|
Author
|
Topic: The Characterization of Intelligent Causation
|
2ndclass
Member
Member # 1979
|
posted 03. April 2007 13:21
quote:
If the processing or output of a system changes from time t=0 to time t=1 with the result at time t=1 being goal compatible, and if the change is faster or more efficient than can be explained by a random search, then the change involves intelligent causation. If the criteria is not satisfied, the change does not involve intelligent causation.
Everything that isn't random meets your criteria, as has been explained to you for a year now. A particle flying through space is moving toward its "goal" in a way that is far more efficient than randomly wandering around. (I put the word "goal" in quotes because anything can arbitrarily be considered a goal in your framework, since you have haven't put any constraints on it.)
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 03. April 2007 13:44
Quote: Please give me 1) an example of something you believe can be modelled using your model of intelligence, and 2) an example of something that cannot. For (1), show the model.
Your question is ambiguous. Are you asking about what can and what can not be modeled using the framework presented or are you asking about the difference between a model of intelligent behavior and a model of unintelligent behavior? Clearly what can and what can not be modeled are not the same as intelligent versus non-intelligent.
IP: Logged
|
|
Daniel Smith
Member
Member # 3004
|
posted 03. April 2007 14:50
It seems to me that the ID movement rightly avoids trying to define the undefinable - "Intelligence".
Rather than do that, they concentrate on known intelligent designs and look for similarities in biological "designs".
The (good) reason for approaching the problem this way is because we already have a working model for intelligence in action - humans. We can readily observe the results of intelligent (a given that does not need defined) design.
The main problem with trying to define intelligence is that so much of it is subjective: How do you measure "creativity", "ingenuity", "artistry", etc.? How do you define "beauty"? These all fall under the "eye of the beholder" and "you know it when you see it" category. Yet humans regularly use the former to produce the latter. And, some people have none of these talents - so to try to lump everything under a broad definition of "intelligence" is fruitless.
So, as I see it, the only scientific way to deduce intelligent causation for biological complexity is to compare objects of biological complexity to objects known to be the result of intelligent causation.
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 03. April 2007 15:02
quote:
It seems to me that the ID movement rightly avoids trying to define the undefinable - "Intelligence".
Rather than do that, they concentrate on known intelligent designs and look for similarities in biological "designs".
But what constitutes "known intelligent designs"? Does it include everything produced by humans? If so, then that fact constitutes a partial definition of "intelligence", which you say is undefinable. On the other hand, if not all human artifacts are intelligent designs, then how do we determine which are and which aren't without a definition of "intelligent"?
Humans make spheres, e.g. balls. Gravity makes spheres, e.g. planets. Does this similarity indicate intelligence on the part of gravity?
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 03. April 2007 16:48
quote:
Your question is ambiguous. Are you asking about what can and what can not be modeled using the framework presented or are you asking about the difference between a model of intelligent behavior and a model of unintelligent behavior? Clearly what can and what can not be modeled are not the same as intelligent versus non-intelligent.
You say you can model processes, and by seeing how the model acts, you can say if the process is intelligent or not, right?
In order to answer the question posed in the OP, we would need a model of human design abilities, and a demonstration that this model meets the criteria for a model of an intelligent process (more efficient than random search). Then we would need a model of how some biological complex mechanism arose, and that too must be shown to pass the criteria for an intelligent process.
It is my belief that neither model can be built, of course - nobody can build a mathematical model of how people design a watch, and nobody can build a mathematical model of how a flagellum came to exist. You obviously disagree, and you think that you can build models like this.
In order to resolve the matter, you simply need to provide these models. Since both processes are quite complex, the model would presumably be very, very large, and I assume the model will be represented as some sort of computer code (pseudo code, or code in an actual computer language). So perhaps you can provide a link to a document that contains these models.
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 04. April 2007 11:21
Aiguy,
Quote: You say you can model processes, and by seeing how the model acts, you can say if the process is intelligent or not, right?
Not quite. I am saying that using a well established paradigm or structure, I can model a wide range of behaviors including behaviors of humans, evolutionary change behaviors and behaviors of computers. I am then saying that using this framework, I can develop a formal definition of intelligent causation that can be used to classify behaviors, or the models of behaviors, as either intelligent or non-intelligent.
Then following standard scientific methodologies, I validate the definition by confirming that it classifies behaviors generally recognized as intelligent as intelligent and behaviors generally recognized as unintelligent as unintelligent. I can then apply the definition to the evaluation of behaviors in dispute and determine if based on the definition proposed the behaviors are or are not intelligent.
