|
Author
|
Topic: Does intelligence imply “motive”?
|
gedanken
Member
Member # 594
|
posted 07. November 2003 11:29
Also, if you read my first few posts, the "motive" aspect is taken as one of many aspects of the "designer" which may be important. This thread is about the importance of considering such criterion as motive in judging the cause of events. If in a particular case "motive" per se is not useful, but other characteristics of the designing system were of importance, that would be as relevant to this topic as the "motive" aspect per se. (I am not prejudging what the reader may consider a "designer".) But for human "designers", about which many examples used apply, the term "motive" seems useful as a predictor of behavior, and relevant to examples given.
(By the way, the thread title is somewhat off the mark of the real subject matter, the title is not representative of the questions of the thread.)
Furthermore I would point out again that "motive" is not sufficient to establish a joint prior probability of "design" cause of event. But it is precisely establishing of joint prior probability of "design" cause that is needed to find out if the EF (or its variants) will be reliable in a given case--wherein "reliable" means that the answer given will stand up in the future.
But consider the "Caputo" case often given as an example of application of the explanatory filter. Here we clearly know of what we would consider "motive", as well as means and opportunity for Caputo to have effected the events. Now if one were to consider "Captuo" (a person) to himself be some sort of automaton, with "motive" taken as a descriptive only aspect and not precise in formal theory of "Caputo" as automaton--then I guess the point is meaningful. But I don't think that such distinctions are necessary. But then in such a "formal" rendering of the situation, one could consider whether state transitions of the 'designer' "Caputo" could lead to state transitions of the event, and probabilities thereof. It does not matter if the state transition probabilities of the agent are called "motive" or not, rather it matters their probability and their linkages to events by connecting state systems.
--
By the way, the idea that a test can be "reliable" (in sense of no false positives), yet can be flipped on its head with the tiniest bit of information gained in the future, seems odd. But it often takes only the tiniest "missed distribution" to upset the answer given by the EF in many cases.
Dr. Dembski, (from quote previous page) in his forthcoming book, says:
quote:
Contrary to Wilkins and Elsberry, the risk of further knowledge upsetting a design inference has nothing to do with the filter’s reliability. The filter’s reliability refers to its accuracy in detecting design provided we have accurately assessed the probabilities in question (see chapter 12). Wilkins and Elsberry purport to criticize the filter’s reliability but are in fact criticizing its applicability (see chapter 14). They’re like someone who dismisses a calculator as unreliable after a friend, seeking to know what “9 times 9” is, gets the wrong answer by accidentally punching “6 times 6.” If that person were dead-set on dismissing the calculator’s usefulness but were pressed to admit that the friend had made the error, then the calculator hater might insist that nobody can be trusted to use the calculator accurately. That’s essentially what Wilkins and Elsberry have done with the Explanatory Filter.
So consider a calculator with tiny or illegibly labeled buttons. Since I am involved with product "design", I see this sort of excuse given quite often--for example a computer program is so hard to use that one often fails to get correct performance, but the programmers blame the users for its lack of "reliability".
[ 07. November 2003, 12:35: Message edited by: gedanken ]
IP: Logged
|
|
Danpech
Member
Member # 163
|
posted 08. November 2003 00:24
We each have a natural EF (NEF). If the NEF is only probablistic (i.e., fallable), then is there any sense in saying that a person misapplies the NEF? If we can sensibly say that a person misjudges an object as designed, then are we admitting that, only were the person sufficiently sensible, he would never misjudge any non-designed object as designed? Is a person's judgment conceivably superior to his own NEF?
Can we say, without doubt, that the NEF is merely probablistic for the 'big' design questions?
What is the logical relationship between the 'little' design questions (both the fairly easily determined problems like in forensics, and the nearly-impossible-to-determine problems like whether a specific random rock on yonder hill was placed there by a human) and the 'big' design questions?
Was the NEF designed? If not, then what? Or, did humans design it?
If the NEF can be misapplied, then does a correct application eliminate false identification? Surely, one can simply fail to use the NEF for inferring design, simply by 'willful disregard' for the idea that the object in question is/was designed.
Of course, one can, by ignorance, make a formal version that is not quite true to the NEF. But, if the NEF can be misapplied, then so can any true formal version of it.
[ 08. November 2003, 00:39: Message edited by: Danpech ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 08. November 2003 03:36
Denpech, we have 'natural' answers to many questions, and thus 'natural' procedures for answering those questions.
