|
Author
|
Topic: Does intelligence imply “motive”?
|
gedanken
Member
Member # 594
|
posted 14. September 2003 10:02
Mike said:
quote:
Hey, I gave the rock/statue pictures the toddler test.
Exactly! (I mean by that the Easter Island figures contain clues that are basic to representationalism, experienced even by a toddler who has made primitive drawings. But then Mike goes on to suggest what I consider to be an incorrect conclusion—that somehow the point gives difficulty for my presentation.)
(Also my “next post” mentioned above must be delayed – I must report – due to it consisting now of a 30+ page paper which truly needs to be presented in PDF form. It was already such when I started this thread, but I realize that a substantial part is needed for that presentation, and also that this thread has given input that needs to be incorporated. I don’t think it would be sufficient to try to put portions in ASCII text with as much symbology as it contains. Also it may not be quite suitable for ISCID in its present form. I need to finish it, deal with its disposition, then will report either it or summarize it when ready. This will take some time, so I am going to answer intervening questions.)
Now Mike should look back over my thread in which I emphasized that I think that ID recognition is very natural and can even be well supported scientifically. (One has to resolve difficulties of precise definitions of terms like “intelligence”—but that is doable.)
My main point in several posts is that auxiliary information that is available in our experience of “statues” is very helpful. However I don’t think that auxiliary information in our experience is captured in “the specification of ‘human anatomical form’”, combined with the eliminative test.
Mike, you really need to read my posts in greater detail, because I am going to virtually completely agree with the individual statements you gave. They do not contradict my thesis in any way. In fact the child’s recognition of faces, combined with our experience at artistic drawings which contain only the bare minimum lines necessary for a child (or ourselves) to recognize the face or body, is part of my argument.
It is not the essence of “human form” that is recognized, it is the essence of our human pattern recognition of the human form—there is a distinction. Mike, look back for the word “representation” and read and understand surrounding argument and context in my posts. A cubist “representation” of a face could have such a nose as the stone face—it would be an essential part. But once again there is much more to the argument, I shall not try to repeat it in few words.
If you wish to give a different specification, I’ll work with that one.
----
[Added in edit]
Let me start out with a variant “specification”: The human recognition of the human anatomical form. (That is more closely what Mike has done tests on.)
My point is that in the logic flow version of the EF, without the use of the “extremal set” optimization, we have a crude measure. All the stone faces fit either Dr. Nelson’s, or this latest specification. What this demonstrates is that the specification is one that is easily met, and thus one that has relatively high probability. Thus the Easter Island figures meeting that specification is irrelevant, no ID is inferred by the formal procedure! (That’s my point—people recognize the ID clearly, yet the formal procedure does not!!!!!!!! Can I add some more exclamation points!!!!!)
So now deal with the “extremal set” optimization in the “generic chance elimination” procedure. We definitely have the potential for a better result—precisely because we have a possibility of a continuous measure rather than an on/off switch type of specification of the pattern can capture more features. NOW put the specification in terms of a completely ordered function of the observation! Or at least it has to be in terms of a pattern recognition set classifying function which can be parametrically optimized. AND it has to be a function that was specified independent of the event! Mike, deal with the issue of constructing a such a function! Without doing so, without having the optimizing effects, we have not improved the situation and still have a high probability match for the specification, thus no inference.
Remember furthermore that I did not claim that the problem in the Easter Island and associated random figures is a false positive of the formal procedure. Rather this is an instance of ID enthusiasts claiming that the formal procedure does “infer design” when in fact it does not. This is a different kind of failure—a failure to represent the inference procedure correctly.
Then get back to this new specification I proposed: “The human recognition of the human anatomical form.” I claim that it smuggles information into the argument that it does not make explicit. That which is most easily humanly recognizable as human form (even for a child) is that which has high likelihood of similar construction characteristics as is a child’s drawing. It is either a direct human figure (not a representation and thus “natural”) or it is a representation optimized for human figure recognition (and thus of raised probability of human construction). Choice of these sorts of specifications for test smuggle this type of information into the actual procedure, and falsely bias it use as an inductive “test” of the EF procedure reliability.
