Uncommon Descent Serving The Intelligent Design Community

How to Lose a Wittgensteinian Battle

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

Philosophy is a battle against the bewitchment of our intelligence by means of language.

Ludwig Wittgenstein, 1953, aphorism 109

My recent exchanges with Jeffrey Shallit illustrate this aphorism. Our disagreement is not over the substance of the matter. Instead, our disagreement hinges on Shallit’s abuse of language to make a trivial point. Shallit and I disagreed over whether an excerpt from Hamlet’s soliloquy could be considered “random” in any meaningful sense of that word. In the course of that exchange Shallit said this:

Barry, and all ID advocates, need to understand one basic point. It’s one that Wesley Elsberry and I have been harping about for years. Here it is: the opposite of ‘random’ is not ‘designed’.

The problem with Shallit’s assertion is that neither he nor Wesley Elsberry get to decide what “random” means. In linguistic theory words acquire meaning in a language by convention among the speakers of that language, not by diktat, and as I will demonstrate below, in the English language “random” does in fact mean the opposite of “design.”

In order to determine whether “random” is the opposite of “design” we must first establish what those two words mean. Wikipedia defines “random” as follows:

Randomness means lack of pattern or predictability in events. Randomness suggests a non-order or non-coherence in a sequence of symbols or steps, such that there is no intelligible pattern or combination.

Thus, a random string of text is one in which there is no intelligible order, coherence, or pattern.

Webster’s Dictionary defines “design” as follows:

1. to prepare the preliminary sketch or the plans for (a work to be executed), especially to plan the form and structure of;
2. to plan and fashion artistically or skillfully;
3. to intend for a definite purpose;

Any string of text that results from “design” will definitely have an intelligible order or pattern.

Therefore, Shallit is wrong. “Random” is in fact the opposite of “designed.”

Shallit insists, however, that Hamlet is in fact “random” as that term is used in algorithmic information theory. For what he means by this, Wikipedia again:

Algorithmic information theory studies, among other topics, what constitutes a random sequence. The central idea is that a string of bits is random if and only if it is shorter than any computer program that can produce that string (Kolmogorov randomness)—this means that random strings are those that cannot be compressed.

In his first post Shallit ran both a string of keyboard banging gibberish and Hamlet through a computer program,

If we want to test this [i.e. randomness] in a quantitative sense, we can use a lossless compression scheme such as gzip, an implementation of Lempel-Ziv. A truly random file will not be significantly compressible, with very very high probability. So a good test of randomness is simply to attempt to compress the file and see if it is roughly the same size as the original. The larger the produced file, the more random the original string was.

Here are the results. String #1 is of length 502, using the ‘wc’ program. (This also counts characters like the carriage returns separating the lines.) String #2 is of length 545.

Using gzip on Darwin OS on my Mac, I get the following results: string #1 compresses to a file of size 308 and string #2 compresses to a file of size 367. String #2’s compressed version is bigger and therefore more random than string #1: exactly the opposite of what Arrington implied!

What is going on here? Despite the facetious title of my third post Shallit is not barking mad. Nor is he stupid. Why on earth would an obviously intelligent person write a sentence like “[Hamlet’s] compressed version is bigger and therefore more random than [gibberish]”?

Please see the Wittgenstein quotation above. The simple and obvious fact of the matter is that the string from Hamlet does not conform to the English word “random” to even the slightest degree. The string was carefully designed. Therefore, it has zero randomness. Hence, it cannot be “more random” than any string of text that displays any randomness whatsoever. Certainly it cannot be “more random” than a string of gibberish. But in his eagerness to discredit my analysis, Shallit lost sight of that fact. In short, he lost the battle against the bewitchment of his intelligence by means of language.

Sure, the compressed version of Hamlet is bigger than the compressed version of gibberish. And if one insists on defining relative randomness in terms of relative compressibility Hamlet is “more random.” Here’s the problem with that approach. It is glaringly obvious that Hamlet is not in any degree “random” whatsoever as that word is used by English speakers. Therefore, by its very nature it is not subject to a relative randomness analysis except to the extent one observes that it is totally non-random and any string that is even partially random is therefore more random. So what did Shallit accomplish when he insisted that under his esoteric definition of “random” Hamlet is “more random” than gibberish? He made a trivial mathematical point, and in the process made himself look foolish.

My advice to Shallit. Next time you are fighting Wittgenstein’s battle against the bewitchment of your intelligence by means of language, fight harder.