I believe, it should be obvious that based on the definition being proposed, lots of behaviors in humans, evolution, computers, and lots of behaviors in biological systems will qualify as intelligent or as involving intelligent causation. Do you question such a conclusion?
Note that based on the criteria defined, I do not claim that the capacity to classify behaviors is, by itself enough to justify a scientific definition of intelligent causation. I am suggesting that the definition must also be useful in formulating non-trivial predictive theories and that it must do so better than any proposed competing definition.
Quote: In order to answer the question posed in the OP, we would need a model of human design abilities, and a demonstration that this model meets the criteria for a model of an intelligent process (more efficient than random search). Then we would need a model of how some biological complex mechanism arose, and that too must be shown to pass the criteria for an intelligent process.
This is sort of correct. To answer the question in the op, based on a ‘current’ knowledge basis, you would need to model an elementary form of human design behavior and show 1) that behavior satisfied the definition and 2) that a non-trivial predictive theory can be formulated for the behavior. Then you would have to identify a simple or elementary form of evolutionary design behavior and show that 3) that behavior satisfied the proposed definition and 4) that a non-trivial predictive theory could be formulate for the behavior. Having validated that the definition applies to both elementary human behaviors and elementary evolutionary behaviors, testing and analysis would continue to determine if the definitions applied to more complex behaviors and to determine if there were better definitions that did or did not apply to both human and evolutionary behaviors.
Quote: In order to resolve the matter, you simply need to provide these models. Since both processes are quite complex, the model would presumably be very, very large, and I assume the model will be represented as some sort of computer code
Why would think you need to start by analyzing complex behaviors rather than starting by analyzing the elementary or relatively simple components of the complex behaviors?
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 04. April 2007 12:56
LE,
quote:
AIGUY: You say you can model processes, and by seeing how the model acts, you can say if the process is intelligent or not, right?
LE: Not quite. I am saying that using a well established paradigm or structure, I can model a wide range of behaviors including behaviors of humans, evolutionary change behaviors and behaviors of computers. I am then saying that using this framework, I can develop a formal definition of intelligent causation that can be used to classify behaviors, or the models of behaviors, as either intelligent or non-intelligent.
I can't see any difference between these two statements. I guess there's some ambiguity about whether you are saying you can tell when the process you are modelling is intelligent, or when the model itself is intelligent: "can be used to classify behaviors, or the models of behaviors, as either intelligent or non-intelligent".
Aside from that, I think I said the same thing.
quote:
Then following standard scientific methodologies, I validate the definition by confirming that it classifies behaviors generally recognized as intelligent as intelligent and behaviors generally recognized as unintelligent as unintelligent.
So here you are saying that you test out your approach to building these models by making sure the result match our intuitions, right? What is "generally recognized" as intelligent, without using your formal definition, must be recognized via just an informal method, right?
quote:
I can then apply the definition to the evaluation of behaviors in dispute and determine if based on the definition proposed the behaviors are or are not intelligent.
Ok, I think I understand.
quote:
I believe, it should be obvious that based on the definition being proposed, lots of behaviors in humans, evolution, computers, and lots of behaviors in biological systems will qualify as intelligent or as involving intelligent causation. Do you question such a conclusion?
Of course I question it, until I see some of these models you are talking about. I've never seen a formal model of human behavior before.
quote:
Note that based on the criteria defined, I do not claim that the capacity to classify behaviors is, by itself enough to justify a scientific definition of intelligent causation. I am suggesting that the definition must also be useful in formulating non-trivial predictive theories and that it must do so better than any proposed competing definition.
So here you are saying that once you have your model of human behavior built, you can actually predict what human beings are going to do in a given situation? Wow! That could come in handy.
quote:
Why would think you need to start by analyzing complex behaviors rather than starting by analyzing the elementary or relatively simple components of the complex behaviors?
Ok, fine, let's see a model of a relatively simply human behavior, and a relatively simple example of the generation of a biological structure. Is your model written in actual code, running on a computer?
IP: Logged
|
|
Daniel Smith
Member
Member # 3004
|
posted 04. April 2007 13:45
2ndclass:
quote:
But what constitutes "known intelligent designs"? Does it include everything produced by humans? If so, then that fact constitutes a partial definition of "intelligence", which you say is undefinable.
You concede my point when you admit that "everything produced by humans" would only constitute a "partial definition" of intelligence.