Now the “toddler” recognition specification is quite different from a specification of “human form”. In this case the specification is “toddlers will recognize as form that will be labeled in ‘human form’ categories by toddler upon viewing”. This is an interesting specification—precisely because it lends itself to recognizing representationalism. Representational art is optimized toward recognition as almost an archetype, certainly to being widely recognized by humans using natural human processes. (We can almost catalogue the recognition aspects in terms of brain area that is stimulated. I believe there were some studies on infants recognizing faces, for example, and brain stimulation maps.) In this form we have finally found a restrictive specification that classifies the Easter Island figures. (But how often has a toddler or child recognized a random pattern as though it was designed? For example see a min in clouds, on an inkblot? Differentiating the specific set of pictures given here is not sufficient to indicate that this particular specification gives a low probability rejection region.) But more important is the degree that it smuggles in the knowledge representational figures as a factor of likely construction. Toddlers regularly construct representational artworks of human forms, and also recognize such. This is information that leads to knowledge that construction could be likely, including motive, by humans. Only by smuggling such knowledge into the specification can we find a specification that approaches differentiating the Easter Island figures with the Explanatory Filter.
Pim makes some interesting points. I would also like to remind people that I think that “scientific” ID recognition procedures will not be strictly eliminative. We can improve them by not dogmatically adhering to a requirement that an ID recognition procedure be eliminative, such as considering aspects of the “intelligent agent” like “motive”. That is the main subject I have carved out for this thread.
[ 14. September 2003, 14:10: Message edited by: gedanken ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 14. September 2003 15:02
I'm a little confused by the profusion of exclamation marks. The GCEA isn't supposed to identify all instances of intelligent design. Rather, it's a conservative test that will miss many cases of ID but in so doing will also avoid assigning ID to things that are not ID--at least that is the claim.
Or were you making a different point?
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 14. September 2003 15:32
Thanks, Rex, I’ll make this a short summary of main points—the exclamations were to emphasize that the point above is ignored by ID enthusiasts but is significant, and I had meant as presented in context. This case is not of a false positive. It is a misrepresentation of the EF as having inferred design in the Easter Island figures. In fact one cannot have a high rate of “false positives” without missing natural cause “distributions” (or an alpha significance threshold near or greater than the actual but unexamined joint probability of intelligent designer action). If the random features seem to match the specification, it shows a known and recognized raising of probability of rejection region, not a false positive.
Easter Island figures could not possibly be a “false positive” since it is accepted as ID.
For direct specification of “human anatomical form” Easter Island figures don’t show good conformance, and random stone faces show that we have some conformance by probabilistic means at low level of conformance. “Toddler test” is not accurate measurement of this specification since it does not test actual geometric or relational conformance.
For distinct specification of “representational” artistic human form we still have problems. Easter Island had higher match—but so do some random stone faces. “Toddler test” is more highly accurate measurement of this specification since it tests recognition of apparent representation.
Both cases raise the probability of meeting a comparable formal specification by probabilistic means, and thus raise the probability in the rejection region “extremal set”—certainly way above the UPB. Thus the EF does not infer design in this case. They do so because known random processes can easily make matches at significant (but somewhat low) probabilities—exemplified by the random figures. (But of course the significance measure alpha as low as UPB is not needed in this case, but that cannot be known by eliminative means since that would mean considering properties of the likely intelligent cause.)
In the long run the best match is to truly capture the essence of representationalism in the Easter Island figures (not a specification that has been given by any ID enthusiast). This has highest differentiating power—precisely because it smuggles in the likelihood of human creation. This is closest specification to what is recognized by “toddler test”, though somewhat lacking objectivity.