For each of them, we could say that if only we got all the issues correct, we would not have made any error in answering our question. So?
But the EF (that is the formal EF in various forms presented in Dr. Dembski's books) does not provide a theory of how design occurrs in the instance given. In fact it explicitly eschews such description. It does so because it is formally "eliminative". (In other words users may 'naturally' add in thier own judgments and ways of estimating aspects of this, but the procedure is strictly defined in formal terms as step by step procedure. But there are many unclear definitions which harm the objectivity--but that is another subject, let's assume for the moment they are resolvable.)
Since the EF (in its formal presentation) ignores the structure of the designer interaction completely, the EF is not a theory of the world. It is only a test. Tests differ from theories of operation in that a test is not a description of how the activity in question operates in some detailed manner. Since the EF eschews knowledge of the details of how the design occurs, it cannot be a description of how the activity of design occurred.
But as such, as a "test", we then have very little we can do to judge it. We cannot use predictions made by the "test" to understand structural aspects of design--precisely because there are none. All that can be done is to evaluate the "test" in terms of its usefulness in determining the very answer that it gives--precisely because that is all it gives, and it gives no additional information other than the test 'answer'.
So the only way to evaluate the EF (or variants) is to see how well it works in practice, precisely at providing the answer that it supposedly provides. Thus we can, for example, test how well EF answers hold up over time. How useful is the EF procedure at predicting future evaluation of the "design" aspect?
Now it turns out that because of the "ProbRes" factors, the EF is incredibly sensitive to missing information. It is much more sensitive to such missing information than explanation based theories of events, because the very factors that are used to control for coincidence also are multipliers of the probability from "missed distributions". That makes the EF especially unreliable--more so than methods in which the structure of how events occurred are evaluated. And that is because evaluating the structure of how events occur allows for avoiding the multiplier effects necessary to control for coincidence.
IP: Logged
|
|
Ermete22
Member
Member # 970
|
posted 08. November 2003 06:13
Waht does IMO mean?
Of course you are right from your point of view. I read it with a lot of attention.
On this very interesting subject the term "point of view" has a very deep meaning,
which is interesting to discuss.
You can ask different questions, which raises different problems:
My simple humble question looks like this "how can I attribute an intellligence to a system I am
observing?"Being a theoretical physicist (actually a repented one for many
phylosophical reasons), I keep the data as they are, that is a possibly long string of observed visible actions.
Of course there exist a lot of technical tools to analyse long string of data. In particular
we can consider such a set from two points of view, a static one, which corresponds to the set of
alecited global actions by the system, something like the collection of the pictures of the
elicited global postures of the system. This is a static view of the repertoire of the observed actions.
Here the only reasonable problem is "is this a language, in the formal sense, grammar, sintax..."
As you see, this is already a linguistic problem, and my theory is necessarily a linguistic one, where
"motivation" sits in its own right". So you are right about linguistic theories of possible intelligence of the system.
Making things simple (and hoping I'm correct, as the subject is complex),I would agree with
your statement if it referred to the theories about intelligence. We need here a
theorem, or principle, of which I'm not aware, stating that
"the motivational concepts, useful and legally sitting in a theory
of intelligence emerging from the observer's data collection, implies that motivations are
still a meaningful concept when we intend discussing the internal computations by
which the system elicits its actions". Of course such a theorem does not exists. This is after me
not such a definite instrument of critic to your point of view. Maybe someone, you maybe,
will be able to demonstrate it. In any case there are a lot of good ideas which are not theorems.
I simply find it a disturbing absence.
The other point you raise is about the nature and meaning of fitness functions in
genetic search algorithms. This is a very subtle subject. I frankly do not agree on your statement “All GA algorithms are based on goals or motive” It seems to me that GA more simply, invent viable effective computations, which can affect the system’s behaviour embedded in an environment. Such new effective computations can affect or not the behaviour of the system, and, consequently its successful permanence, as a species, in such an environment. In most cases , the vast majority I suspect, the newly introduced effective procedures will produce very strange, negative effects. I do not intend being being a bothering pseudo-philosopher, but simply underlie that GA cannot know a-priori what’s good or bad for their owner. They cannot be based therefore on a goal or a motive, also for the very simple reason that they have no memory bits where storing such motives. I agree with you on the contrary, that, when planning the use of a GA tool to solve a problem, you insert the motive within the fitness function you choose, and that such a fitness function stays for your theory about the system under analysis. Then you are perfectly correct in this case by stating: “ All GA algorithms are based on goals or motive”.
