ISF Logo   IS Forum
Forum Index Register Members List Events Mark Forums Read Help

Go Back   International Skeptics Forum » General Topics » Religion and Philosophy
 


Welcome to the International Skeptics Forum, where we discuss skepticism, critical thinking, the paranormal and science in a friendly but lively way. You are currently viewing the forum as a guest, which means you are missing out on discussing matters that are of interest to you. Please consider registering so you can gain full use of the forum features and interact with other Members. Registration is simple, fast and free! Click here to register today.
Tags atheism , stephen hawking

Reply
Old Today, 12:45 PM   #3841
Belz...
Fiend God
 
Belz...'s Avatar
 
Join Date: Oct 2005
Location: In the details
Posts: 76,412
Originally Posted by Myriad View Post
The issue raised by the OP is the question of generalizing from experience.

We examine a volcano and find that it's caused by subterranean heat and pressure, rather than by the anger of a volcano god. We examine fifty volcanos and find that all of them are caused by subterranean heat and pressure, and not one by the anger of volcano gods. But then a new volcano erupts somewhere that we've never examined before. Can we say with any confidence that that one's also caused by subterranean heat and pressure, and not by the anger of a volcano god?

Most people (nowadays) would say yes, though some of them would also caution against absolute certainty, and suggest we should probably go there and take a few measurements.

Now instead of fifty volcanos, let's say we've investigated the causes of fifty different natural phenomena. In all cases where we could identify causes, we found causes other than gods. How justifiable, how rational, is the conclusion that no natural phenomena are caused by gods?

On the one hand, the pattern established by experience so far is clear. On the other hand, the fifty-first, fifty-second etc. phenomena that remain unexplained aren't just randomly chosen phenomena that we haven't got around to looking at before. They're unexplained because for whatever reasons, they're harder to examine or explain than volcanos and the rest of the previous fifty. Those reasons might be complexity, scale, remoteness, etc. but one of the conceivable reasons certain things might be hard to explain is the involvement of gods.

The question of how far we're justified in generalizing isn't specific to science. It applies to all learning. It's a fundamental issue in all learning.

Under-generalizing is "not getting the concept," not seeing the pattern or "not seeing the forest for the trees." When I studied machine learning in the early 80s it was the crux of the "open minded learner" problem demonstrating that built-in biases were necessary for all but the crudest of learning. A completely open-minded learning algorithm could be told that for an input of 1, the correct output is 1; for an input of 2, the correct output is 2; ditto for 3, 4, 5, and 7. But if then told to figure out the correct output for 6, it could only output "unknown" because it hadn't been taught that specific rule. In other words an open-minded learner cannot learn; or more precisely, it can learn only in the sense of being able to parrot specifically taught instances. That falls under our definition of learning (a multiplication table, for instance) but doesn't encompass all that we mean by learning. To go farther the learning algorithm needs a "model" which amounts to a system of biases. For instance if programmed to find the shortest computation that relates output to input for all known instances, the learning algorithm will no longer be "open minded," instead having a strong bias toward simpler rules, but will quickly learn the rule "output = input," and will offer the output of 6 for an input of 6.

Over-generalizing is "jumping to conclusions" or "mistaking the map for the territory" or "comparing apples to oranges" or "every problem seeming like a nail." Fictional AI-gone-wrong scenarios usually feature something to that effect. HAL is programmed to "protect the mission" and tries to do so by killing the astronauts. Colossus is programmed to "defend the nation" and does so by taking over the world.

Both possibilities loom in every learning situation. Both are the basis for jokes, sometimes but not always about silly things children do. From my own childhood, a comedy bit on Laugh-In by Totie Fields stuck in my memory: a toddler-age character (much like Lily Tomlin's "Edith Ann" character, but different) recites a series of laments: "Nobody TOLD me I shouldn't paint the baby. They said not to paint the walls, or the floor..." That's failure to generalize, which we associate with cognitive immaturity. (Fields' character appeared to be regressing in age with each transgression in the series.) In other family comedy bits, toddlers do things like try to feed the goldfish by dumping dog food in the bowl. That's generalizing too far, which we often associate with inexperience. It's the reason why "a little learning is a dangerous thing." It's usually the reason for the problems the "FNG" causes at work. FNGs know the dozen fundamental rules for how things should be done, but not the eight hundred exceptions.

Clearly the issue goes way beyond early childhood learning. Every romance is about learning not to generalize. "Can't you see Broody McLeatherpants isn't like all the other vampires?" Every recovery narrative is about learning to generalize. "I've just realized that every time I reach this point and go get a bottle of cheap whisky and open it up and drink it, it doesn't help! Maybe I should try something else instead."