Using other features comparatively can increase the comparison of Easter Island figures as best explanation argument, but these features cannot be used in eliminative argument because they must remain “independent of the event” which might be hypothesized comparatively. But due to restrictions of “independence” the features that are actually compared miss much of the information available, even intuitively, in the Easter Island figures, thus missing many opportunities to restrict and lower the extremal “rejection region” probability.
Making the smuggled side-knowledge of likely human creation explicitly and formally known (in our examples given so far in this thread) would make a superior filter (but not entirely eliminative). Not doing so typically causes one of four kinds of failure in many circumstances:
- Formal EF procedure does not capture “design” and has little value. (This is the known and accepted failure of false negative.)
- Formal EF procedure does not capture “design” and is misrepresented as doing so.
- Formal “specification” smuggles in sufficient side-knowledge of causation to make distinction clear, but in doing so is not really “independent of the event”.
- Missed joint probability distributions of natural non-intelligent causes may outweigh the un-studied joint probability distributions of intelligent causes, thus causing the Filter to exhibit high rate of false positives.
The last is the actual reliability failure which requires the kind of example like the slot machine with unexplained string of “D” symbols.
[ 14. September 2003, 22:16: Message edited by: gedanken ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 14. September 2003 22:27
Makes sense. I agree. Thanks for the clarification.
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 24. September 2003 01:30
From another thread which discussed this subject, Alonso said the following. It suggests an explanation that might possibly not have been entirely clear due to the length of this thread. I am going to amplify on a critical point, at the risk it is redundant with what I’ve already said.
quote:
Ged writes:
quote:
In any case now we have established that the “specification” of the human form can be met with high probability. Thus that specification cannot be used as a claim of inferring design with the EF, whether the EF is reliable or not. To do so is to make a false claim of a positive result. Now of course we can recognize that the Easter Island figures were done by humans, but the way we do that is not to simply recognize that they are of a specification of a human form—having just agreed that specification has high probability of being met. Rather we need a more detailed specification, and when we add more details we are no longer “independent of the event”.
First, it is no surprise that specifications can arise with high probability. Since, high probability events and low probability events are distinct from specifications. Consider for example randomly banging on your keyboard and accidentally writing the word "the". Nonetheless, it is the marriage of independent patterns and low probability events which causes us to infer design. There is no need to add more information the independent pattern, such as with the Easter Island statues, because the independent pattern itself is embedded within a complex state such that inference to design is trivial (which is why people don't think that the human faces on clouds they think they see are designed). Complexity breaks the ambiguity. Thus, it is completely possible to calculate the probability of an event arising by chance, and independently identifying the event.
(Some typos corrected on both our parts)
First let’s take the discrete version of the EF, in which we do not have a function “f” or parameter gamma which we can optimize for “extremal set”.
In the Easter Island figures, the “specification” is matching the human form in some manner.
We know that the specification has to include “in some manner” because the Easter Island figures are not an exact match for the human form.
But now we recognize that the specification also applies to other rock formations. It may not fit as exactly as one might intuitively wish to consider, but that is in fact a subjective issue. Also there may be an issue of “completeness”. But completeness is only a matter of probability. Since there are several examples of the level of completeness of the stone faces, the probability may be relatively high of such a construction appearing even more complete by random processes. This is true even if there are actually no examples on the Earth, because the issue is of the probability, not of the actual existence of the event. So the probability of very rough figure matches is 1.0, since we have several. The probability of more complete figures appearing can be given as several orders of magnitude lower. I’m going to pull a number out of the air of 10^-20. This is significantly lower than the probability of 1.0 that we have observed of finding the more crude face figures, but is an intuitive estimate that I shall give without much more justification at this time. I am only giving this probability because I want to avoid an argument that the examples given of stone figures would need a small degree of more completeness to actually pass the “toddler test”, for example.
