Earth has captured a second moon, says NASA

Originally posted by DeepThought
Think of it in risk terms. Each assumption in a theory has some chance of being wrong. The chance of the theory being right is the probability that none of the assumptions are wrong. So theories with lots of assumptions are more likely to be faulty than ones with fewer assumptions, simply because they rely on fewer things that can be wrong.

Well done. However. 🙂 The chance of a given assumption being wrong (or right) depends on a whole bunch of factors, different for each assumption. A thousand well-reasoned assumptions outweighs a simplistic single assumption that happens to depend on a bunch of let's say unexpected factors. I think the razor is a type of fallacy.

Originally posted by apathist
Well done. However. 🙂 The chance of a given assumption being wrong (or right) depends on a whole bunch of factors, different for each assumption. A thousand well-reasoned assumptions outweighs a simplistic single assumption that happens to depend on a bunch of let's say unexpected factors. I think the razor is a type of fallacy.

It's not supposed to be rigorous. Suppose we find some unexplained phenomenon. Two rival theories emerge and both explain the effect to acceptable precision. The theories are incompatible, based on different principles and cannot both be true and there isn't some sort of duality between them. We have no empirical basis for separating them and can't pretend they're really the same theory in different forms. If we are forced to choose one rather than the other then we'll choose the simpler, this is when Occam's razor should be invoked - when there is no better way of separating two theories. An alternative view is that there is no reason to choose one or the other if there is no reason to. People will naturally tend to use the one that is calculationally less intense, which is probably going to be the simpler one. So there's an element of rationalizing behaviour that is what happens anyway in the Occam's Razor argument. Yes it's a formal logical fallacy, certainly nothing can be proven using Occam's razor, but I think it's a reasonable approach when there isn't a better option.

Originally posted by apathist
The chance of a given assumption being wrong (or right) depends on a whole bunch of factors, . ...

I believe the term "..with all else being equal.." applies here for oc-razor.
Note oc-razor is not supposed to guarantee giving the correct theory the highest probability because oc-razor, just like the principle of induction, is probabilistic.

Originally posted by DeepThought... this is when Occam's razor should be invoked - when there is no better way of separating two theories. ...

So the razor gives a way of separating the theories. That doesn't make it a good idea.

Originally posted by humy
I believe the term "..with all else being equal.." applies here for oc-razor.
Note oc-razor is not supposed to guarantee giving the correct theory the highest probability because oc-razor, just like the principle of induction, is probabilistic.

Well, are there stats showing the razor works more often than not, to a statistically significant degree? Note the strange list of requirements before we apply the razor, as provided by DT:

Two rival theories emerge and both explain the effect to acceptable precision. The theories are incompatible, based on different principles and cannot both be true and there isn't some sort of duality between them. We have no empirical basis for separating them and can't pretend they're really the same theory in different forms. If we are forced to choose one rather than the other then we'll choose the simpler...

Sheez, when has that even ever occurred?

Originally posted by apathist
Well, are there stats showing the razor works more often than not, to a statistically significant degree? Note the strange list of requirements before we apply the razor, as provided by DT:

Two rival theories emerge and both explain the effect to acceptable precision. The theories are incompatible, based on different principles and cannot both be ...[text shortened]... han the other then we'll choose the simpler...

Sheez, when has that even ever occurred?

The much vaunted anthropocentric principle, in its weak form, has had one success - a resonance of Carbon 12 was predicted on the basis that otherwise helium could not be fused into carbon in stars and we couldn't exist.

I'm not aware of Occam's razor ever being used in the way I described. More plausible is in the way paradigms die. Classical mechanics started to need too many fudges to explain relativistic effects. Einstein's theory of relativity was adopted pretty quickly because it wasn't contrived, really the only basis for this is Occam's razor. It had become the simpler theory. A similar thing happened with geocentrism when Galileo proposed his heliocentric theory. To properly map the movement of planets in the sky a geocentric model has to have the planets undergo retrograde motion, which there is no good explanation for in the model. The heliocentric theory doesn't have this problem so it is the simpler theory. I don't think anyone at the time consciously invoked Occam's razor, but in accepting the heliocentric model over the geocentric one they we basically doing just that, it is the simpler explanation. One way of looking at Occam's razor is as a description of the behaviour of scientists rather than an actual theoretical tool.

Originally posted by DeepThought... One way of looking at Occam's razor is as a description of the behaviour of scientists rather than an actual theoretical tool.[/b]

The only place I've seen the razor used is as an argument against a creator god. I'm atheist, but I value critical thinking more that I value scoring a cheap shot.

Goddidit. Cannot get simpler or involve fewer assumptions than that!

And so the razor is useless.

oh god, forgive me that cheap shot

Originally posted by apathist
Goddidit. Cannot get simpler or involve fewer assumptions than that!

And so the razor is useless.

