
25
25
2011
12:04 PM
12
12
04
PM
PDT
Mung:
Please describe the message and the measurement and explain how the measurement demonstrates that there is 0 bits of information in the message.
Elizabeth Liddle:
OK, let’s take a channel that is dominated by noise at a frequency of 50 hertz. We listen to this channel. We fail to detect any departure from the 50 Hz tone. We therefore receive 0 bits of information.
In one sentence the 50Hz is noise. In the next sentence the 50Hz is a tone. So if all that is present is a 50Hz tone, it's the noise that's being dominated, not the tone. But where's the message? Well, you say we cannot detect the message, because we have not detected any departure from the 50 Hz tone. Yet you appear to argue that from this line of reasoning we can conclude that we've measured the message and found no information in it.
We therefore receive 0 bits of information.
We haven't received any message. So we haven't measured a message that contains 0 bits of information. Try again, please.Mung
July 25, 2011
July
07
Jul
25
25
2011
11:49 AM
11
11
49
AM
PDT
That is what I need you to tell me, in sufficiently unambiguous terms that I can derive a set of measurable variables from it.
Dr Liddle, before this goes any further, would you please clarify your thought for me. In statistical research, a variable is often a measurement of the relationship between at least two different characteristics. For instance, a researcher might want to know the relationship between the number of persons in a population who are mothers of children under the age of five years, and those who are also over the age of 40. The researcher might then take that relationship and compare it over different cultures to gain an insight into those various cultures. That is an example of one such variable in statistical research. Another example would be a variable such as the asking for the amount of physical exercise within a given population. The variable comes into play in the form of defining what constitutes exercise: is it working out in a gym or is it also riding a bike to the grocery store. Such variables have to be established in order to either include or exclude them from a measurement. Even so, these examples can be categorically different than asking for the presence of a particular characteristic within a population – where measuring the presence of the characteristic constitutes no more than simply finding that particular distinction. In that sense, there is no variable except the distinction between ‘yes’ and ‘no’. An example of this would be to ask someone if their age is over 50, or if maximum circumference of their head is less than 60 centimeters, or if they currently possess a valid pilot’s license. These are three such examples of queries that have no variables. There are also others which have virtually no variable – meaning the variable is so small as to be statistically meaningless – such as ‘are you a male’ or “have you ever broken your tibia’. Even though these types of queries have no variation, or no statistically-viable variation, they are each completely valid queries within a sample. They are measured in research all the time, just as they are. Thus far, you have said several times that the conceptual definition we developed is satisfactory, but simply needed to be operationalized. In that process we may find the need to query for characteristics that do not have variation – in other words, they either exist or they do not. And if that is the case, such a situation does not invalidate the query by anyone’s conceptualization of research. In that vein, what exactly (in concrete terms) is the variable you are expecting to measure in the presence of information?Upright BiPed
July 25, 2011
July
07
Jul
25
25
2011
09:04 AM
9
09
04
AM
PDT
Dr Bot: The big problem with it, that I have, is that if you make the nucleotides of DNA, the "digital" part of a "computer", the "computer" itself is the whole population over time, because it is at the level of replication that the "switches" (nucleotide) are reset, and the output is different phenotypes. And if you include the environment in the analogy, it is the environment that sets the switches (or at least the mean switch positions) in the population and the output is an adapted population. However, if you regard the organism as the computer, then the "digital" part is the genes, which are switched on and off according to incoming chemical and electro-chemical signals, and the output is the right protein at the right time. Of course sometimes the nucleotides are reset within the organism itself, but that usually means bad news!Elizabeth Liddle
July 25, 2011
July
07
Jul
25
25
2011
03:18 AM
3
03
18
AM
PDT
I just don’t find it a very useful metaphor. But regarding the information content – of course. I fully agree that DNA has information content.
I believe - cynic that I am ;) that it is useful in helping with the argument from analogy - 'It is digital, which is like a computer, computers are designed.'DrBot
July 25, 2011
July
07
Jul
25
25
2011
03:06 AM
3
03
06
AM
PDT
kf: I'll have to get back to your posts later, but meanwhile, you ask:
If not,then why do you seem so reluctant to accept that DNA is digital information bearing based on string data structures?
I completely accept that DNA contains information. I just wouldn't describe it as digital - it seems alphabetic to me (aka a "string" as you say"). But I guess in the sense that alphabetic characters can be demonstrated by ascii code then you could describe it as digital, and the tRNA as the translator from digital to ascii. I just don't find it a very useful metaphor. But regarding the information content - of course. I fully agree that DNA has information content. In fact I think "database" is a good description. Also a "library".Elizabeth Liddle
July 25, 2011
July
07
Jul
25
25
2011
02:43 AM
2
02
43
AM
PDT
1 2 3 4 9

Leave a Reply