Now the key: The EF uses a typical alpha cutoff of 10^-150. This probability given of 10^-20 is off by a factor of 10^130. One would have to show a lower probability of say 10^130 lower than given to actually “infer design” by matching the condition of “complexity”.
We have three conditions, specification, complexity, and independence. We must meet all three conditions to “infer design” according to the EF. But remember that the probability used for figuring “complexity” is that of matching our “specification” in a stone figure by random non-intelligent geological processes. Most importantly, this is not the probability of the structure we observe on Easter Island being configured as we see it! It is the probability of meeting the specification, not the probability of what we observe on Easter Island!
Why is this such an important aspect of the EF? Why is it the probability of meeting the “specification” (which must be demonstrated to be given “independent of the event”) that must be measured, and not the complexity or probability of what we see?
The reason is simple. Consider a coin toss of 500 coins. That particular pattern of heads and tails – taken as a sequence of bits – would only occur with a probability of less than 10^-150. Yet if we look at that sequence, it would be what we see. It is “complex” in that sense. And if we simply say what we saw, we would have “specified” it. What failed is the third element—being “independent of the event”. We are not allowed to pull in aspects of what we see just because they are present. If we were allowed to do so, we could easily compose a “fabrication!” That which we consider must simultaneously be part of the “specification”.
The specification, that which is being measured for probability, must be the part that is given independent of the event. It cannot simply be what we see in the situation. If you see more than is captured in the “specification”, and think that one can lower the probability by rewriting the “specification” then by all means do so. But write it down, give the new specification! Because this new specification must be “independent of the event”!
So in fact we do not “infer design” using the EF, with this example, given any specification that has been shown here.
So this is an example of ID enthusiasts holding out the Easter Island figures as though they would give a positive result in the EF. So far this has not been demonstrated, and to claim so is to misrepresent the situation.
Now what is wrong—we humans clearly look at the Easter Island figures and just intuitively know that they were designed, probably by humans. Where is the problem?
Alonso also said in that other thread:
quote:
Remember that Dembski's math simply reflects how we infer design from everyday life.
This is what is wrong. ID enthusiasts actually believe this. But it is not the case.
I propose that we recognize the similarity of the Easter Island figures to artworks that we have created ourselves and seen others create. Read back over this thread for more information on this aspect—I’m not going to repeat that argument which took pages.
The key point is that we have greater knowledge of how such figures are created, and we mentally and immediately imagine a scenario of its creation. We even can immediately judge some “motive” for such a creation—done by a human. (There being various desires for wanting to have representation of the human form.)
Now without doing actual mathematics, we can easily imagine a scenario of human art work. This is an immediate “Gestalt”. It is not a carefully worked out piece by piece argument, it comes almost whole to mind when we see such a case.
And that is just the point. We don’t do an “eliminative” consideration of whether a natural process could cause that. In fact any consideration of natural processes comes substantially after the first recognition. For example in the stone faces the first thought may indeed be of representation there as well. But then and only then do we start to get into details, like how would a human create such a figure, and how would natural processes also be likely to create such appearances. By considering how we could search for a long time and then finally find another pattern that matches our specification we realize that this could be of high probability by natural process. And then we would reject it as caused by the human, because we know that our comparison of hypotheses does not fare well. But we may do this intuitively, rather than completely working out all the numbers. To actually work out the numbers would be expected to confirm the notion.
This is the key. Dr. Dembski’s EF is not how we “infer design from everyday life”. Rather we do intuitive comparison to the “best explanation”, even if we are not so careful, complete, and scientific as might be desired.
But let’s continue the discussion of details.
As has been presented before, as we look at the Easter Island figures, we continue to mount a list of aspects that compare favorably to human construction, but would significantly reduce any probability of random construction. As we grow this list of human-construction characteristics, the probability calculation of random construction would grow lower and lower by multiplicative powers. In other words we can easily reach a probability of construction in all the details we are considering of lower than 10^-150, for example.
So now with all this additional information—don’t we now have a successful application of the EF?