It seems to me that what emerges from the comparison between your point of view and mine is a very interesting question: “What are the critical conceptual problems when one compares theoretical implementations of AI and the attribution of intelligence to autonomous existing (in general living) systems? Can these problems give rise to s new, more enlighten, synthesis?
Carlo
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 08. November 2003 09:24
Carlo,
IMO= in my opinion, IMHO= in my humble opinion
Quote: My simple humble question looks like this "how can I attribute an intellligence to a system I am observing?"
I am undoubtedly in the minority, but I believe this question has a scientific answer. The answer starts with the relationship- "design implies design processes which implies an entity (or entities) containing the design processes’.
It is possible to develop objective mathematical standards for differentiating design from non-design. There may not be any one single criteria that can draw an absolute line between all instances of design and all instances of non-design, but there are clear objective standards which can be used in most instances distinguish between designed objects and processes and non-designed objects and processes. If you are interested, we can discuss some simple objective mathematical criteria for distinguishing design/non-design.
If we start with designed objects, the next step is to consider design processes. For the discussion here consider 1)a snowflake, 2)a birds wing, and 3)an airplane wing. It is easily demonstrated that all three objects exhibit mathematical properties that would qualify them as designed.
For the snowflake, we know, so I am told, the physical process responsible for generating observed patterns or designs. We know the design is the result of a permanent natural process. We would generally agree snowflake design process is not intelligent. Since the snowflake design is the result of interactions between the snowflake and its environment, we can say the design process is contained in the snowflake. Since we would agree that the snowflake design process is not intelligent, we would conclude that the snowflake is not intelligent.
For the airplane wing, we know that in most instances the process responsible for generating the observed design was a human design process. We would generally agree such processes are intelligent. Furthermore, we would generally agree that significant portions of the design process occurred inside the human head or design. Therefore we would conclude that humans are intelligent objects. [Humans are essentially the benchmark or operational definition of intelligent objects.]
This brings us to the most ambiguous case, the wing of a bird. Using the same types of objective mathematical criteria used to determine that snowflakes and airplane wings are designed, we should be able to agree that birds wings are designed.
The next issues are to determine 1)what type of design process produced the bird wing and 2)do the design processes qualify as intelligent. I propose as a concept or criteria, that a design process is intelligent if it is purposeful or goal-oriented. I can add other criteria to the definition of an intelligent design process such as ‘a design process is intelligent if it is creative, progressive and dynamic’, but the existence of a purpose, goal, or function is the probably the most easily identified or defining characteristic of an intelligent design process. Wings of both airplanes and birds serve the function or goal or purpose of flying.
Using the definition of intelligent design as purposeful or goal-oriented design, the wing design is a purposeful design process performed by living systems. All living systems perform this type of goal-oriented design therefore all living systems are intelligent objects. Except for man made machines, there do not appear to be any non-biological examples of creative, progressive goal-oriented design processes.
In answer to your question, I would thus suggest that using the ‘progressive, creative, goal-oriented design process criteria, you can recognize an object as intelligent if it is alive, or if it is man made and satisfies certain criteria.
Note that in a GA or any type of search engine the criteria ‘find the optimal design’ or ‘accept design A or B if A has a higher fitness value’ define goals. A system will wander aimlessly unless it has such a goal or purpose. This is true of all AI systems.
Note that if you equate ‘motive’ and goal or purpose, then intelligence implies motive. If you view ‘motive’ as involving emotion or consciousness, then motive is only associated with certain limited types of intelligence.
Quote: "What are the critical conceptual problems when one compares theoretical implementations of AI and the attribution of intelligence to autonomous existing (in general living) systems?
IMO, the critical conceptual problem is the issue of causation. If you accept as a legitimate part of science, hypotheses based on the Aristotelian concept of teleological causation, then you can construct testable predictive scientific hypotheses of intelligent behavior and you can reproduce such causal relationships and intelligent behavior on a computer. This is in practice the conceptual solution used in all successful AI simulations. IMO, an interesting example of where practice is far ahead of theory.
