Ockham's Razor

[bunny]

Arising from a recent post I wish to make a half-baked claim on ockham's razor which has rolled around in the back of my mind for a while but never been properly articulated.

I think that ockham's razor is something we have found to be useful in science because it seems like nature is created from a few simple principles combining in complicated ways. That is, we come up with a bunch of explanations and when we look into it deeper the more ontologically extravagant ones fail.

I dont see why this should be grounds for believing that the cause of the universe must be ontologically simple. It seems to me that adopting Ockham's razor is not a starting point - it is a result of seeing it work so often in studying different elements of nature. It doesnt follow to me that it can be extended to the creation of the universe.

[pilliwinks]

I'm not sure anyone's claiming it's simple, just that there's no need to make it more complicated than it is.

Of course if you have unexplained phenomena, then you know your explanation is insufficiently complicated, and you need to invoke Occam's less well known Hair Restorer.

In the case of universe creation, I guess we don't/can't know whether we are invoking explanations that are more or less compex than they need to be, so we should probably use Occam's even less well known Shrug.

[CallMeIshmael]

[ QUOTE ]
I dont see why this should be grounds for believing that the cause of the universe must be ontologically simple. It seems to me that adopting Ockham's razor is not a starting point - it is a result of seeing it work so often in studying different elements of nature. It doesnt follow to me that it can be extended to the creation of the universe.

[/ QUOTE ]


Is OR's often used to explain the universe? Ive never heard it used there before, but Ive also never really read that much about the creation of the universe (beyond the basic BB stuff).


Are you referring more to the origins of life?

[bunny]

Clearly there are people who believe that if the universe can be explained purely scientifically then occam's razor means we should not believe in a God. I dont know that this follows.

[bunny]

No I meant the origin of the universe.

[chezlaw]

[ QUOTE ]
Clearly there are people who believe that if the universe can be explained purely scientifically then occam's razor means we should not believe in a God. I dont know that this follows.

[/ QUOTE ]
Dunno if anyone believes that but if they do their wrong. If you could explain the existence of the universe scinetifically then it just means that the existence of the universe isn't a reason to believe in god.

but that's true even if the universe cannot be explained scientifically.

Ockhams razor is way over-rated anyway. If there's disagreement about two equaly valid explanations then usually both sides can appeal to the razor as which explanation is simpler/elegant is often a matter of opinion.

chez

[bunny]

I think this is not an uncommon belief. I think Copernicus has expressed this view (apologies if I have misunderstood). Something along the lines of:

There is an explanation of the world that has no God in it. There is an explanation of the world including God as a creator. We should prefer the first since it is ontologically parsimonious.

This seems like applying ockham's razor (I still dont know how to spell it) to me.

[atrifix]

[ QUOTE ]
There is an explanation of the world that has no God in it. There is an explanation of the world including God as a creator. We should prefer the first since it is ontologically parsimonious.

This seems like applying ockham's razor (I still dont know how to spell it) to me.

[/ QUOTE ]
This seems right to me. But I don't agree with applying Ockham's razor blindly. What it is is a useful methodological principle, not a self-evident axiom. One could just as easily appeal to the opposite methodological principle (that we should prefer theories that are more complicated rather than less), but practice shows that that is generally a bad idea.

I have no idea how to spell it, either, although I know Occam and Ockham are the two accepted spellings. There are also some others (Ockam, say). Any English scholars?

[bunny]

[ QUOTE ]
[ QUOTE ]
There is an explanation of the world that has no God in it. There is an explanation of the world including God as a creator. We should prefer the first since it is ontologically parsimonious.

This seems like applying ockham's razor (I still dont know how to spell it) to me.

[/ QUOTE ]
This seems right to me. But I don't agree with applying Ockham's razor blindly. What it is is a useful methodological principle, not a self-evident axiom. One could just as easily appeal to the opposite methodological principle (that we should prefer theories that are more complicated rather than less), but practice shows that that is generally a bad idea.


[/ QUOTE ]
This is my point - practice has shown that it is a bad idea when studying nature (as nature has turned out, time and again, to prefer the simple solution). I dont think this gives any evidence that it should be used to decide between two competing theories for the creation of the universe (which I take as being qualitatively different from other processes).

[nietzreznor]

[ QUOTE ]
Arising from a recent post I wish to make a half-baked claim on ockham's razor which has rolled around in the back of my mind for a while but never been properly articulated.

I think that ockham's razor is something we have found to be useful in science because it seems like nature is created from a few simple principles combining in complicated ways. That is, we come up with a bunch of explanations and when we look into it deeper the more ontologically extravagant ones fail.