The answer is a resounding NO! We have no more done so than if we accumulated probabilities by multiplying the steps of the coin flip, noting each aspect in succession.
Where is the difference? The difference is in the “specification” being given “independent of the event”! In our comparative approach, we do not have to have a “specification” for comparison of the natural process that is “independent of the event”! If we have a hypothesis of the specific event characteristics, we can use those for comparison.
And furthermore we can loosen our threshold on the natural process. If we know that the probability is high of human construction, for example, of something which matches the observation, then we only need to compare to a level of probability derived from that, not to the level so low as 10^-150 or so that the EF requires.
This is very similar to the steps in the EF, in which we had to use an independently gained characteristic and calculate its probability by natural process. Just now we are not limited by the “independence” constraint—because we can use any factor form the specific intelligent agent hypothesis. We can do so because we are using details of our consideration of the intelligent agent, rather than a total abstraction of intelligence. Rather than rely on the entire world of experience of intelligent agents, we can focus on specific characteristics that we know of particular types of agent activity. Precisely the type of activity that we hypothesize for our event! Precisely the activity of humans creating a representational form out of stone, with a “motive” to impress with their size and expression! And we can form conjunctions of those aspects—something that we cannot do if we require that the specification to be given “independent of the event”.
[ 24. September 2003, 09:55: Message edited by: gedanken ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 24. September 2003 01:46
Good points!
Many of the points you make about specification are also made by Dembski in TDI and/or NFL. Unfortunately, many ID proponents seem to have forgotten some of them. (Including, strangely, Dembski himself, at times.)
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 24. September 2003 20:58
Ged writes:
quote:
So this is an example of ID enthusiasts holding out the Easter Island figures as though they would give a positive result in the EF. So far this has not been demonstrated, and to claim so is to misrepresent the situation.
Before I fully respond to your post, since we both agree that a specification exists, I would like to see a demonstration that the specification does not exist in a low probability state. I think that to claim that it does not, not only flies in the face of common sense (how we infer design in everyday life) but is disingenous without a demonstration.
I am arguing that the Easter Island statues actually would yield a positive result (design) when it runs through the EF.
Just as a bullet point summary of the position I will argue in this thread:
It is completely false to say that the designer is an "unknown". It is hypothesized the designer is an advanced human-like intelligence.
I still have not even an inkling of an idea how motive is at all relevant to this discussion.
Ged's "unknown" non-intelligent cause is nothing but an unfalsifiable promissory note which does not detract from any design inference.
I'm not sure how much time I will have in the next couple of days (something very major came up), but I'll try not to leave this thread hanging for too long.
[ 24. September 2003, 21:09: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 24. September 2003 22:45
Alonso,
For the “specification” of an appearance of a human form, we accept that the stone face appears to be a match to a human form, and that it was randomly created. The chances of random forms matching “human form” is therefore about 1.0. That is not very low.
Now I am sure that neither you nor I think that is appropriate description of the Easter Island figures.
The problem is that the “specification” must be found “independent of the event”. Take the arrow shot at the wall example, and draw a bullseye around it. We could “fabricate” the appearance of low probability of so-called ‘specification’ by describing the event after its occurrence.
Now an obvious thing to do in the Easter Island figures is to start describing the features of the figures in detail, and noting that those aspects would have low probability each. Then furthermore note that the product of those low probability events alone would multiply to a very low probability. And of course we could probably show that with some degree of accuracy that it was a very low number, since by multiplication we can keep adding features until we reach 10^-150.
But that would be a problem of the “bullseye”. We could just as well start describing the positions of grains of sand in detail (as I just discussed describing the features of the stone figures) and those would multiply to the low value. So that is not sufficient.