Comments
In comparison, Shallit’s analysis is impartially applicable to both strings. He tests both strings by an independent quantifiable measure.
Put the elephant through the meat grinder and you've got a way of measuring elephants while ignoring the elephant in the room.Mung
October 13, 2014
October
10
Oct
13
13
2014
06:13 PM
6
06
13
PM
PST
E Seigner:
Since Shallit proved the opposite that Barry intended to instruct...
No, all Shallit did was prove that he is deluded. It is very telling that ES just blindly accepts what Shallit did and then turns around and tells us our methodologies are subjective.Joe
October 13, 2014
October
10
Oct
13
13
2014
08:07 AM
8
08
07
AM
PST
kairosfocus
PS: The design inference process is more than adequately clarified and working, for those who do not have problems recognising that a key Shakespeare text is English.
(First off, funny that the little sentence that's conveying a meaning and is even on topic should be prefixed with "PS" while the main text is utterly unintelligible and off the chart from every angle.) Of course English text is in English and speakers of English have no trouble recognizing that this is tautologically so. But what makes this situation a "design inference process"? What is the inference in question and what is its process? Is there anything intelligent here besides the tautology? More importantly, what is the quantifiable measure in ID theory that reliably demonstrates that there is an inference going on and that it's heading the correct way? What is the bit-value of an English text versus of a non-English text? The two strings can be framed up in any number of ways. In one sense, one is gibberish because it's not English while the other is English. In another sense, both strings draw from the same set of characters and they have been produced in the same manner by the same agency, so they are not too different. Who is to tell which considerations are more relevant to the "design inference process" and why? Let's look at another example. Let's say there are two lamps attached to an electrical circuit. The circuit makes the lamps go on and off with perfect regularity. What does this blinking mean? It's perfectly regular and entirely determined by the circuit. Is there a message? Why? Why not? What's the message? Can you tell? What's the bit-value of the complexity and what's the specification that you base your answers on? How did you arrive at the specification? What other specifications did you try and how did you eliminate them? Now, let's consider that the lamps are yellow and they are installed on the left side of a motor vehicle. The vehicle is being operated by a driver and the vehicle is in traffic. In accordance with the traffic code, the blinking lamps here mean "This vehicle wants to turn left." My conclusion is that the context is doing all the work of "design inference process" while the bit-value - which you mostly even fail to compute - tells exactly nothing. Furthermore, given any thorough metaphysical framework I can think of, the number cannot even in principle tell anything that would not be known otherwise. To put it another way, it tells you whatever you want it to tell. To one person it tells one thing, to another person it tells some other thing, so what makes your opinion better than someone else's opinion?E.Seigner
October 13, 2014
October
10
Oct
13
13
2014
07:42 AM
7
07
42
AM
PST
SteRusJon #124
As I have a headache from trying get through to ES, I have nothing more to say directly to him.
I think everyone who has tried to communicate with her so far has had the same experience. I stayed with it far too long.Silver Asiatic
October 13, 2014
October
10
Oct
13
13
2014
05:50 AM
5
05
50
AM
PST
ES: The English language has long been reduced to dictionary status. Long before that, English speakers knew what was good English -- and in fact the cited passage is one of the all-time greatest writings in English by perhaps the greatest writer in English ever. The text and its specification as English text, Elizabethan style, is nit in doubt; your responsiveness to the easily observable facts is. And so far, we are seeing a plain FAIL. Please, think again, and reassess what post modernism and the sophomoric form Kantian ugly gulch between the inner and outer worlds are leading you into. Please. KF PS: The design inference process is more than adequately clarified and working, for those who do not have problems recognising that a key Shakespeare text is English.kairosfocus
October 13, 2014
October
10
Oct
13
13
2014
03:52 AM
3
03
52
AM
PST
SteRusJon
ES "How did you determine that English dictionary was the right specification? Where was the analysis? All I see is a jumping to a conclusion. How did you test the specification?" How did I determine the English dictionary was the right specification? I went down to the Library of Congress and beginning with the dictionary section of the reference section I checked for an entry in each dictionary for each delimited element in both texts, tabulated the ratio of elements found in the dictionary to the total number of elements. Well, not really. In principle I could have. Barry’s example was intended to be instructional and therefore, obviously English text compared to text with no connection whatsoever to any language. He purposely took the grunt work out as it is a teaching moment. For ES to get hung-up on the point is obvious non-concessional hyperskepticism on his part. I am very confident that 100% vs nearly 0% is sufficient justification for my conclusion. Only a hyperskeptic would call that “jumping to a conclusion.”
Long story short, you did not test the other string at all. You did not even think of any specification which it could fit, much less test them. You saw two strings, and recognized in one of them something you liked, and judged the other by the one you liked, without testing. A basic noob fail. In comparison, Shallit's analysis is impartially applicable to both strings. He tests both strings by an independent quantifiable measure. SteRusJon