I dont see why this should be grounds for believing that the cause of the universe must be ontologically simple. It seems to me that adopting Ockham's razor is not a starting point - it is a result of seeing it work so often in studying different elements of nature. It doesnt follow to me that it can be extended to the creation of the universe.

[/ QUOTE ]

Ockham's razor has nothing to do with whether or not explanations for the universe, or anything else, are simple or complicated. It just means that you ought to pick the simpler of competing explanations for a given phenomena, etc., all things considered.

I don't beleive it is basing this on anything empirical, either--it has nothing to do with how simple or complicated the real world is. Physics is growing increasingly complex (to my understanding), yet we should still use ockham's razor. Ockham's razor is like burden of proof--if you explain something using x, y, and z, and i explain the same thing with x, y, and z, but also need assumptions of a, b, and c, then the burden is on me to show why my theory is better or yours is inadequate.

[bunny]

[ QUOTE ]
Ockham's razor has nothing to do with whether or not explanations for the universe, or anything else, are simple or complicated. It just means that you ought to pick the simpler of competing explanations for a given phenomena, etc., all things considered.

[/ QUOTE ]
I dont understand this distinction - perhaps I didnt say it very well but this is what I meant. We should choose the simplest explanation if faced with two competing theories (being a realist I maintain that the reason we should choose the simpler is because it is true, not for some pragmatic reason that it is easier to calculate or something).


[ QUOTE ]
I don't beleive it is basing this on anything empirical, either--it has nothing to do with how simple or complicated the real world is. Physics is growing increasingly complex (to my understanding), yet we should still use ockham's razor. Ockham's razor is like burden of proof--if you explain something using x, y, and z, and i explain the same thing with x, y, and z, but also need assumptions of a, b, and c, then the burden is on me to show why my theory is better or yours is inadequate.

[/ QUOTE ]
The reason physics is getting more complicated is that we are learning more about it - ontologically it is getting simpler in that more and more effects are explained with fewer and fewer fundamental facts. I think it is justified empirically because the only reason we do it is that it has worked before.

I think it has everything to do with how simple the world is. If our theories are getting simpler and simpler, but the world is more complicated and we are missing the full picture (but getting the right answers) I think that is a bad thing (albeit a bad thing we are currently oblivious to).

[atrifix]

[ QUOTE ]
Physics is growing increasingly complex (to my understanding), yet we should still use ockham's razor.

[/ QUOTE ]I don't think this is true, although I'm not a physicist. It's true that, say, the mathematics that govern physics are extraordinarily complex and difficult. But that's not so much what Ockham's razor is about. Ockham's razor says that we should not commit ourselves metaphysically to entities if the alternative between existence and nonexistence is equally likely (roughly). Thus Einstein's theory of relativity has a very complex geometry but he also explains things in more metaphysically simple terms than Newton did, there is no need to suppose things like 'force'. Similarly for string theory and others. But perhaps I am wrong.

[bunny]

I think you are right and you said it better than I did. The point isnt simplicity of conception or use it's ontological simplicity.

[nietzreznor]

[ QUOTE ]
I don't think this is true, although I'm not a physicist. It's true that, say, the mathematics that govern physics are extraordinarily complex and difficult. But that's not so much what Ockham's razor is about. Ockham's razor says that we should not commit ourselves metaphysically to entities if the alternative between existence and nonexistence is equally likely (roughly).

[/ QUOTE ]

This is basically what I meant, but you said it better.

[Copernicus]

[ QUOTE ]
I think this is not an uncommon belief. I think Copernicus has expressed this view (apologies if I have misunderstood). Something along the lines of:

There is an explanation of the world that has no God in it. There is an explanation of the world including God as a creator. We should prefer the first since it is ontologically parsimonious.

This seems like applying ockham's razor (I still dont know how to spell it) to me.

[/ QUOTE ]

I dont think Ive ever said this, though it sounds like the way my reasoning would go. If theory A has 3 elements that explain a phenomenon, and theory B has the same 3 elements plus god to explain the same phenomenon, then to accept Theory B there needs to be some "redeeming value" for adding to the complexity. (And awe about complexity is not redeeming value).

[bunny]

[ QUOTE ]
[ QUOTE ]
I think this is not an uncommon belief. I think Copernicus has expressed this view (apologies if I have misunderstood). Something along the lines of:

There is an explanation of the world that has no God in it. There is an explanation of the world including God as a creator. We should prefer the first since it is ontologically parsimonious.