So now lets talk about a requirement that there is some existing, independent statement of the given aspect. (The idea being that all “aspects” taken in whole make the entire “specification”, but that individual aspects are discussed separately.) Take another case, a coin flip. Once again start accumulating individual bits of information. First we see a “head”, 0.5 probability. But we have seen “heads” specified before—so we use that as a specification. Then next coin (it was either head or tail). Obviously we have seen that before independently of the event, so we specify that. After 500 in a row, we have 500 details accumulated in the sequential 500 coin flips, each of which was a detail that has been given somewhere separately of the event. Problem—the 500 coin flips had less than 10^-150 probability and yet we described them with descriptions that have been seen before independently. The problem is of course that we have not seen them described simultaneously and independently!.
What I am showing is that one cannot simply take the features that one sees and expect to derive a specification that is “independent of the event”. One must in fact find that specification—as the conjunction of all its aspects—given “independently of the event”.
With that background, now I ask the same question I seem to find valuable in many threads:
What is the specification?
(By that I mean the specific specification that we want to take as having “low probability”. I’ve already given a high probability specification, that one was not good enough, so I need a better specification. Given that we can proceed.)
[ 24. September 2003, 23:22: Message edited by: gedanken ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 25. September 2003 10:47
Rex, I need help! (And no comments from the peanut gallery, please.)
I'm trying to figure out exactly how the GCE procedure eliminates "fabrications."
e.g. on NFL P.72, how does P(E|Hi&K) = P(E|Hi) eliminate a fabrication of the background knowledge K simply being a writing out of the pattern of 500 sequential coin flips after they occurred?
This is subtle, as we have varying time frames of information involved.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 25. September 2003 17:17
Ged writes:
quote:
We know that the specification has to include “in some manner” because the Easter Island figures are not an exact match for the human form.But now we recognize that the specification also applies to other rock formations.
This is not true actually and shows a misinterpretation of what specified complexity is. Here is Dembski:
quote:
Given an event which is described by a pattern, which is one of a range of possible events; a pattern is detachable from the event if it can be formulated without knowing which event actually occurred.
I can describe for you the elaborate shapes of eyes (complete with pupils), mouths, and even a hat without having ever looked at the easter island statues. That the shapes are more exaggerated then ours is irrelevant. The intracies of the shapes are imbedded in the complexity of the event. On the other hand, I have to actually see the rock formation that shows an ambiguous human form in order to know what exactly the specification is, it is post-hoc. Thus the latter would most likely be a fabrication, or a simple form of specification.
Thus I would say that the EF would certainly yield a positive result for the easter island statues. This is further corroborated by the toddler test that Mike Gene gives, in which the child refers to one picture as "a rock" and the other as a "man rock".
Ged writes:
quote:
It may not fit as exactly as one might intuitively wish to consider, but that is in fact a subjective issue. Also there may be an issue of “completeness”. But completeness is only a matter of probability. Since there are several examples of the level of completeness of the stone faces, the probability may be relatively high of such a construction appearing even more complete by random processes.
No, that a random process would cause something like the easter island statue is quite outlandish. This is why the statues themselves represent a complex event.
Ged writes:
quote:
This is true even if there are actually no examples on the Earth, because the issue is of the probability, not of the actual existence of the event. So the probability of very rough figure matches is 1.0, since we have several.
This makes no sense. How can something not exist, but at the same time, we have several?
If you are talking about ambiguous human forms in several rock formations, the fact that some may see several is irrelevant to the probability of a highly complex specified event like the easter island statues.
Ged writes:
Ged writes:
quote:
We have three conditions, specification, complexity, and independence.
This is already wrong from the get-go. There are only two conditions. The independant pattern is a specification.
Ged writes:
quote:
Most importantly, this is not the probability of the structure we observe on Easter Island being configured as we see it! It is the probability of meeting the specification, not the probability of what we observe on Easter Island!
Oddly, you have just said the same thing twice. Calculating the probability of the structure on Easter Island is precisely calculating the specification that we attribute to the pattern we see.