[ 08. November 2003, 09:28: Message edited by: warren_bergerson ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 08. November 2003 18:50
A key point I wanted to establish was that motive could be useful in establishing the particulars of the details of a specific case. While only one of a number of issues, it is a recognition of importance of using more than an eliminative inference method. But motive is not sufficient--one needs sufficient theory of events to reduce the effects of coincidence on the inference method, and one needs sufficiently complete information on the intelligent agency scenario to constrain prior probabilities and likelihoods.
[ 08. November 2003, 18:55: Message edited by: gedanken ]
IP: Logged
|
|
Ermete22
Member
Member # 970
|
posted 09. November 2003 04:22
IMO (I like it!) the following is not completely wrong:
Design-non design differentiation by maths is for sure something of great interest, and cannot but attract my curiosity. Let me first apologise for being an academician: it’s not a joke as I’m very well aware of the sins of my category. So I apologise in advance.
A snowflake is the result of the interaction between the sticking properties of , say molecules and the dynamical characteristics of the environment. During its growth the screening effect exerted by the most exterior portion of the snowflake is the responsible for many characteristics of the final shape, together with the available symmetry of crystallisation of H2O. The features of the outside molecules flux is important as well.
To make it simple, snowflakes are the results of:
1) environmental conditions
2) symmetries associated with their elementary constituents.
An observer could initially suspect some responsible for their attracting shapes, but, unless looking for esoteric responsible agent, would conclude, as you suggest, that no design process is needed to get snowflakes.
Avoiding a whole class of dangerous (and boring) “distinguo” , I agree that plane wings are evidently the result of a human project. Some man has conceived them.
The rub, as you anticipate, comes with birds wings.
IMO we cannot even imagine a designer. At least in the sense we quietly used for the planes wings. We could consider human breeding birds with particular final targets, and making use of genetic transmission of certain characteristics. That scenario does not differ so much from the above except for the techniques adopted.. Bioengineers are selling as feasible such kind of “intelligent design”, without actually having the tools to do that, at least for the moment.
Ignoring birds breeders and selling biotechnologists (which is good at least for health), let’s go back to your message.
IMHAO lets restate the whole affair:
1) birds wings are real physical objects
2) negating that they represent efficient tools for flying would be simply stupid
3) the evolutionary story of birds wings can be quite well reconstructed, in many details
4) birds wings are the results of such evolution.
5) Penguins and ostriches are birds
Whatever bird species has now reached an average, recognizable, wings shape (zoology). All bird species evolved from reptiles in a very long time period.
I DISAGREE WITH YOU IF YOU MEAN THAT THE VERY FIRST GENETIC CHANGES, WHICH CAN BE SEEN NOW, AS ANTICIPATING THE FUTURE EVOLUTION OF BOTH BIRDS AND BIRDS’ WINGS CONTAINED IN ANY SENSE ANY PROJECT. On the other hand such changes are necessary elements for reaching the nowadays situation. Seeing backward the whole genetic story NOW, you can SEE, as an observer, a project. But what to say to about the vast majority of unregistered (and unregistrable) evolutionary failures?
Without willing incorporate morality and economy in this analysis, a project should, in a normal conception of the term, have been able to bypass a decent percentage of failures obtained by randomly chosen genetic modification. But this is non very honest to be stated here.
Systematic random exploration, will finally, but not necessarily produce something that works well, but to adopt the term “project” or worse “intelligent project” to what emerged from it seems to me an excessive semantic move at minimum.
Evolution plays by physical, chemical laws, a more complex game than that plaid by water and turbulence in snowflakes creation. Evolution’s stories are much longer (in time I mean), but, as you use saying in USA, the name of the game is the same. What exists now is, tautologically what has successfully evolved respecting the rules.
Your statement: “the existence of a purpose, goal, or function is the probably the most easily identified or defining characteristic of an intelligent design process” I intend as “Purposes, goals, or functions can sit in linguistic theories about systems. They are epistemic concepts in their own right, which can be used in the limit of their nature. They do not grasp the underlying dynamics, particularly when referring to an overwhelming, complex, non linear dynamics as intelligence, not because they are per-se wrong tools.
Carlo
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 09. November 2003 07:57
Carlo,
Quote: The rub, as you anticipate, comes with birds wings.
IMO we cannot even imagine a designer. At least in the sense we quietly used for the planes wings.