The issue of this thread is not how science should be done, but how we should or should not learn from it. Some want to limit our learning to only the specific results investigated, like the "open minded learner" algorithm that can only repeat what it's been taught. "Be open-minded! That fifty-first volcano might have a god in it!" Others want to take models that might be shaky or incomplete and extrapolate them literally to the ends of the universe (I'm looking at you, cosmologists, though some other examples come to mind, such as the "central dogma of genetics" that proved over-generalized).

More specifically, some see only a small distinction between "no volcano gods" and "no gods at all," while others see a huge gulf. The first view can be justified on the basis of volcanos once being among the natural phenomena that were very consistently and solely attributed to the actions of gods. (They're even among the events that insurance companies technically term "acts of god" to this day.) The second can be justified on the basis of the still-unexplained things such as conscious experience, the origin of life, and the origin of the universe being, in present day eyes, so much very more mysterious than mere volcanos.

Hawking's viewpoint adds a bit of a new wrinkle: the suggestion that even if the second view is correct, and there is a huge divide between "no volcano gods" and "no gods at all," that science has succeeded in crossing it. That is to say, "no gods exist" is a reasonable generalization to have learned by now.
Excellent post, Myriad. You rule.
__________________
Master of the Shining Darkness

"My views are nonsense. So what?" - BobTheCoward


Belz... is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old Today, 01:09 PM   #3842
Skeptic Ginger
Nasty Woman
 
Skeptic Ginger's Avatar
 
Join Date: Feb 2005
Posts: 70,398
Originally Posted by David Mo View Post
According to you there is a science that is not published in scientific papers. But if that science is science it will work with the scientific method. If not, say how we can know what other method follows.
Do you know how silly this sounds? Gee, how did anyone figure anything out before there were science journals?

Originally Posted by David Mo View Post
If that supposed science has not studied the subject of God, we are not interested. If it has studied it, say how we can know and contrast.
If neither one or the other, stop teasing us.
God beliefs are studied all the time and very much published in science journals. Beyond beliefs, there is no evidence of anything left to study.
__________________
Restore checks and balances no matter your party affiliation.
Skeptic Ginger is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old Today, 01:11 PM   #3843
Skeptic Ginger
Nasty Woman
 
Skeptic Ginger's Avatar
 
Join Date: Feb 2005
Posts: 70,398
Originally Posted by JoeMorgue View Post
Nothing anybody has said is even close to that. Stop lying.
To be fair, I have though I don't think it's what David is referring to. There is a ton of research that isn't published.
__________________
Restore checks and balances no matter your party affiliation.
Skeptic Ginger is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old Today, 01:14 PM   #3844
Skeptic Ginger
Nasty Woman
 
Skeptic Ginger's Avatar
 
Join Date: Feb 2005
Posts: 70,398
Originally Posted by xjx388 View Post
... But that came about because scientific explanations supplanted religious explanations and most people are smart enough to realize that this probably means religious ideas are invalid.
Just as scientific explanations supplant an awful lot of philosophy as we gain a better understanding of how brains make moral and other subjective choices.
__________________
Restore checks and balances no matter your party affiliation.
Skeptic Ginger is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Old Today, 01:23 PM   #3845
Skeptic Ginger
Nasty Woman
 
Skeptic Ginger's Avatar
 
Join Date: Feb 2005
Posts: 70,398
Originally Posted by Myriad View Post
The issue raised by the OP is the question of generalizing from experience.

We examine a volcano and find that it's caused by subterranean heat and pressure, rather than by the anger of a volcano god. We examine fifty volcanos and find that all of them are caused by subterranean heat and pressure, and not one by the anger of volcano gods. But then a new volcano erupts somewhere that we've never examined before. Can we say with any confidence that that one's also caused by subterranean heat and pressure, and not by the anger of a volcano god?

Most people (nowadays) would say yes, though some of them would also caution against absolute certainty, and suggest we should probably go there and take a few measurements.

Now instead of fifty volcanos, let's say we've investigated the causes of fifty different natural phenomena. In all cases where we could identify causes, we found causes other than gods. How justifiable, how rational, is the conclusion that no natural phenomena are caused by gods?

On the one hand, the pattern established by experience so far is clear. On the other hand, the fifty-first, fifty-second etc. phenomena that remain unexplained aren't just randomly chosen phenomena that we haven't got around to looking at before. They're unexplained because for whatever reasons, they're harder to examine or explain than volcanos and the rest of the previous fifty. Those reasons might be complexity, scale, remoteness, etc. but one of the conceivable reasons certain things might be hard to explain is the involvement of gods.

The question of how far we're justified in generalizing isn't specific to science. It applies to all learning. It's a fundamental issue in all learning.