This seems like applying ockham's razor (I still dont know how to spell it) to me.

[/ QUOTE ]

I dont think Ive ever said this, though it sounds like the way my reasoning would go. If theory A has 3 elements that explain a phenomenon, and theory B has the same 3 elements plus god to explain the same phenomenon, then to accept Theory B there needs to be some "redeeming value" for adding to the complexity. (And awe about complexity is not redeeming value).

[/ QUOTE ]
Well I'm glad if I put words in your mouth they were at least palatable.

I dont see why this should be the case though when it comes to an attempt at explaining the universe. It seems to me that the reason we apply occam's razor is that it has worked in the past in similar situations (ie when studying phenomena of nature). I dont see why it should necessarily follow for such a different phenomenon as the creation of the universe (which is qualitatively different from all other processes we consider since it didnt occur within the universe).

[chezlaw]

[ QUOTE ]
[ QUOTE ]
I think this is not an uncommon belief. I think Copernicus has expressed this view (apologies if I have misunderstood). Something along the lines of:

There is an explanation of the world that has no God in it. There is an explanation of the world including God as a creator. We should prefer the first since it is ontologically parsimonious.

This seems like applying ockham's razor (I still dont know how to spell it) to me.

[/ QUOTE ]

I dont think Ive ever said this, though it sounds like the way my reasoning would go. If theory A has 3 elements that explain a phenomenon, and theory B has the same 3 elements plus god to explain the same phenomenon, then to accept Theory B there needs to be some "redeeming value" for adding to the complexity. (And awe about complexity is not redeeming value).

[/ QUOTE ]
Yes, if ABC -> something then that's prefered to
ABCD -> something

my point is that rarely happens and ockhams razer is used to decided between:
ABC -> something
XY -> something

where claims of simplicity tend to be arbitary.

The other point is that as an unexplained universe is no reason to belive in god then adding an explanation cannot be a reason to not believe in god. This follows from ockhams razor /images/graemlins/smile.gif

chez

[godBoy]

http://www.xs4all.nl/%7ejax/blog/PrettyBluePlanet.pps
I find this quite powerful, I don't see it as wishful thinking.

It's a powerpoint presentation, so you need to have that installed.

[Copernicus]

[ QUOTE ]
http://www.xs4all.nl/%7ejax/blog/PrettyBluePlanet.pps
I find this quite powerful, I don't see it as wishful thinking.

It's a powerpoint presentation, so you need to have that installed.

[/ QUOTE ]

powerful?

[hmkpoker]

[ QUOTE ]
http://www.xs4all.nl/%7ejax/blog/PrettyBluePlanet.pps
I find this quite powerful, I don't see it as wishful thinking.

[/ QUOTE ]

Well, there sure wasn't any thinking involved.

[MidGe]

[ QUOTE ]
I find this quite powerful, I don't see it as wishful thinking.

[/ QUOTE ]

Yes, just a pity that it so one sided and restricted to the aesthetics. It shows nothing of the pain at ground level. It is a way of dehumanising the experience in the usual fashion of those that want by all means to justify their belief in a "good" god responsible.

[Chips_]

I think you are on the right track bunny. I had similar thoughts when the movie Contact starring Jody Foster popularized the idea of applying Occam's razor to the idea of God and creation.
Actually just removing the whole question of God behind creation, the history of our understanding of the cosmos has really been to move to more and more complex (and accurate) understandings of the real scene. At first we think the moon+sun both go around the earth because that seems like the simplest explanation. Then we find out about the solar system, the planets, come up with theoires to explain why the same face of the moon always faces the earth, explain the perihelion shift of mercury with the help of Mr Einstein etc. All would seem increadibly complex to the person looking up in the sky in 2000 BC. The thing is that we may be at about the equivalent of 2000BC in understanding the mechanics of how the Universe was created. So Occam's razor is better applied to other things I think. It's not a good argument either for or against a creator of the Universe in my view.

[guesswest]

OR is useful, it's just important to remember it's inductive, so it'll never be conclusive. Inductive reasoning is plenty useful, it's how we conclude the sun will rise tomorrow (or at least was how we got to that conclusion for most of human history), but it always has the admitted potential to end up getting it wrong when a new instance of the same phenomena arises. It has no place in a 'proof', except as an intermediary step, and was never intended for use in that way.

If I observe that my watch has a digitial display, then see another watch that does, and another - OR tells me all watches have digital displays.

[chezlaw]

[ QUOTE ]
If I observe that my watch has a digitial display, then see another watch that does, and another - OR tells me all watches have digital displays.

