"Here is a point on which we may not agree though: What about the domain of the a priori? Propositions that are true or false merely based on the meanings of the terms involved. A propositions like "this triangle has 4 sides" seems to me like it has a probability of 0, and "all bachelors are unmarried" seems to me like it has a probability of 1, just based on the meanings of the terms involved. These propositions don't involve ideas we have to check against some perceptual data--they relate properties of language and definitions, and they seem to be false and true, respectively just based on their definitions. I'm comfortable assigning probabilities of 0 or 1 to these kinds of propositions, and I don't see this as necessarily in conflict with model agnosticism. What's your take? "

Hrm, thats a good point, let me see how I feel about it. Ok, it seems to me what your talking about two things. First the internal consistancy of your models can be tested for logical consitancy against themselves. On this level of models, true and false are not attempting to make definitive statments about reality, rather, the attempt seems to be to have the ideas that you have about reality match as consistantly as possible with the data you seem to recive from the outside world. Since no set of ideas is likley to be complete unsupported by yet other ideas, true and false in this context seem to be a measure of use/consitancy relative to the rules of logic as we understand them, and on this level the ideas of true and false are valid only within a specific context. All Ideas may be inherantly valid in some sense, all are most definatly not equally useful in all contexts (such as being a living, breathing, human animal). Now the second part of this goes beyond testing your models for internal consistancy by comparing their predicted results against "reality". In this sense true and false only can refer to the ablity of a model to predict, and not to ultimatly judge reality, which we can not percieve whole cloth. Now some of these constants seem to come from "objective" reality. A triangle to me by definiton seems to be a triangle to you. On the other hand, a triangle is just a two dimentional abstaction we have built out of our third dimention based five senses. If we could peer more deeply into the heart of things I think we might find that the concept "triangle" only represents a fraction of the actual process. In this sense true and false should not be assumed to be attempting to describe reality in a non relative context. So either way the terms only can be usefully applied to your own internal systems I think. I don't believe in triangles, or assume that they are an objective part of the universe. On the other hand, I must admit, to my personal system of senses, and to that of other human beings as far as I can tell, it does seem true that a triangle always has 3 sides. I guess I just object to the use of the words true and false In a non relative question. I hope that awnsers what your getting at, you ask good questions.

"Here is a point on which we may not agree though: What about the domain of the a priori? Propositions that are true or false merely based on the meanings of the terms involved. A propositions like "this triangle has 4 sides" seems to me like it has a probability of 0, and "all bachelors are unmarried" seems to me like it has a probability of 1, just based on the meanings of the terms involved. These propositions don't involve ideas we have to check against some perceptual data--they relate properties of language and definitions, and they seem to be false and true, respectively just based on their definitions. I'm comfortable assigning probabilities of 0 or 1 to these kinds of propositions, and I don't see this as necessarily in conflict with model agnosticism. What's your take? "

Hrm, thats a good point, let me see how I feel about it. Ok, it seems to me what your talking about two things. First the internal consistancy of your models can be tested for logical consitancy against themselves. On this level of models, true and false are not attempting to make definitive statments about reality, rather, the attempt seems to be to have the ideas that you have about reality match as consistantly as possible with the data you seem to recive from the outside world. Since no set of ideas is likley to be complete unsupported by yet other ideas, true and false in this context seem to be a measure of use/consitancy relative to the rules of logic as we understand them, and on this level the ideas of true and false are valid only within a specific context. All Ideas may be inherantly valid in some sense, all are most definatly not equally useful in all contexts (such as being a living, breathing, human animal). Now the second part of this goes beyond testing your models for internal consistancy by comparing their predicted results against "reality". In this sense true and false only can refer to the ablity of a model to predict, and not to ultimatly judge reality, which we can not percieve whole cloth. Now some of these constants seem to come from "objective" reality. A triangle to me by definiton seems to be a triangle to you. On the other hand, a triangle is just a two dimentional abstaction we have built out of our third dimention based five senses. If we could peer more deeply into the heart of things I think we might find that the concept "triangle" only represents a fraction of the actual process. In this sense true and false should not be assumed to be attempting to describe reality in a non relative context. So either way the terms only can be usefully applied to your own internal systems I think. I don't believe in triangles, or assume that they are an objective part of the universe. On the other hand, I must admit, to my personal system of senses, and to that of other human beings as far as I can tell, it does seem true that a triangle always has 3 sides. I guess I just object to the use of the words true and false In a non relative question. I hope that awnsers what your getting at, you ask good questions.

Ok, tribes seems a little laggy right now, sorry if this becomes a double or tripple post.

"The agnostic position in failing to discount ideas with no observational support or logical basis can not effectively counter these "crazy" ideas."

I see no need to counter ideas in order to dismiss the need to seriously consider them.

"The agnostic shugs his shoulders and says well, might be true, I better run for cover."

This is the fundamental mistake most people make about agnosticism; the assumption that just because somthing might possibly be true, that therefore is likley to be relevant, and even further, that all critical thinking tools used to analzye how useful of an idea it is (agnostics tend to think in terms of use rather than truth) must therefore be thrown out the window. Nothing could be less correct.