Also, note I had challenged ES to give me chance generated text that would be incorrectly classified as designed. Nothing was offered.
As it follows from above, you simply go by hunch to classify "chance generated text". You are not someone who could be trusted to classify anything. And we don't even agree on what "chance generated text" means. There have been at least two analyses, yours and Shallit's. The results are absolutely incompatible. The key is obviously not in the strings per se, but in the method or specification by which they are analyzed. And this is what I have been talking about all along. What is the method? Why precisely this specification and none other? How did you eliminate other specifications? Your response is not getting us closer to the answer, because you didn't even consider any other specifications and you thought nobody would notice. SteRusJon
ES: "Nowhere have you *proved* that Shallit’s measure of randomness is irrelevant, you simply assume and assert it is," I made no attempt to *prove* irrelevancy. I justified my opinion that Shallit’s methodology was the incorrect methodology to uncover reason for the obvious difference in Barry’s instructional example.
As the person to whom Barry's instructional example was directly addressed, I am obliged to tell you that Barry's example was not very instructional due to its embedded presuppositions that were problematic for me and which I had brought out explicitly even before Barry entered the discussion without bothering to read up on it properly. Furthermore, Barry's instructional example went conclusively down under the very moment Shallit made an actual computation on it, precisely as ID theory was supposed to be able to do and which I had been requesting an example of. Nobody else has explicitly attempted to quantify the strings. I don't wonder why. Since Shallit proved the opposite that Barry intended to instruct, it's clear - as it was clear to me from the beginning - that the ID theory rests entirely on the presuppositions and "fine-tuned" definitions, not on the calculations. The strings or data per se are irrelevant. The methodology should be applicable to any and all data. So I ask: What's your methodology? Shallit has one that enables impartial quantification of strings. Where's yours? SteRusJon
I have not seen any challenge to Shallit’s “randomness” calculations. Where does he ask for justification for the “randomness” methodology?
It so happens that your approach, walking into the Library of Congress, picking up an English dictionary, etc. is an alternative to his approach. Comparing the results of his test and yours without any prior judgement, it looks on the face of it that there's no support for any notion of "intelligence detection" or "design inference". The presented results are absolutely inconclusive to infer any causal history to the strings, intelligent or otherwise. This happens to confirm my presupposition that such detection or inference is, as a matter of principle, unquantifiable altogether. You really need an airtight methodology to convince me otherwise. SteRusJon
I do not “assume and assert” my specification is correct. I observe that it works. Barry’s example is a test bed of sorts. It is an example of known design verses known non-design. My method unambiguously indicated the true state and Shallit’s failed to correspond to reality. Whatever it is that Shallit’s method demonstrates, one thing is obvious- it does not demonstrate that Hamlet was designed.
Actually, Shallit demonstrated that the so-called random string turned out the same amount of bits as Hamlet. His results of compressibility were 38,6% for string #1 and 32.7% for string #2, so the difference is small, probably statistically irrelevant. Which happens to perfectly conform to reality: Both were composed by a person. You assume and assert that there should be a difference, but since it's obviously true that both strings were composed by a person, I don't get why you would assume and assert it. On the other hand, if the measured amounts for the strings were widely different whichever way, and then, given another specification, the measured amounts for the strings were again different in the other way, it would again go to prove the complete futility of the ID theory. Please get your methodology straight and clarify - to yourself - what it is you intend to prove, if anything.E.Seigner
October 13, 2014
October
10
Oct
13
13
2014
02:51 AM
2
02
51
AM
PST
F/N: If you want to generate true random numbers, try a zener diode based ckt: http://holdenc.altervista.org/avalanche/ (To flatten off, if desired, use something like a Johnson counter ckt and use the random seed to manipulate it.) More with onwards: http://en.wikipedia.org/wiki/Hardware_random_number_generator Using a GM counter to directly get flat-random o/p using an interval counter tuned to the expected inter-pulse interval: http://www.ciphergoth.org/crypto/unbiasing/ (Sky noise can also be used) Here's a nice paper: http://wayback.archive.org/web/20110722112209/http://www.letech-rng.jp/SNA+MC2010-Paper.pdf In short, there is a can-do. To get truly random ASCII text, feed the bit pattern into a txt file then simply read it off. But all of this is tangential, I did this because obviously my repeated references to Zener ckts has been ignored through the zero concessions to IDiots policy. (Maybe, I need to point out that I have had occasion to teach electronics, with some modicum of success . . . ) On the main point, the fact that a zener random ckt is not naturally flat, but is in fact random, should suffice to shoot down one of the flying wild geese. Randomness in real world contexts does not demand, flat random. Yes, for statistical sampling, for proper randomness each pop member should have an even chance of being selected. That's for a particular purpose. Next, randomised blind text typing is sufficiently at-random for toy example