[/ QUOTE ]
Is that OR?

I thought OR is a philosophical device to decide which explanation of 'all watches having digital displays' is more reasonable

Inducing that 'all watches have have digital displays' is science (of a sort).

chez

[guesswest]

My understanding of OR is something like - if you have to assume new evidence for which there is no current indication in order to explain an alternative hypothesis, the working assumption should be that such evidence isn't available and thus the hypothesis should be rejected?

I'll admit to having forgotten the original wording of OR - but that's how I've seen it applied in philosophy, in support of idealism for example.

I await correction /images/graemlins/smile.gif

[chezlaw]

[ QUOTE ]
My understanding of OR is something like - if you have to assume new evidence for which there is no current indication in order to explain an alternative hypothesis, the working assumption should be that such evidence isn't available and thus the hypothesis should be rejected?

I'll admit to having forgotten the original wording of OR - but that's how I've seen it applied in philosophy, in support of idealism for example.

I await correction /images/graemlins/smile.gif

[/ QUOTE ]
I fear another athiest/agnostic debate /images/graemlins/smile.gif

From wiki

[ QUOTE ]
The common form of the razor, used to distinguish between equally explanatory theories, can be supported by appeals to the practical value of simplicity. Theories exist to give accurate explanations of phenomena, and simplicity is a valuable aspect of an explanation because it makes the explanation easier to understand and work with. Thus, if two theories are equally accurate and neither appears more probable than the other, the simple one is to be preferred over the complicated one, because simplicity is valuable.


[/ QUOTE ]
That's how I've always understood it but there's tons more on wiki and elsewhere.

chez

[guesswest]

I was sorta under the impression that was the pop culture misapplication of OR. I may be wrong in that regard though, not claiming to know much about this.

But in either event, I think it'd still apply to the watches example. It'd just be a case of what the question was, I was imagining a question like 'what is a watch?' -> 'something with a digital display (amongst other things obviously)'.

[chezlaw]

[ QUOTE ]
I was sorta under the impression that was the pop culture misapplication of OR. I may be wrong in that regard though, not claiming to know much about this.

But in either event, I think it'd still apply to the watches example. It'd just be a case of what the question was, I was imagining a question like 'what is a watch?' -> 'something with a digital display (amongst other things obviously)'.

[/ QUOTE ]
I think that's the ockham version of ockham's razor /images/graemlins/smile.gif

I can't see how it applies to the watch example. Generalising from all observed watches to all watches is not what OR allows. To apply OR we need two competing theories that explain all the data (all observed watches are digital) and then to argue that one is simpler than the other.

chez

[guesswest]

We automatically do have two competing theories though, that watches are something with a digital display or that a watch can be something with a digital display or some other kind of display additionally - my understanding of OR is that it'd dictate we accept the former since all evidence indicates that, ie the existence of non-digital watches would be an unwarranted assumption.

I possibly chose a bad example because of the semantic issue of knowing the word 'watch' to start with - take a new noun, a zigazigahhh. All the zigazigahhh's we see have tails, so OR doesn't allow for the assumption that zigazigahhh's without tails might exist.

[chezlaw]

[ QUOTE ]
We automatically do have two competing theories though, that watches are something with a digital display or that a watch can be something with a digital display or some other kind of display additionally - my understanding of OR is that it'd dictate we accept the former since all evidence indicates that, ie the existence of non-digital watches would be an unwarranted assumption.

I possibly chose a bad example because of the semantic issue of knowing the word 'watch' to start with - take a new noun, a zigazigahhh. All the zigazigahhh's we see have tails, so OR doesn't allow for the assumption that zigazigahhh's without tails might exist.

[/ QUOTE ]
OR doesn't resolve this (at least not simply):

Theory A: All ziggies have tails
Theory B: Some ziggies might not have tails

In both theories there is an assumption.
A: we haven't missed tailless ziggies
B: we might have missed tailless ziggies

To prefer one with OR we need to show that A or B is simpler in some way. This may be possible with much more detail but in general neither theory wins without a fight.

chez

[guesswest]

Well, I think assumption B would be more like 'tailless ziggies exist', just to rephrase more forcefully, so I think it'd be the weaker theory by OR since it's an assumption with no evidentiary basis.