Ged writes:
quote:
The reason is simple. Consider a coin toss of 500 coins. That particular pattern of heads and tails – taken as a sequence of bits – would only occur with a probability of less than 10^-150. Yet if we look at that sequence, it would be what we see. It is “complex” in that sense.
Complex in what sense? In that it's just what we see?
Ged writes:
quote:
And if we simply say what we saw, we would have “specified” it.
Thats called a fabrication.
Ged writes:
quote:
I propose that we recognize the similarity of the Easter Island figures to artworks that we have created ourselves and seen others create. Read back over this thread for more information on this aspect—I’m not going to repeat that argument which took pages.
There was no argument that addressed the fact that the design inference basically reflects what how we usually infer design in everyday life anywhere in this thread. Note, the easter island statues is similar to how humans actually look like and dress and not to artwork, but thats irrelevant.
Ged writes:
quote:
And that is just the point. We don’t do an “eliminative” consideration of whether a natural process could cause that.
I will now present an example of how we do the exact opposite of what Ged suggests in the above quote which will show how we actually do an "eliminative" consideration when inferring design.
Most people will recall the Mars Face fiasco a while back. Pictures of the surface of Mars caught photos of what looked like a human face structure, sort of like the Sphinx.
Here is how one article described the inference to design by many people:
quote:
The feature's resemblance to a human face is intriguing, and many people see in that resemblance not an accident or trick of light, but rather the work of intelligent extraterrestrials.
source
Because of the face's apparent strong resemblance to a human face, they eliminated accident and/or trick of light.
This gets even more interesting and relevant to the design inference. When photographs were taken 10 times sharper than the original image, what was revealed?
A natural landform. Thus, now they conclude it was not designed, but why? Because there is no specification, no complexity, because a natural geological processes can easily produce landform. As you can see, I went through every single step in the EF in describing this event, showing how the EF maps to how we usually infer design in everyday life.
[ 24. February 2006, 16:24: Message edited by: Nel ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 25. September 2003 17:35
quote:
quote:
We have three conditions, specification, complexity, and independence.
This is already wrong from the get-go. There are only two conditions. The independant pattern is a specification.
I feel that this thread is very readable and has very well developed arguments. The readers that want to read it in detail will understand. Those who only want to try to confuse issues will have to make their own way.
With regard to the above quote, we have clearly that a pattern that is "given independent of the event" can be a specification, as per Dr. Dembski. Thus inherent in "specification" is that it must be "independent". Thus these are three requirements of importance, the pattern (which will be referred to as "specification"), the "complexity" in which is defined as low probability, and the "independence" issue. These are three issues, even if "independence" is subsumed into the most technical definition of "specification". There was nothing of substance to the argument wrong in that statement.
But I don't see a "specification" written out for our examination, so that we can check that it is found separately of the event, that we can simultaneously check for the the three aspects. This time I will use more exact Dembski terminology, which will not change the argument one whit. That is where is the "pattern" which simultaneously has "independence of the event" and which has "low probability".
In fact the above post, with all its length, does not actually give a "specification" that we can verify for all the important components.
So I will ask again:
Please give a complete and unambiguous specification for the Easter Island figures.
Make all the arguments you wish, but clearly delimit what you are considering to be the "specification".
quote:
Oddly, you have just said the same thing twice. Calculating the probability of the structure on Easter Island is precisely calculating the specification that we attribute to the pattern we see.
This statement makes my point that the pattern does not appear to have been given independent of the event. This is a key point. Failing to understand this point is why ID fails to be logical. The EF requires setting up of a rejection region, which is then tested in general to be met, not a calculation that includes aspects of the event E that are not part of the "specification". The distinction between what is in the "specification" and what is in the event E is critical--aspects of the event E not in the "specification" cannot be composed into the "complexity" probability calculation or else one is creating a fabrication.
On answers to other questions or points, simply read through the thread.