didactic purposes, showing enough evident lack of intelligently directed purposeful contingency to be illustrative. If one wants in future annotate as a toy example. The red herrings having been addressed, we can now draw attention to what has been ducked and side tracked from for weeks: the two text sequences are patently and instantly recognisably different, i.e. the difference is observable. S2 conforms to English text, S1 does not and shows lack of purposeful control sufficient to justify a chance conclusion. Both are complex 535 ASCII characters per my Libre Office count. One, shows functionally specific complex organisation and associated information, S2. As expected, the design inference filter process, however implemented, indicates design. This conforms to known history. Yet another illustration of its reliability in positive detection of design, its intended purpose. The same, says of S1, not designed. Maybe, that is wrong, there is a code lurking that is disguised as toylike randomness. So what? the design filter is not designed to have a low incidence of negative errors, it is happy to be conservative so that when it does rule design it is extremely reliable. As is again shown. For, the design inference process is not intended to be a universal decoder algorithm, which in any case is not expected on theory of computation. Likewise, it is fairly easy to construct a metric model of the design inference process, here on solar system scale: Chi_500 = I*S - 500, bots beyond the solar system threshold Where I is an index in bits of information carrying capacity per the usual ways we so measure S is a dummy variable defaulting 0 [chance] and set to 1 if there is a good reason to identify functional specificity [in this case English language text] 500 is a threshold in bits If Chi_500 goes positive, design is inferred. Simple, and directly related to other considerations in light of what the solar system can do in 10^17 s with 10^57 atoms, and a rate of action equal to fast chem rxns. But of course there will be endless selectively hyperskeptical objections and side tracks, as there is a zero concession policy in force. The fallacy of the closed hyerskeptical mind, in a nutshell. Time for a fresh start . . . KF PS: Another rain band passed, and there is wind in gusts but not great strength yetkairosfocus
October 13, 2014
October
10
Oct
13
13
2014
12:47 AM
12
12
47
AM
PST
As I have a headache from trying get through to ES, I have nothing more to say directly to him. Lest anyone else think I have no response to his latest comment directed to me, I submit the following. ES wrote:
The specification being English dictionary? Well, you didn’t do the proper analysis. How did you determine that English dictionary was the right specification? Where was the analysis? All I see is a jumping to a conclusion. How did you test the specification?
How did I determine the English dictionary was the right specification? I went down to the Library of Congress and beginning with the dictionary section of the reference section I checked for an entry in each dictionary for each delimited element in both texts, tabulated the ratio of elements found in the dictionary to the total number of elements. Well, not really. In principle I could have. Barry's example was intended to be instructional and therefore, obviously English text compared to text with no connection whatsoever to any language. He purposely took the grunt work out as it is a teaching moment. For ES to get hung-up on the point is obvious non-concessional hyperskepticism on his part. I am very confident that 100% vs nearly 0% is sufficient justification for my conclusion. Only a hyperskeptic would call that "jumping to a conclusion." Also, note I had challenged ES to give me chance generated text that would be incorrectly classified as designed. Nothing was offered. ES queried
How did you ensure there’s no other better more relevant specification?
If the significant portion of the elements had not been found in the dictionary, I could have some sympathy for ES asking this question. Again hyperskepticism on his part. In reality, for design detection where the case will not be as cut and dried as Barry's instructional example, this is a perfectly valid concern. I am not opposed to making sure, to the extent possible, that a positive for design is well founded. In fact, there is additional indication of design in the example such as, conformance to sentence structure conventions. There other reasons beyond the unambiguous 100% verse nearly 0% result. ES charged
Nowhere have you *proved* that Shallit’s measure of randomness is irrelevant, you simply assume and assert it is,
I made no attempt to *prove* irrelevancy. I justified my opinion that Shallit's methodology was the incorrect methodology to uncover reason for the obvious difference in Barry's instructional example. In my opinion, any methodology that leads one to conclude that two differing things are not different when you are seeking to understand how they differ is the wrong methodology to use. ES's hyperskepticism is showing, again. He is unable to hold himself to the same level of skepticism as he holds me. I have not seen any challenge to Shallit's "randomness" calculations. Where does he ask for justification for the "randomness" methodology? I have not seen even a hint of concern that Shallit may be doing it all wrong. Why does Shallit get a pass? ES continued
and you assume and assert your specification is right, but the question was all along: How? Why? By what measure? By what method?