The term "intelligence" has too broad a definition to be scientifically useful.
quote:
On the other hand, if not all human artifacts are intelligent designs, then how do we determine which are and which aren't without a definition of "intelligent"?
All human artifacts are known to be the result of intelligent agency (that is, if we can agree that all humans are intelligent).
quote:
Humans make spheres, e.g. balls. Gravity makes spheres, e.g. planets. Does this similarity indicate intelligence on the part of gravity?
We have never seen gravity make a planet. How do we know that gravity - acting alone - can make planets? ![[Big Grin]](biggrin.gif)
But (as per Dembski's explanatory filter) if it can be shown that gravity can make a planet, then that rules out the necessity of intelligent causation. If we find a spherical rock, we don't automatically assume human design - since natural forces can produce such things. If we find a pyramid however, we can safely assume intelligent causation.
[ 04. April 2007, 13:52: Message edited by: Daniel Smith ]
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 04. April 2007 14:07
Daniel,
quote:
You concede my point when you admit that "everything produced by humans" would only constitute a "partial definition" of intelligence. The term "intelligence" has too broad a definition to be scientifically useful.
If the term "intelligence" is not scientifically useful (and I couldn't agree more), then the meaning of "intelligent causation", or "intelligent designs", which clearly depend on the meaning of "intelligence", are likewise not scientifically useful.
quote:
All human artifacts are known to be the result of intelligent agency (that is, if we can agree that all humans are intelligent).
This is not a proposition of fact, but only of arbitrary definition. If we cannot decide what constitutes intelligence in a way that can be unambiguously tested, we can't possibly know what is or is not intelligent. In our casual usage of the word, of course we refer to human beings as intelligent, or as intelligent agents. It is not at all clear, however, what else might or might not be an intelligent agent. Which other animals or systems that we know of are intelligent agents? Without an operationalized definition of intelligence, there is simply no fact of the matter at all.
quote:
We have never seen gravity make a planet. How do we know that gravity - acting alone - can make planets?
Because our detailed models of gravitational effects explain and predict the shape of planets (to some degree of certainty). Are you suggesting that if we can't prove that, we resort to an explanation of intelligent design for the shape of planets? But it's obvious that all sorts of things make spheres. Bubbles on the ocean are spherical... are those created by intelligent processes?
quote:
But (as per Dembski's explanatory filter) if it can be shown that gravity can make a planet, then that rules out the necessity of intelligent causation. If we find a spherical rock, we don't automatically assume human design - since natural forces can produce such things. If we find a pyramid however, we can safely assume intelligent causation.
Purely an argument from ignorance, I'm afraid. The explanatory filter assumes that everything we can't explain must share an identical cause, which is utterly unfounded. We cannot explain how proteins are folded into correct 3-D conformations inside of cells. Nor can we explain how people manage to design watches. By the explanatory filter, we are forced to conclude that the same thing is responsible for both phenomena. I find that rather hard to believe.
[ 04. April 2007, 14:07: Message edited by: aiguy ]
IP: Logged
|
|
2ndclass
Member
Member # 1979
|
posted 04. April 2007 16:26
quote:
You concede my point when you admit that "everything produced by humans" would only constitute a "partial definition" of intelligence.
The term "intelligence" has too broad a definition to be scientifically useful.
If your point is that the term "intelligence" is undefinable, then no, I don't concede that. Every word is definable. On the other hand, if your point is that the current definition of "intelligence" isn't scientifically useful (except as applied to entities that can be tested with standard IQ tests), then I certainly concede that.
quote:
All human artifacts are known to be the result of intelligent agency (that is, if we can agree that all humans are intelligent).
I take that to mean that "known intelligent designs" includes all human artifacts.
You say that IDists "concentrate on known intelligent designs and look for similarities in biological 'designs'." But this method tells us that planets are intelligent designs.
quote:
But (as per Dembski's explanatory filter) if it can be shown that gravity can make a planet, then that rules out the necessity of intelligent causation.
Are you saying that IDists apply Dembski's method after applying the above method? (The two are very different.)
quote:
If we find a pyramid however, we can safely assume intelligent causation.
Yes, because human artifacts are intelligent designs, by your definition. But to make that inference for pyramids on Mars, we would need to expand our definition of "intelligent designs". Can we say that the set "intelligent designs" is defined to consist of (1) human artifacts, plus (2) objects that are similar to human artifacts, as long as they pass through Dembski's EF?