No; oc-razor works against Goddidit!
That is because Goddidit is not one assumption but many; for Goddidit there has to also be a God which is an object not with just once defining attribute but many; It assumes there exists an object we call God that, depending on your exact religion, has a list of attributes attached to it similar to this one;

1, is the creator of the whole entire universe
AND
2, is conscious with mind
AND
3, is all powerful
AND
4, is all loving (despite letting some very young children suffer horribly for no apparent reason)
AND
5, is supernatural (and this in turn assumes the supernatural exists which in turn assumes natural law can be violated)
AND
6, depending on your religion; came down to Earth and represented himself as Jesus to talk to us.
AND
7, depending on your exact religion; made the whole universe including the Earth only a few thousand years ago.
AND
8 rewards you in heaven for being good (this in turn assumes the afterlife which in turn assumes the supernatural which in turn assumes natural law can be violated)
AND
9 punishes you in hell for being bad (this in turn assumes the afterlife which in turn assumes the supernatural which in turn assumes natural law can be violated)
AND
10, wants you to believe he exists else punishes you for not doing so (despite being all powerful and yet not showing evidence of his existence to you)
AND
11, be motivated to do that particular Goddidit
...etc...

And thus, because of all those assumptions and the ANDs above, it only takes ONE of those assumption to be wrong for that particular Goddidit to be wrong thus, according to oc-razor, that Goddidit is highly assumptive and should be dismissed as such; oc-razor passes that test and is vindicated!

Goddidit. Cannot get simpler or involve fewer assumptions than that!

you couldn't be more wrong. See above lists of assumptions Goddidit generally composes of.

Originally posted by humy
No; oc-razor works against Goddidit!
That is because Goddidit is not one assumption but many; for Goddidit there has to also be a God which is an object not with just once defining attribute but many; It assumes there exists an object we call God that, depending on your exact religion, has a list of attributes attached to it similar to this one;

1, is the c ...[text shortened]... te]
you couldn't be more wrong. See above lists of assumptions Goddidit generally composes of.

Einstiens Pantheistic God needs none of those assumptions.

Originally posted by joe shmo
Einstiens Pantheistic God needs none of those assumptions.

correct.
However, the overwhelming majority of people, including myself, would say what they mean by the word "God" is certainly not 'Einstein's God' since what he meant my the word "God" was merely "everything" or "the whole universe" which isn't at all what most reasonable people mean by the word at all. He could be validly accused of severely abusing the word "God" and he was, to his great announce, often severely and inevitably misunderstood by his unwise use of the word "God" (and nobody is perfect) as most people would say that most of us simply don't mean that by the word "God" at all but mean something entirely different and usually involving the supernatural.

As for how this relates to oc-razor;
Since 'Einstein's God' means merely "everything" or "the whole universe" and doesn't involve the supernatural, and since there exists "everything" or "the whole universe" isn't an assumption, oc-razor favors the 'theory', if you can call it that, that "everything" or "the whole universe" exists (as required). So this isn't a mark against but rather a mark in favor of oc-razor that it favors 'Einstein's God'. It is also a mark in favor of oc-razor that it disfavors the Christian 'God' or any other supernatural or personal 'god' or supernatural 'Goddidit'.

incidentally, don't know if I already mentioned this to you but I am currently writing a book about my current research into a new type of logic (called tie logic) that will solve the problem of induction and mathematically define and vindicate oc-razor (amongst many other things such as revolutionize A.I. and statistical analysis ) and I hope I will get it published some time in the latter half of next year.
I will try and remember to explain to you and everyone here oc-razor properly and much better than what I have done here (I made my explanation here vastly too simplistic) when that time comes and then I am sure you would agree that oc-razor is valid. Just like with most people who doubt the validity of oc-razor, I suspect most of your doubt of oc-razor validity is as a result of several severe misunderstandings of what it actually says; for example, contrary to what many websites and people say, it does NOT say "the simplest theory is correct"!

Originally posted by humy... See above lists of assumptions Goddidit generally composes of.[/b]

As if any scientific assumption doesn't imply an even longer list of assumptions as well.

Plus god could've done it with none of the assumptions you mentioned.

Why should we assume that nature will tend to submit to our least assumptive theories? That belief itself makes a huge unjustified assumption! Reality is what it is, and I see no reason to assume reality is under some sort of compulsion to adhere to our beliefs or desires, period. The razor is misguided.

Originally posted by apathist
Why should we assume that nature will tend to submit to our least assumptive theories? ....

because the least assumptive theory is the most probable.
If we have two exhaustive mutually exclusive theories, theory X and Theory Y, and theory X makes an assumption A that theory Y doesn't make but theory Y doesn't make any assumption that theory X makes, then theory Y is the least assumptive. Therefore, theory Y is the most probable because, regardless of what the probabilities are of each assumption the two theories both make, for theory X to be true, that extra assumption A has to be true and there must be a non-zero probability that assumption A is false. If you just do the maths, you see that means theory Y is more probable.
Oc-razor favors theory Y as required since theory Y is less assumptive than theory X.

Note oc-razor doesn't guarantee the correct theory (yet another common misunderstanding of oc-razor) but merely favors the least assumptive one and thus, occasionally, the more assumptive one will be the correct. Thus oc-razor will occasionally favor the incorrect theory but that doesn't invalidate oc-razor since it doesn't guarantee the correct theory but rather is probabilistic. In that narrow sense, oc-razor is similar to the principle of induction.