[ 25. September 2003, 18:13: Message edited by: gedanken ]
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 25. September 2003 17:36
Ged writes:
quote:
e.g. on NFL P.72, how does P(E|Hi&K) = P(E|Hi) eliminate a fabrication of the background knowledge K simply being a writing out of the pattern of 500 sequential coin flips after they occurred?
I don't have the book on me at the moment, otherwise I would help you more (I'll add to this post tomorrow). By showing that one must specify the event post-hocly is good enough. Who is the "peanut gallery"?
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 25. September 2003 17:42
Ged writes:
quote:
With regard to the above quote, we have clearly that a pattern that is "given independent of the event" is a specification, as per Dr. Dembski. Thus inherent in "specification" is that it must be "independent".
No, the specification is the independant pattern. They are not two seperate things.
Ged writes:
quote:
But I don't see a "specification" written out for our examination, so that we can check that it is found separately of the event, that we can simultaneously check for the the three aspects.
This is probably due to the fact that you are not very familiar with the EF. You can seperate the pattern of nose, mouth, eyes, hat, etc from the probability that the entire formation arose by chance, pointing to the shapes which cannot be accounted for by chance or regularity of natural laws.
Ged writes:
quote:
Please give a complete and unambiguous specification for the Easter Island figures.
The pattern of nose, mouth, eyes (even the pupils), hat is independant of the event (in fact I think I already gave this unambiguous specification).
Ged added in edit:
quote:
This statement makes my point that the pattern does not appear to have been given independent of the event.
I can easily give the pattern independant of the event. I have seen the intricate shapes of noses, eyes, mouth, etc way before I saw the easter island statues.
Ged writes:
quote:
This is a key point. Failing to understand this point is why ID fails to be logical. The EF requires setting up of a rejection region, which is then tested in general to be met, not a calculation that includes aspects of the event E that are not part of the "specificaiton".
But that is what makes the design inference reliable. From No Free Lunch:
quote:
S identifies a rejection function f and therewith a rejection region R that includes E and that is an extremal set of f. R is therefore of the form T^gamma = {w in omega | f(w) >= gamma}....or delta...where gamma and delta are real numbers. Typically gamma is chosen as large as possible and delta as small as possible so that
Ged writes:
quote:
On answers to other questions or points, simply read through the thread.
I just read through the thread, and there are no answers to any of my points that can be found. However, the most important question, who is the peanut gallery? Does motive imply intelligence?
[ 25. September 2003, 18:05: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 25. September 2003 17:48
quote:
The pattern of nose, mouth, eyes (even the pupils), hat is independant of the event (in fact I think I already gave this unambiguous specification).
Yes, once again, this is a high probability event, as Alonso has demonstrated for us. Remember it is not the pattern of the specific Easter Island figures, rather it is the pattern in the "specification" sentence given. Read the sentence with no knowledge of Easter Island figures (e.g. forgetting, not looking) and then find the probability of finding that event.
(Alonso, if that "specificaiton" was not sufficient, propose another.)
I am satisfied with my presentation. Readers will have to read it over to understand the details.
[ 25. September 2003, 17:57: Message edited by: gedanken ]
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 25. September 2003 18:01
Ged writes:
quote:
Yes, once again, this is a high probability event, as Alonso has demonstrated for us.
Actually the pattern of nose, mouth, eyes (even the pupil), hat is a low probability event. That is why when those characteristics are found, design is inferred.
Ged writes:
quote:
Remember it is not the pattern of the specific Easter Island figures, rather it is the pattern in the "specification" sentence given.
Which is the same thing. The pattern must match the event observed.
Ged writes:
quote:
Read the sentence with no knowledge of Easter Island figures (e.g. forgetting, not looking) and then find the probability of finding that event.
Since the event includes intricate shapes that match a human face, of course the probability will be low.
Ged writes:
quote:
(Alonso, if that "specificaiton" was not sufficient, propose another.)
No, my specification was more than sufficient.
[ 25. September 2003, 18:02: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
|
Printer-friendly view of this topic