I do not "assume and assert" my specification is correct. I observe that it works. Barry's example is a test bed of sorts. It is an example of known design verses known non-design. My method unambiguously indicated the true state and Shallit's failed to correspond to reality. Whatever it is that Shallit's method demonstrates, one thing is obvious- it does not demonstrate that Hamlet was designed. That was the challenge. Is there room for increased rigor with the tool I employed? My God! Yes! I invite ES to help make design detection a rigorous, reliable endeavor. Design exists. Why not participate in filling a tool box with tools suited for the job? But we can't even get ES to give it half a chance. He is committed to hyperskeptical obstructionist tactics to defend his worldview. I wonder if ES would have pitched the Rosetta Stone onto the rubble pile. He has no interest in detecting design. All rocks are just rocks! To quote Barry from above, *sigh* ES now has the last word, he may do with as he wishes. StephenSteRusJon
October 12, 2014
October
10
Oct
12
12
2014
06:49 PM
6
06
49
PM
PST
E.Seigner:
A random sample does not come about haphazardly or carelessly. It is meticulously and thoroughly defined in statistics, just like all scientific concepts are.
IOW, by design. Thank you.Mung
October 12, 2014
October
10
Oct
12
12
2014
06:41 PM
6
06
41
PM
PST
SteRuSjon
I made the proper analysis. Detected 100% conformance to a independent specification for one text verses a nearly 0% conformance to the identical specification
The specification being English dictionary? Well, you didn't do the proper analysis. How did you determine that English dictionary was the right specification? Where was the analysis? All I see is a jumping to a conclusion. How did you test the specification? How did you ensure there's no other better more relevant specification? Nowhere have you *proved* that Shallit's measure of randomness is irrelevant, you simply assume and assert it is, and you assume and assert your specification is right, but the question was all along: How? Why? By what measure? By what method? Silver Asiatic
Hilarious. Not one word about how the method ensures randomness or what random means. Let’s forget that I make my living professionally with data sampling and instead watch you copy/pasting things you just discovered yesterday.
You talked about it as if you never had any idea about data sampling, as if you never read anything about it, yesterday or any other day. You talked about it as if you could generate random strings by haphazardly banging your keyboard. But if you know this is not the case, then don't pretend it is. That would not be detrimental to your scientific rigor, not hilarious at all. I don't care what random is by itself. The discussion was how to detect design, how to distinguish it from other things, such as randomness or laws of nature. The question was how randomness and design relate to each other, but now it appears that random has so many disparate definitions that it barely has any relation even to itself. All the worse for design.E.Seigner
October 11, 2014
October
10
Oct
11
11
2014
11:39 PM
11
11
39
PM
PST
ES
sampling process comprises several stages: Defining the population of concern Specifying a sampling frame, a set of items or events possible to measure Specifying a sampling method for selecting items or events from the frame Determining the sample size Implementing the sampling plan Sampling and data collecting Data which can be selected
Hilarious. Not one word about how the method ensures randomness or what random means. Let's forget that I make my living professionally with data sampling and instead watch you copy/pasting things you just discovered yesterday.Silver Asiatic
October 11, 2014
October
10
Oct
11
11
2014
08:44 PM
8
08
44
PM
PST
ES: Way to go! Zero concession policy intact. You wrote "Now, why would someone not grant the distinction of the strings the way Barry posited it?" I say, "It is because Shallit has a zero concession policy, as well." Barry was trying to lead a horse to water for a refreshing drink with an obvious and, he thought, an unambiguous didactic scenario. Shallit dug himself in with all four shoes and found a "gotcha". I made the proper analysis. Detected 100% conformance to a independent specification for one text verses a nearly 0% conformance to the identical specification for the (proxi for) chance generated text. No recognizable conformance to any other specification for the text string (that is a proxi for chance generation) of comparable length of the Hamlet text either. Result: True positive detection of the designed text. Negative (possibly false) detection of design for the (proxi for) chance generated text. Care to venture if I would get a false positive for any truly chance generated text of such length you may care to send my way? Any methodology that fails to reflect the obvious difference between the two text strings must be the wrong methodology for detecting that type of difference. Shallit's method could not differentiate between the two. Mine did. Who should you be patting on the back? Shallit is not stupid. He saw the obvious and knows the real difference between the two texts and chose to ignore it. Neither he, nor you will likely ever concede the point. StephenSteRusJon
October 11, 2014
October
10
Oct
11
11
2014
05:55 PM
5
05
55
PM
PST
ES:
Shallit picked up the strings and actually calculated the randomness.
Shallit calculated the randomness of a 100% non-random string? *sigh*Barry Arrington
October 11, 2014
October
10
Oct
11
11
2014
04:09 PM
4
04
09
PM
PST
@SteRusJon
If the designed group is less complex than the chance group, there must be something other than complexity that allows you to detect design. What do you think that something is?
The problem of the "challenge" always lied in the premise. It presupposes that one of the strings is designed and the other is random. However, for someone who doesn't grant that the distinction between the two strings is that of design versus random, the end-question does not make sense and it doesn't even begin to be a challenge. Now, why would someone not grant the distinction of the strings the way Barry posited it? Well, for one because the discussion was about defining and measuring the distinction, not merely assuming or asserting it. And voila, Shallit picked up the strings and actually calculated the randomness. The measurements ended up in favour of those who hadn't granted the distinction posited in the challenge. This is a shame because ID theorists are, according to themselves, experts in telling random stuff apart from designed stuff and they can calculate it too.E.Seigner