One of the important issues in discussing design and intelligence is the question of location. As the example of the airplane wing shows, the ‘design’ and ‘design process’ are not necessarily located in the same place. As the snowflake example illustrates, a design process is not always clearly located inside a single object. The term designer as I am using it here, refers to the location of the design process. Since the concept of ‘location of a design process’ is or can be ambiguous, the concept of designer will also be ambiguous.
Note that if we look at the design process responsible for an airplane wing, the location of the process is not nearly as clear as it at first appears. Complex design processes such as those involved with airplane wings involve directly or indirectly many different individuals. The final wing design developed by a particular set of engineers is ultimately the end product of the work of thousands and even millions of individuals whose work may span centuries. As with the snowflake, the ‘human’ design process also involves the use of a variety of processes and devices located outside the human.
IMO, it is generally confusing and misleading to suggest that design processes are something occurring in some type of self contained system called a designer. It is, IMO, more productive to complex design processes as being contained in large systems involving many interacting components operating over long periods of time. The primary question, IMO, is the relationship between design and design process. The size, shape, and location of the design process is a secondary issue.
Quote: I DISAGREE WITH YOU IF YOU MEAN THAT THE VERY FIRST GENETIC CHANGES, WHICH CAN BE SEEN NOW, AS ANTICIPATING THE FUTURE EVOLUTION OF BOTH BIRDS AND BIRDS’ WINGS CONTAINED IN ANY SENSE ANY PROJECT.
I find it interesting that this issue of pre-cognition keeps coming up because it was resolved a very long time ago. Purposeful, teleological or goal-oriented behavior does not involve pre-cognition. Going back to the days of Aristotle, so I have been told, it was recognized that a future event-a goal or purpose could not logically be the cause of ‘striving for a goal behavior’. Science and logic has (almost) never accepted the concept of causation working backwards in time. The concept of purposeful or goal-oriented causation used in science has never been based on precognition. In the behavioral sciences, goal-oriented or purposeful behavior is explained in terms not of future goals, but in terms of ‘goal expectations’. In GA systems, a goal is not a precognition of the wing to be designed, it is a rule of the form ‘select the option with the higher fitness value’.
In very general term, the goal in an evolutionary process is a rule of the sort ‘select the option with the greater expectation of survival’. Eliminate this goal or goal criteria and life forms would wander( and diverge) as aimlessly as a GA system without a goal or goal criteria.
If you accept the obvious conclusion that bird wings ‘exhibit design compatible with flying’, and the design process responsible for wing design involved a goal or goal criteria, then you accept that wings are a design produced by (as defined) an intelligent design process.
If we accept that evolution is an ‘intelligent design process’, then the question that remains to be addressed in ‘how complex and how powerful is the intelligence associated with the design process. Can the design of birds wings be explained by a very simplistic process of ‘random genetic variation and natural selection’ or is the process involved far more powerful and complex. Mathematical and AI analysis of designs and design processes clearly show that evolutionary change process are vastly more powerful and efficient than suggested by current so called evolutionary hypotheses.
Quote: Without willing incorporate morality and economy in this analysis, a project should, in a normal conception of the term, have been able to bypass a decent percentage of failures obtained by randomly chosen genetic modification.
Interestingly enough, it appears fairly easy to demonstrate mathematically that biological systems do in fact avoid a huge portion of potential failures. Biological variations are not random, but designed. As a simple example, when an offspring has a different overall size than its parents, the design factors are not randomly distributed, but designed or coordinated to be compatible with the changed body size.
Quote: They do not grasp the underlying dynamics, particularly when referring to an overwhelming, complex, non linear dynamics as intelligence, not because they are per-se wrong tools.
There is a fundamental choice in analyzing design and design processes. You can a priori insist that design and design processes must be reduced to physical-chemical processes as can be done with snowflakes. I call this the physical science orthodoxy. Alternatively, you can models and hypotheses based on purpose and teleology. I call this the engineering orthodoxy.
Many will and have argued that the physical science orthodoxy leads to a ‘true’ scientific understanding. Unfortunately, it also guarantees that the scientific analysis of design, intelligent design processes and topics relating to biological systems will be a failure(in the sense of not producing predictive hypotheses).