IP: Logged
|
|
LifeEngineer
Member
Member # 3446
|
posted 05. April 2007 11:52
Quote: So here you are saying that you test out your approach to building these models by making sure the result match our intuitions, right? What is "generally recognized" as intelligent, without using your formal definition, must be recognized via just an informal method, right?
No. I don't think anyone is naive enough to believe that it will ever be possible to reach a consensus on what is and what is not intelligent behavior. The development and validation of definitions begins by determining if the proposed definitions work on phenomena that are clearly intelligent behavior/information processing and on behavior and information processing that is clearly not intelligent. The next step, the step that I did not address in my earlier response, involves isolating simpler and simpler components of intelligent and non-intelligent behaviors. By the time you have isolated relatively simple components of complex intelligent behaviors, you are probably dealing with information processing that not everyone would intuitively recognize as intelligent.
The behaviors that I initially identified clearly involving intelligent causation were human goal directed decisions with the general form ‘select the best available option’. More specifically, I started with dynamic goal directed decision making where the ‘best option’ could change over time or could be different for different decision makers. Intelligent causation, based on my intuitive view, was involved if the decision process or logic could rapidly or efficiently change to reflect a change in the best available option.
Unintelligent causation was involved if the logic could not change to reflect the change in ‘best option’ or if the change in logic was no faster than what would be achieved by a trial and error search process.
This ‘efficient adaptive change in information’ processing is an intuitive concept of intelligent or ‘my’ intuitive concept of intelligent causation. Clearly not everyone will agree nor is it necessary for everyone to agree with this intuitive concept.
The process is 1) develop or identify AN intuitive concept of intelligent causation that 2) is demonstrably and reasonably consistent with at least some intuitive concepts of intelligent causation, and 3) translate this intuitive concept into a formal mathematical/scientific definition and 4) show that the formal definition is useful in formulating hard science predictive theories.
There are lots of different and often logically incompatible ‘intuitive concepts’ of intelligence and intelligent causation. It is useful to understand which intuitive concept is being in developing a particular definition, but there is not much to be gained by debating the relative values of different intuitive concepts.
IP: Logged
|
|
Melvin H. Fox
Member
Member # 1684
|
posted 05. April 2007 12:34
Humans can seek out, collect, store, organize, synthesize, analyze, and characterize information regarding the present reality. Humans can project these characterizations of reality into the past or future and make predictions concerning events that have occurred or will occur. Humans can do these things because they are intelligent. These characterizations of reality are not consistent from human to human. They are not even consistent for the individual human.
I submit that if humans are heuristically programmed machines logical coherence dictates these characterizations of reality should be converging on the true character of reality. Instead the characterizations are diverging. Even “scientific” characterizations are diverging. One need go no further than this forum to get a good dose of diverging scientific conclusions about reality.
I will give you one extreme yet common example of every day divergence. Just months ago a student in one of my calculus classes came to me and complained about points taken away on a test. She had integrated incorrectly because she tried to apply the product rule of differentiation when she was integrating a product. When I told her she could not do this she replied “why not?” After receiving my detailed explanation her retort was, “who says?” It made sense to her to use the product rule when integrating, so she concluded it was the “right” thing to do and who is this Prof. to tell her otherwise? After all, she got an answer.
quote:
If logical coherence isn't a requirement for something to be true, then there's no point in discussing anything. Everything's true and everything's false. Are you sure you want to go that route?
I thought long and hard about the above comment. No, 2ndclass, I don’t want to go down that route. I have free-will but not unlimited power. If that is the route that human behavior is headed down, I can’t stop it. I can try to explain it. My explanation is free-will.
It seems aiguy would like to explain the inconsistencies and resulting divergence by chance alone. I am aware of his statement:
quote:
So what explains inconsistencies in human behavior? Two things: First, even if we were perfectly deterministic machines, a human being need never respond the same way twice, since the totality of our previous experiences plus our innate characteristics would determine our responses to our circumstance. And second, perhaps just chance.
But, differences in previous experiences and innate characteristics are not determined by fixed law. If not purposefully specified by some intelligence, then chance is the only other explanation left to us by aiguy. Since the scientific definition of chance is, “humm, we don’t really know what causes it”, then simply admitting we don’t really know what causes biological complexity is answer for the query.
-Mel
IP: Logged
|
|
Daniel Smith
Member
Member # 3004
|
posted 05. April 2007 14:45
aiguy:
quote:
We cannot explain how proteins are folded into correct 3-D conformations inside of cells. Nor can we explain how people manage to design watches. By the explanatory filter, we are forced to conclude that the same thing is responsible for both phenomena. I find that rather hard to believe.