"The atheist says, it is just a dream man, cool down, have a drink and we will talk about it later. "

While the agnostic says "Dont worry about it man, it doesnt seem all that likley, and even if it were true it shouldnt make much difference on how you choose to live the rest of your life, so lets have a drink, look at this thing from all sides, and as soon as your less worried, lets look at whatever else seems interesting."

Rene, sorry I’ve taken so long to get back to you. Here are some thoughts on some of the things you said.

<”The atheist position is basically this: I do not believe in anything until and unless there is observational support and/or some sound logical basis.”>

This view does not sound necessarily atheistic—this is an epistemological view that mixes elements of rationalism with elements of empiricism. While it sounds eminently reasonable on the face of it, when we look closer, we find that there has been observational support for many ideas/theories that ultimately turned out to be false. Nearly any theory in the history of science not currently accepted is an example of exactly this phenomenon. We can see with hindsight why observationally supported theories like the phlogiston theory of chemistry or the caloric theory of heat were ultimately bad models. The scientists were making the wrong kinds of observations, drawing conclusions based on too little data, and didn’t have instruments sophisticated enough to capture the phenomena in enough detail. Clearly over the years, we get better at doing science, but nonetheless this has to make one wonder how many of the observation-supported theories we now take as “true” or “accurate” will turn out to be deeply flawed models of reality.

You characterize agnosticism as saying
<”we have no knowledge and can have no real knowledge, so essentially all ideas are equally viable to them.”>

So you take the agnostic to be saying that the probability of any and all propositions is exactly .5. I suppose this would count as a type of agnosticism, but its not the type I wish to defend here. In fact, this philosophy strikes me as utterly self-defeating because then the probability of agnosticism being the best choice would only be .5, just like atheism, fundamentalist islam, or any other theory. This is philosophical suicide. The model agnosticism that I wish to defend is the idea that our theories, our mental representations, our linguistic labels, are models of the world--less like sets of propositions that have truth values, and more like maps. But one does not say that a map of California is “true” or “false”, one says what it accounts for and how accurately (to the best of your finite knowledge) it represents what it is accounting for. This philosophy lends itself well to thinking in probabilities rather than absolutes. And thankfully there are probabilities other than .5.


<”The theist does not believe that our reason can accurately determine anything about the world and therefore they feel comfortable to rely on faith and put a great deal of emphasis on "authority" to supply them with what they should think.”>

I think this is probably an accurate statement with respect to the “masses” of religious people. But it does not apply so nicely to some intellectual theists. Some of the best formal logicians of the 20th century were Christians, and they thought that reason favored Christianity. I’m not saying I agree with them, I find it exceedingly unlikely that Christanity is anything like an accurate model of the universe, but I also can’t deny their superior mastery of formal logic. Personally I understand this as demonstrating that logic is a whore for our assumptions.

You agree with my idea of treating theories and ideas as models, and then you say that the important part is that they are based on observation. I totally agree. Observational success is arguably the main basis for favoring a model. But then you say that agnosticism implies that “nothing can be truly known and no idea should be rejected.” The model agnosticism that I am defending at the moment does not imply that no idea should be rejected. Models that do not correspond well with what we observe seem less likely, and have less reason to be taken seriously and less reason to be pursued by research. We don’t need to throw scarce government grant money investigating the polytheistic theory that every platypus is a deity. This obviously makes no sense of anything we observe and is therefore most likely an atrociously bad model of the external world. I imagine that your thought at this point is that theism is just as unlikely as the polytheistic platypus theory. And maybe you’re right. Maybe any sort of theism is just a bad model of the world, motivated by all the wrong kinds of observations. The theistic hypothesis can take many forms and I find them all pretty unlikely, but I’m not terribly convinced of any other story of the origin of stuff either, so… I guess I don’t really have a preferred model. Rene, I remember you saying that you are skeptical of the big bang theory, so I’m curious. Do you have a preferred model of the origin of stuff?

You conclude that agnosticism cannot combat fundamentalism because the agnostic never makes up his mind about anything because all options are equally likely. I hope I’ve made it clear that the type of agnosticism I’m talking about does not suffer from this flaw. Model agnosticism confronts fundamentalism by making a person aware that whatever ideas they happen to take as “truths” are generalizations made by a nervous system with finite capacities of observation that tends to make the best models it can given its informational inputs and its computational quirks. We can observe how an artificial neural network makes a set of generalizations and learns to recognize patterns in the world based on its prototypes, but we can also see the ways in which these networks make systematic errors when their learned prototypes are insufficient to represent the world in enough detail. Its hard not to see an analogy to the history of human thought here, especially given the remarkable architectural and funcational similarities of these aritifical neural nets to what we know of brain physiology. When we really get it into our heads that all models of phenomena in the external world are tentative generalizations based on finite experience, it’s difficult to sustain fundamentalism, which is largely based on accepting the proclamations of authority, as you said. Rather than thinking in terms of “X is true” or “X is false” one can learn to think in terms of “X is a relatively accurate (or inaccurate) model for these reasons…" It is difficult to sustain any sort of fundamentalism when you understand that your brain is a builder of models, not a container for truths, and accept your own fallibility.

I know probalby not everything I've said here is crystal clear, so if I need to elaborate on any of these points, let me know.