IP: Logged
|
|
Ermete22
Member
Member # 970
|
posted 09. November 2003 11:41
Dear Warren,
I beleive we are sitting on the opposite sides of something. meaning by that that there exists no profitable information exchange between us. Like between those who have a religious faith and those who do not share it. We fortunately are on the same side for what regards the curiosity about the world and the desire to understand it within our limits. And this is, in my opinion the most important point. I honestly cannot discuss your points as I literally do not see them. Being most of all a mathematician and a theoretical physics, I'm use to other more rigid (too rigid possibly) categories than the one you make use of. I do not mean in any sense you are wrong, and even less that I am right. Marbe you are saying truth, but a language I cannot understand. So best luck and I assure you, all my respect as, as I said we are sitting on the same side of curious part of the species.
Carlo
IP: Logged
|
|
brianjohnson
Member
Member # 979
|
posted 09. November 2003 17:57
i wish to pose a question for gedanken, regarding the early posts (pages 1 & 2)
you have stated that is not likely a criterion to be applied to recognize the easter island statues as designed that also rejects the old man of the mountain WITHOUT USING THE ACTUAL OBJECTS UNDER INSPECTION. i agree with you (for now), but then i ask, what does one infer from this.
i am thinking about the ability to discern design from non-design in general. isn't it appropriate to create a class of objects such that all objects of that class are only objects that have been designed by human beings for specific usage (i.e. tools yes, but not splatter-art). also, to create a class of non-human designed objects such that no objects of this class can also be in the other class.
AFTER we have established these classes, couldn't we then apply these distinctions to non-human artifacts for the sake of determining if there are a class of objects that more accurately resemble the first class and not the second? And from there, couldn't we infer that these objects are likely objects of deliberate design?
so, while we cannot create criteria to determine before hand if the easter island statues are really designed, it does seem that a post-facto, classifications are not an undesireable substitute.
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 09. November 2003 22:11
Since this thread is now in the process of wandering far from gedanken's original ideas, I don't feel too bad about interjecting a request. warren_bergerson wrote quote:
Mathematical and AI analysis of designs and design processes clearly show that evolutionary change process are vastly more powerful and efficient than suggested by current so called evolutionary hypotheses.
Are there references to these "mathematical and AI analyses" that one can look up and read? Absent that, it's not at all clear in what respect(s) "evolutionary change processes" (by which I assume warren_bergerson means processes operating in biological systems) are allegedly clearly shown to be more powerful and efficient than current evolutionary theory suggests, and thus it's impossible to independently verify the claim.
RBH
[ 09. November 2003, 22:12: Message edited by: RBH ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 09. November 2003 23:37
Please, all, on "wandering", I have a great deal of additional material coming that is directly on topic! I would really like to get this material in this thread. I am sorry that it will arrive so slowly, but I have many priorities ahead of this.
But the material (which builds on previous posts in many ways) is the product of over an elapsed year of research (only occiasional spurts of actual work) in the mathematical probability relationships of the EF.
So I appreciate if we can stay on topic without space filling, as I think the material to come will be quite important with regard to previous material. Already at 6 pages, with probably as many to go in the next months, I would deeply appreciate restraint.
Very specifically, the topic is not "does intelligence imply motive." Perhaps I should request the moderator to extend the title to:
Does intelligence imply "motive"? Considerations of reliability of and improvements to the Explanatory Filter and similar design inference methods.
(The problem is that is so long, but I would not want to change the first few words as there are several outside links that use the title. "Motive" is an example of non-eliminative characteristics of the "design" process or agency which can constrain probabilities thereof, and can thus inform on the reliability of the EF in given circumstances.)
Thanks
--
With regard to BrianJohnson's question:
(Yes, I think this is on topic.)
I don't completely understand, but I do use a basic concept of classifying events. (And I use "events" rather than objects because that is the standard ID terminology, though the potential construction of the object is usually the event in question.) We can use classification of events wherein different classifications might differ in terms of universal agreement as to their "intelligence". For example various processes may or may not be accepted as "intelligent" by different readers. As long as the definition being used is clear, all is well, but if the definition of "intelligent" is not clear one can have problems.
But the topic I discuss it does not rely on a specific definition of "intelligent", rather only on a clear means of making such a distinction. In fact any lack of clarity in making such distinctions only reduces the reliability of the EF or other 'design' inference, and cannot improve the reliability.
I do not quite understand the usefulness of defining "classes" of objects this way, unless one is refer to some sort of classification that could be given by an omniscient being and is used as definition for comparison to what is measured by inference procedures. (In other words an inference procedure cannot use such a classification of object in question as input, as the entire problem is to classify the object or event as output of the classification procedure based on observation, and that would be circular reasoning.)