October 11, 2014
October
10
Oct
11
11
2014
03:01 PM
3
03
01
PM
PST
Silver Asiatic
Again, in statistical research, we generate a random sample. It’s a pretty common understanding and usage in scientific research. Clearly, we don’t hear anything from critics like Shallit when research is done on a random sample of a population.
A random sample does not come about haphazardly or carelessly. It is meticulously and thoroughly defined in statistics, just like all scientific concepts are. Wikipedia introduces it this way: The sampling process comprises several stages: Defining the population of concern Specifying a sampling frame, a set of items or events possible to measure Specifying a sampling method for selecting items or events from the frame Determining the sample size Implementing the sampling plan Sampling and data collecting Data which can be selected Silver Asiatic
"Also, as Dieb points out, it illustrates how difficult it is to create a string which has equal probability of each any character in each position." Another fascinating tangent. It could lead to a very fruitful discussion on how to generate random numbers...
This is not another tangent. It is the same topic of randomness versus design. Despite all attempts at deviation and redefinition, it's the central point of the original context where Barry first produced the two strings.E.Seigner
October 11, 2014
October
10
Oct
11
11
2014
01:57 PM
1
01
57
PM
PST
In response to Steno in #95 above In Barry's "Tale of Two Text" originating post #274 of the thread “On not putting all your theological eggs into one basket” Barry asked a question in reference to the two text examples. I quote the question with context.
Both groups are complex. One group was constructed through random strokes on a keyboard. The other was designed. Can you tell which is which? Certainly you can. And just as certainly it is not the degree of complexity that allows you to tell the difference. If anything, the random group is more complex than the designed group. If the designed group is less complex than the chance group, there must be something other than complexity that allows you to detect design. What do you think that something is?
Where, pray tell, is there a challenge to identify how they share randomness? The challenge is to identify what makes one obviously different than the other. That challenge requires the ID opponent to consider non-physical realities such as ideas, concepts, messages, functionality, meanings, specificity etc. A place that the ID opponent refuses to go. ID opponents- What properties or characteristics are present in Hamlet’s soliloquy that are not present in any random text string that suits your fancy? How can we go about putting those empirical observations onto a firm scientific footing? Are you going to advance science or continue to obstruct? Steno, you assert that the only way one can know that Hamlet is designed is to know it was designed. That is patently false. It is obvious that the passage is composed of individual elements (delimited by spaces) that conform to a specification (a comprehensive English dictionary.) That specifying dictionary lays out, for all to see, the arbitrary linkage between the ASCII text elements of the passage and concepts the elements represent. 100% of the delimited elements of the passage is mapped into the specifying dictionary. Barry's "random", "haphazard", "gibberish" or whatever else you might wish to call it or your own "whatever suits your fancy random" text will conform to that specifying dictionary with nearly 0% match. Where it does it is almost always trivially so as in matching one, two or, maybe, three letter words without any delimiting of elements. That is the correct way to begin to analyse the challenge Barry originally set forward. There is more than that. I have not broached sentences, syntax, relevance of the individual thoughts to the whole, etc. These are all signs of specificity and not, in anyone's wildest imagination, possible by any set of random events. A scientific mind would want to understand the ins and outs of all that and seek to formalize that understanding for further application to the world around us. If that formalization has anything to say about the questions of origins, a seeker of truth would ascent to its decree. Shallit derailed the discussion, I believe deliberately, from the beginning. MF, ES, Steno and others have gleefully piled on to his folly. As I see things, from the way ID opponents staunchly resist conceding one microscopic point and insist on discussing irrelevancies, such as, how "random" a textual sample is that does not contain a single random character, while ignoring/denying the obvious conformity to an independent specification, ID opponents are not in least interested in the truth value of the IDist's proposal. Stephen PS. Eric Anderson and I exchanged some similar thoughts briefly at the tail end of the now stale portion of this multi-OP "KF Cuts to the Chase (Again)" started on Oct.5 I have seen this flavor of denial in years past and recount it, briefly, there.SteRusJon
October 11, 2014
October
10
Oct
11
11
2014
12:42 PM
12
12
42
PM
PST
Shallit:
“random” in the sense of likely to have been generated randomly and uniformly
Interesting definition - "random" in the sense of likely to have been generated randomly (as strange as that might sound). Again, in statistical research, we generate a random sample. It's a pretty common understanding and usage in scientific research. Clearly, we don't hear anything from critics like Shallit when research is done on a random sample of a population. To be consistent, I'd expect him to reject every analysis that uses that sense of randomness. * What purpose did Shallit have in showing that this particular string was less random than the speech from Hamlet MF:
Mostly to show up your limited expertise I guess.
You're probably right. It's an attempt "show up" on an irrelevant tangent (as mentioned several times). I read that as he couldn't refute BA's main point and therefore ran after a distraction in order to prove victorious. It's very much like finding a spelling error in a thesis and then spending several weeks trumpeting that 'weakness' in the paper. I'm just amazed at how childish this one really is.
Also, as Dieb points out, it illustrates how difficult it is to create a string which has equal probability of each any character in each position.
Another fascinating tangent. It could lead to a very fruitful discussion on how to generate random numbers - and with that, perhaps we can all forget about the point that BA made so conclusively (and which has only gotten stronger with further discussion). :-)Silver Asiatic
October 11, 2014
October
10
Oct
11
11
2014
12:42 PM
12
12
42
PM
PST
Mark Frank:
Also, as Dieb points out, it illustrates how difficult it is to create a string which has equal probability of each any character in each position. /blockquote> And as has been pointed out that is a useless definition of random because you first have to know all of the probabilities before making the calculation.
Joe
October 11, 2014
October
10
Oct
11
11
2014
05:53 AM
5
05
53
AM
PST
MF: First, in the course of discussion, it was clear that Mr Shallit did in fact claim that on a compression test, a Hamlet cite was giving a higher randomness index than gibberish. This raises serious questions regarding the definitions he used and the applicability of resistance to compression per se as an index of randomness. Secondly your attention has been repeatedly drawn to the Trevors and Abel paper of 2005, and to the fig 4 that resolves the matter. This of course you have studiously refused to address cogently and on the merits. That consistent unresponsiveness on the merits speaks again to the zero concessions to IDiots policy of disregard for truth, evidence and disrespect for people, and it shows why you have been subject to direct rebuttals and some plain speaking. Where of course, conveniently, you will predictably ignore what I have said here on the excuse you no speaka da Inglish rite. (Onlookers, someone else will have to draw his attention to the matter for him to deign to respond, usually evasively or dismissively, on years long track record.) Never mind, let's focus the substance. In essence, random sequences are highly incompressible, orderly repetitive ones are highly compressible, and functionally coded ones [whether algorithmic or linguistic makes little difference] are intermediate and often towards the highly resistant end. But for very different reasons. As Orgel and Wicken highlighted across the 70's, functionally specified complex organisation/information is distinct from both randomness and order. It (e.g. the text of the last sentence) is aperiodic, so of much less redundancy than an orderly repetitive sequence like sdsdsdsd . . . but it is not as resistant as a flat random distribution for which effectively the typically shortest way to express it is to simply quote it:t0isau4tpw2l . . . Where, as the just previous sentence shows, FSCO/I may even include orderly and random sub sequences. Note, onlookers, this three fold distinction was put on the record by OOL researchers in the '70's and was highlighted by Thaxton et al in the very first technical ID work, TMLO in 1984. It has been consistently cited since and continues to be highlighted today. Refusal to acknowledge this and setting up and knocking over strawmen such as Mr Shallit indulged, are without any reasonable excuse. But, they aptly illustrate the absurd lengths to which the zero concessions to IDiots policy will go, and how once marching orders are sent forth, the usual online foot-soldiers will predictably dutifully toe the line even if it means marching off a cliff into absurdity. That clinging to absurdity is a diagnostic sign that something has gone drastically wrong, and it is a sign that a sea-change will be a-coming. Absurdities backed by ideologies with power can only be sustained so long. At minimum, people will vote with their feet and walk away. As, is obviously happening. ANYTHING that says or suggests -- makes but little rhetorical difference -- that a clip of Hamlet is per some metric or other of a higher randomness metric than gibberish text, is patently absurd. Instead, let us have the good sense to face the point Trevors and Abel have aptly made nine years ago . . . 9 yrs ago. (See the slight compression?) KFkairosfocus
October 11, 2014
October
10
Oct
11
11
2014
01:20 AM
1
01
20
AM
PST
F/N: Strictly, the log metric (or the comparable chain of yes/no q's) measures information-carrying capacity. Common cases of info, such as file sizes, are about functionally specific complex info measured in that way. Functionally specific complex organisation such as for a car engine etc (as AutoCAD shows) may be reduced to structured descriptive strings based on suitable coding that structures and answers a chain of y/n q's. So, discussion on bit strings is WLOG. All of this has been discussed, summarised, explained, highlighted over and over and should be common backdrop absent the zero concessions to IDiots policy that has obviously long been in force . . . one of the rhetorical tactics. Going beyond, once we get to 500 - 1,000+ strings it becomes practically utterly implausible for a config space, needle in haystack search to find the implied islands of FSCO/I . . . multiple, well-matched, properly arranged and coupled parts to achieve function lead to this pattern . . . in the astronomically large sea of non-functional configs. The only observed, empirically reliable and needle in haystack analysis plausible explanation for FSCO/I is design. This is the heart of the design inference on FSCO/I, which on trillions of directly observed cases has a 100% success record as, say, the Internet shows. But to acknowledge that would be to yield the case, so we see everything and anything but that. KFkairosfocus