Under-generalizing is "not getting the concept," not seeing the pattern or "not seeing the forest for the trees." When I studied machine learning in the early 80s it was the crux of the "open minded learner" problem demonstrating that built-in biases were necessary for all but the crudest of learning. A completely open-minded learning algorithm could be told that for an input of 1, the correct output is 1; for an input of 2, the correct output is 2; ditto for 3, 4, 5, and 7. But if then told to figure out the correct output for 6, it could only output "unknown" because it hadn't been taught that specific rule. In other words an open-minded learner cannot learn; or more precisely, it can learn only in the sense of being able to parrot specifically taught instances. That falls under our definition of learning (a multiplication table, for instance) but doesn't encompass all that we mean by learning. To go farther the learning algorithm needs a "model" which amounts to a system of biases. For instance if programmed to find the shortest computation that relates output to input for all known instances, the learning algorithm will no longer be "open minded," instead having a strong bias toward simpler rules, but will quickly learn the rule "output = input," and will offer the output of 6 for an input of 6.

Over-generalizing is "jumping to conclusions" or "mistaking the map for the territory" or "comparing apples to oranges" or "every problem seeming like a nail." Fictional AI-gone-wrong scenarios usually feature something to that effect. HAL is programmed to "protect the mission" and tries to do so by killing the astronauts. Colossus is programmed to "defend the nation" and does so by taking over the world.

Both possibilities loom in every learning situation. Both are the basis for jokes, sometimes but not always about silly things children do. From my own childhood, a comedy bit on Laugh-In by Totie Fields stuck in my memory: a toddler-age character (much like Lily Tomlin's "Edith Ann" character, but different) recites a series of laments: "Nobody TOLD me I shouldn't paint the baby. They said not to paint the walls, or the floor..." That's failure to generalize, which we associate with cognitive immaturity. (Fields' character appeared to be regressing in age with each transgression in the series.) In other family comedy bits, toddlers do things like try to feed the goldfish by dumping dog food in the bowl. That's generalizing too far, which we often associate with inexperience. It's the reason why "a little learning is a dangerous thing." It's usually the reason for the problems the "FNG" causes at work. FNGs know the dozen fundamental rules for how things should be done, but not the eight hundred exceptions.

Clearly the issue goes way beyond early childhood learning. Every romance is about learning not to generalize. "Can't you see Broody McLeatherpants isn't like all the other vampires?" Every recovery narrative is about learning to generalize. "I've just realized that every time I reach this point and go get a bottle of cheap whisky and open it up and drink it, it doesn't help! Maybe I should try something else instead."

The issue of this thread is not how science should be done, but how we should or should not learn from it. Some want to limit our learning to only the specific results investigated, like the "open minded learner" algorithm that can only repeat what it's been taught. "Be open-minded! That fifty-first volcano might have a god in it!" Others want to take models that might be shaky or incomplete and extrapolate them literally to the ends of the universe (I'm looking at you, cosmologists, though some other examples come to mind, such as the "central dogma of genetics" that proved over-generalized).

More specifically, some see only a small distinction between "no volcano gods" and "no gods at all," while others see a huge gulf. The first view can be justified on the basis of volcanos once being among the natural phenomena that were very consistently and solely attributed to the actions of gods. (They're even among the events that insurance companies technically term "acts of god" to this day.) The second can be justified on the basis of the still-unexplained things such as conscious experience, the origin of life, and the origin of the universe being, in present day eyes, so much very more mysterious than mere volcanos.

Hawking's viewpoint adds a bit of a new wrinkle: the suggestion that even if the second view is correct, and there is a huge divide between "no volcano gods" and "no gods at all," that science has succeeded in crossing it. That is to say, "no gods exist" is a reasonable generalization to have learned by now.
I would simplify this. Often you can approach a problem from different angles.

You can ask, are there any real gods?
Or you can ask, what best explains god beliefs?

The first question is based on the fact people believe in gods, it is not based on any actual evidence gods exist.

The second question is based on actual observations. Scientific inquiry usually starts with an observation and proceeds from there.
__________________
Restore checks and balances no matter your party affiliation.
Skeptic Ginger is offline   Quote this post in a PM   Nominate this post for this month's language award Copy a direct link to this post Reply With Quote Back to Top
Reply

International Skeptics Forum » General Topics » Religion and Philosophy

Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -7. The time now is 05:22 PM.
Powered by vBulletin. Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.

This forum began as part of the James Randi Education Foundation (JREF). However, the forum now exists as
an independent entity with no affiliation with or endorsement by the JREF, including the section in reference to "JREF" topics.

Disclaimer: Messages posted in the Forum are solely the opinion of their authors.