That's actually the whole point of the filter - to show that both watches, and protein folding, are most likely the product of design (I'll leave off the word "intelligent" since you have problems with it).
2ndclass:
quote:
Yes, because human artifacts are intelligent designs, by your definition. But to make that inference for pyramids on Mars, we would need to expand our definition of "intelligent designs". Can we say that the set "intelligent designs" is defined to consist of (1) human artifacts, plus (2) objects that are similar to human artifacts, as long as they pass through Dembski's EF?
That would be a good start. The question is a good one. How would we go about determining if pyramids (or any other structures) on Mars were designed? I think we'd have no choice but to compare them to human designs.
If you can describe the tests necessary to determine whether an object on Mars is designed or not, you can then turn around and apply those same criteria to biological complexity.
To be honest, this is where most 'metaphysical naturalists' bug out of the discussion. They see that the tests are rather simple and that when applied to biological complexity, they'll point to design, so they avoid the question, change the subject, or don't respond at all.
IP: Logged
|
|
aiguy
Member
Member # 3736
|
posted 05. April 2007 14:46
Hi Mel,
quote:
Humans can do these things because they are intelligent.
Really? And why do we say that humans are intelligent? Humans are intelligent because they can do these things. This is perfectly circular.
Why are both of these statements equally true, even though they are the converse of each other? This is not the case for other propositions, such as "Humans can see because they have eyes": We cannot say humans have eyes because they can see. What is the difference between these propositions? The first, "Humans can do these things because they are intelligent", is an analytic statement, necessarily true but only by definition. The second, "Humans can see because they have eyes", is a synthetic statement - a statement of contingent fact. We cannot use "intelligence" to explain anything that humans do, because we simply choose to call those things "intelligent".
quote:
I submit that if humans are heuristically programmed machines logical coherence dictates these characterizations of reality should be converging on the true character of reality.
I have no idea why you would say this. My heuristically programmed AI systems don't necessarily converge on the true character of reality, although I would be very happy if they did.
quote:
My explanation is free-will.
That's fine, but it is not demonstrable.
quote:
It seems aiguy would like to explain the inconsistencies and resulting divergence by chance alone.
Actually, I gave two reason, only one of which was chance.
quote:
But, differences in previous experiences and innate characteristics are not determined by fixed law.
Fixed law and chance, as far as we know.
quote:
If not purposefully specified by some intelligence, then chance is the only other explanation left to us by aiguy.
No, fixed law and chance. Saying something is specified by "some intelligence" is no explanation at all - it has no definite meaning that we can tie to our shared experience.
quote:
Since the scientific definition of chance is, “humm, we don’t really know what causes it”, then simply admitting we don’t really know what causes biological complexity is answer for the query.
Yes, that is the definition of chance. It is also the scientific definition of intelligent causation!!!
And yes, I am perfectly satisfied with the answer "we don't really know" what causes biological complexity. However, nobody attempts to explain anything with chance alone. And nobody should attempt to explain anything by "intelligence".
[ 05. April 2007, 14:48: Message edited by: aiguy ]
IP: Logged
|
|
Stephen Wright
Member
Member # 195
|
posted 05. April 2007 16:11
quote:
Humans make spheres, e.g. balls. Gravity makes spheres, e.g. planets. Does this similarity indicate intelligence on the part of gravity? - 2ndclass
Earlier you spoke to a goal states, for undirected and directed events. Planets are formed without purposeful adjustments to a preconceived target state. The resulting structure emerges from a matrix of deterministic (mainly) forces and the available matter. The vectors toward a finished state, of such a process, are solely from the physical forces in the environment, without intervention.
Cybernetics is the study of just these directed adjustments to a target state. Humans or other living things have a capacity to create a virtual state where a ball is conceived. This conception exists only in this probable form. At the start and during formation – appropriate adjustments alter the project toward its target state. Timing is everything; in many cases the beginning of the process is related to special related circumstances occurring in the immediate environment. Energy is spent on the preconception, evaluation of the timing for start-up, gathering data on the development and in enforcing the design’s image, as a standard, during development.
Intelligence is indicated – strongly – by phenomena that exhibit use of feedback to guide events toward a goal or pre-imaged structure. Evidence for these adjustments and of choice in timing would be evidence for intelligence.
IP: Logged
|
|
|
Printer-friendly view of this topic