October 11, 2014
October
10
Oct
11
11
2014
12:48 AM
12
12
48
AM
PST
I seem to be unable to give up my daily insult from Barry. It is clear from #98 he was not satisfied with my response to his question:
What purpose is served by concluding that Hamlet is “more random” than gibberish?
Like most things Barry writes it can be interpreted many ways but let’s pick out two: * In general what use is there for measuring randomness using KC compression Barry – just Google “uses of kolmogorov complexity”. This set of lecture notes seems particularly relevant, look at slide 22. Did you really think this method of measuring randomness became so established without any applications? One application that is particularly relevant to ID is that allows one to assess whether a string was created through independent and equal probability of any character in each position without having to know anything about how the string was created, what functions it can perform or any meaning it might have. * What purpose did Shallit have in showing that this particular string was less random than the speech from Hamlet  Mostly to show up your limited expertise I guess. Also, as Dieb points out, it illustrates how difficult it is to create a string which has equal probability of each any character in each position. The apparently gibberish string is actually much more determined (although not designed) than it appears.Mark Frank
October 10, 2014
October
10
Oct
10
10
2014
11:12 PM
11
11
12
PM
PST
REC @ 102 (quoting Richard Feynman):
This is actually the most information you can get with this particular choice of symbols.
Mung @ 109:
The most information you can get about what?
I think I see the problem.sergmendes
October 10, 2014
October
10
Oct
10
10
2014
09:15 PM
9
09
15
PM
PST
REC:
This is actually the most information you can get with this particular choice of symbols.
The most information you can get about what?Mung
October 10, 2014
October
10
Oct
10
10
2014
09:05 PM
9
09
05
PM
PST
REC, you are not introducing anything we haven't seen before. https://uncommondescent.com/intelligent-design/william-dembski-on-what-information-is/Mung
October 10, 2014
October
10
Oct
10
10
2014
09:01 PM
9
09
01
PM
PST
REC @ 106, So?Mung
October 10, 2014
October
10
Oct
10
10
2014
08:54 PM
8
08
54
PM
PST
https://uncommondescent.com/intelligent-design/william-dembski-on-what-information-is/ "What concept could be more fundamental to intelligent design theory than “information”? None, but what exactly does the word mean in the context that we use it? William Dembski explains: Information....has many meanings. The way I use it is as a ruling out of possibilities. ~20 seconds.REC
October 10, 2014
October
10
Oct
10
10
2014
08:28 PM
8
08
28
PM
PST
1 bit of what? Oh, I don't know. Let's call it "information." And let's create from this an entire field and call it "Information Theory." It doesn't tell us what information is, not even in theory. And it doesn't tell us whether we are in fact measuring "information." And why not? Whatever. We now have this way to measure "something" given a finite set of symbols and a probability distribution, so let's say that our "uncertainty" is at a maximum given certain conditions or that our "uncertainty" has been reduced by this amount, and since uncertainty is always uncertainty about something, let's call this amount "information."Mung
October 10, 2014
October
10
Oct
10
10
2014
07:58 PM
7
07
58
PM
PST
REC, why complicate things without necessity? Flip a coin for goodness sake. Given a finite set of symbols (2 aka H/T or 0/1) chosen at random (equiprobable aka the probability distribution). We can calculate the maximum entropy and assign a measure. 1 bit. whoop dee doo. What is the coin toss about? Is it about the team that gets to select whether to receive the kickoff in a game of american football? Is the fact that the meaning of the coin toss is not derivable from the mathematics somehow relevant to your argument? How so?Mung
October 10, 2014
October
10
Oct
10
10
2014
07:46 PM
7
07
46
PM
PST
*sigh*Mung
October 10, 2014
October
10
Oct
10
10
2014
05:54 PM
5
05
54
PM
PST
"Instead, it points to the attempted application of KC in an area where the results of the method lead to error" You're striking awfully close to classic critiques of a former owner of this blog. "Perhaps the best way to demonstrate the difference between our definition of information and the everyday term is to consider a random message, that is, an N-bit binary string with random bits. If all possible strings are allowable messages, and all are equally likely (which will happen if each bit is equally likely to be 0 or 1), then the information in such a message will be: I = log2(2N) = N (4.24) This is actually the most information you can get with this particular choice of symbols. No other type of message will reach I=N. Now surely this doesn't make sense - how can a random string contain any information, let alone the maximum amount? Surely we must be using the wrong definition of "information"? But if you think about it, the N-bit strings could each label a message, as we discussed earlier, and receiving a particular string singles out which of the 2N possible messages we could get that we are actually getting. In this sense, the string contains a lot of "information". Receiving the message changes your circumstance from not knowing what it was to now knowing what it is; and the more possible messages you could have received, the more "surprised", or enlightened, you are when you get a specific one. If you like, the difference between your initial uncertainty and final certainty is very great, and this is what matters." http://www.astroscu.unam.mx/~angel/tsb/Feynman.htmREC
October 10, 2014
October
10
Oct
10
10
2014
04:55 PM
4
04
55
PM
PST
1 2 3 5

Leave a Reply