Now if one is considering these as "target classes" (e.g. generic classes to which existing objects are to be judged as to membership, rather than assumed), then there are important aspects to consider. I actually believe that archaeologists do, for example, uses concepts vaguely like that. Certain objects give the appearance of belonging to a certain class of objects due to details of construction that would be noted in objects in that class. (Once again, question is whether this object is in the defined "class".)
What is important about that level of analysis is that it is not possible with the EF methodology. The reason is that method is eliminative and does not consider the details of construction by intelligent agents as part of the aspects that are analyzed for probability.
Brian if you read on to later pages this comes up, as the details of theory construction are important in comparative methods. But being comparative is not enough per se, rather one must have details to relate to construction and implementation method in some greater theory (or proposition at least) so that the case can be assessed without as many coincidences possible.
When one has theory of construction by the intelligent agent (including especially all of motive, means, and opportunity, for example) then the objects characteristic "specification" is not nearly as subject to coincidence. This is quite different from the EF method, which looks for a pattern completely without regard to any detailed construction scenarios by intelligent agencey of the particular case in question. All that is considred is non-intelligent process action, and the eliminative method explicitly eschews the consideration of the intelligent process details.
The mathematics that I will be presenting will show that the control for coincidence is key. Dr. Dembski's "ProbRes" factors M and N are part of what establishes a very low level for the significance factor 'alpha' in the GCE procedure. But these very same factors are multipliers of any "missing distributions"--for precisely the same reasons of coincidence probability as the reason for using the factors M and N in calculating the cutoff 'alpha' in the first place.
Thus without a theory of intelligent agent construction and implementation of the event in question, factors M and N are only constrained as calculated by Dr. Dembski so as to control for coincidences. (In other words the factors M and N multiplicatively estimate opportunities for coincidences, which must be controlled for by lowering 'alpha'.) But theories of intelligent agent action in the specific case in question allow for reduction of M and N factors. This can only be done by considering the specifics of the given case with regard to considerations of scenarios of intelligent agent action in the detailed case in question. This is not possible with eliminative method.
Comparison per se (as is shown by Dr. Dembski) is not sufficient--as it leaves the sensitivities to "missing distributions" the same. But what Dr. Dembski misses is that theory about intelligent agent action, constraining probabilities thereof, are necessary to reduce the M,N factors. This is the failure of eliminative method--not the failure to compare per se.
[ 10. November 2003, 00:37: Message edited by: gedanken ]
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 10. November 2003 01:11
gedanken,
I apologize, and withdraw my question of warren_bergerson.
RBH
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 10. November 2003 06:34
While I am dubious of the benefits of arbitrarily framing ‘motive in intelligence’ and ‘explanatory filters’ to preclude discussion of design or design processes, I will respect his request to control his own topic. If RBH wishes to rephrase his question to "Are there references documenting and quantifying the amount or complexity of specific designs? And, are there references showing how much quantified design can be generated by specific design processes in specific periods of time?" Then I will be glad to address his questions in a separate thread.
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 10. November 2003 10:14
Warren, some of what your method addresses is indeed details of mechanism of the designing process, if I am not mistaken. As such, that is on topic. But this thread is also about reliability of inferences--that is the element that I wanted to request be present. See in my OP:
quote:
I am requesting that such conceptual work should begin to include aspects of mechanism in the ID event, so as to give some greater degree of scientific content to ID analysis.
But "scientific content" was expanded as dealing with relibility of inferences, and not just general issues of what is scientific.
So if you have an entirely new method that discusses "mechanism" as opposed to "eliminative" approach that would per se be compatible with the OP. However I would still request that a completely new approach could be dealt with in a new thread, if it is not largely also discussing the other features like "reliability"--and simply leave a link here showing the relatioship to this thread. In fact, any "reliability" aspects developed (in that other thread) might be good opportunities to place another link HERE, showing the relationship to the reliability issue, as it might be highly relevant!
But I appreciate not using this thread for a long expanse on a different new theory, just simply due to the amount of additional information left to present that is very specific to the several pages of development that has occurred here. Thanks
[ 10. November 2003, 10:16: Message edited by: gedanken ]
IP: Logged
|
|
|
Printer-friendly view of this topic