But if not, it's just another bad example - let me reformulate my original point about induction, because I'm noticing I expressed it horribly badly. Take two competing explanations of '11':

A: 10+1
B: 2+3+4+1+1

OR says absolutely nothing about either being true or false per se, it just says B should be rejected in favor of A because it's more likely I made a mistake in B, since there were more steps involved. OR is in that regard just some insulation from the skeptical attack in the form of the argument through error - it proves (and claims to prove) absolutely nothing, it just factors likelihood of arriving at the correct explanation bearing in mind a fallible calculator. And that's what I meant RE that definition too - OR does not say that simplest solution is the best because simple solutions inherently have more validity, it's just because we're less likely to screw up with less factors involved.

To that end OR in practical application manifests as a vanishing point, the simpler the explanation the less the likelihood of error the more likely the conclusion is to be valid, up until the point where x->x - in the same way generalizing from particulars in other kinds of formal argument approaches the truth the greater the number of instances, up until the point where induction is indubitable when assessing every particular. That reasonable?

[chezlaw]

[ QUOTE ]
A: 10+1
B: 2+3+4+1+1

OR says absolutely nothing about either being true or false per se, it just says B should be rejected in favor of A because it's more likely I made a mistake in B, since there were more steps involved. OR is in that regard just some insulation from the skeptical attack in the form of the argument through error - it proves (and claims to prove) absolutely nothing, it just factors likelihood of arriving at the correct explanation bearing in mind a fallible calculator.

[/ QUOTE ]
Seems fair enough. I don't think the value if there is one, in OR is avoiding mistakes. The impossibility of avoiding mistakes isn't really a skeptical argument - skeptical arguments are of the form that even if a method is applied perfectly the answers are not necessarily true. (that's a digression but I ignored it once before and can't resist this time).

Sometimes methods that avoid mistakes are much more complicated - double entry book-keeping springs to mind.

If there is value in OR then its because there's something inherently better about simpler theories. It could just be that we find them appealing but I think the value is that simpler theories focus on the nub of the issue allowing better understanding and more chance of extending the theory.

chez

[atrifix]

I don't agree. OR says that other things being equal, the simpler theory should be preferred. A prehistoric astronomer can't predict the orbit of Alpha Centauri with nearly as much accuracy as a 4-dimensionalist can. And, other things being equal, metaphysics is growing increasingly simple, not more complex--thus Einstein is able to get rid of extraneous things like Newton's "force that pulls from afar".

Consider two explanations of a phenomenon:
creation->x
God->creation->x (can be rephrased as God & creation->x)

If both have equal justification, OR says you should reject the latter.

[atrifix]

This is not my interpretation of OR. In particular, I think that applying OR to A and B is a misapplication of the principle, because neither A nor B posits more entities (assuming that both structures are the same). I don't think OR is used because we are less likely to make mistakes, but rather because simpler ontologies have generally shown to be right more often.

[chezlaw]

[ QUOTE ]
This is not my interpretation of OR. In particular, I think that applying OR to A and B is a misapplication of the principle, because neither A nor B posits more entities (assuming that both structures are the same). I don't think OR is used because we are less likely to make mistakes, but rather because simpler ontologies have generally shown to be right more often.

[/ QUOTE ]
I agree with that but except for very simple cases where something extra that adds nothing can be shaved, its not at all clear which ontologies are simpler.

In your Einstein vs Newton example its not at all clear that Einstein's theory is simpler. It seems more elegant but its all a bit touchy-feely. If they both predicted the same evidence then it would be tough to demonstrate that Newton's was more complicated.

I agree with your point that its a very useful working methodology. Trouble is people keep using it to insist that their theory is better than someone elses and it's rarely a justified use.

and its Ockham, named after William of Ockham. Occam is a programming language.

chez

[atrifix]

[ QUOTE ]
I agree with that but except for very simple cases where something extra that adds nothing can be shaved, its not at all clear which ontologies are simpler.

[/ QUOTE ]I agree, but a lot of simple cases are not trivial. For example, a typical overdetermination argument in philosophy of mind goes something like this: physical interactions are sufficient to explain events. Nonreductivists posit that there are also mental interactions which are sufficient. We have no physical evidence of mental (extraphysical) phenomena, and both have the same predictive power; thus by OR we should reject nonreductivism.

[ QUOTE ]
In your Einstein vs Newton example its not at all clear that Einstein's theory is simpler. It seems more elegant but its all a bit touchy-feely. If they both predicted the same evidence then it would be tough to demonstrate that Newton's was more complicated.

[/ QUOTE ]
Okay, I'll give you that one. Take string theory, though. The mathematics that goes into it is exceptionally complex, but it reduces all of these other entities--minds, bodies, atoms, quarks, gluons, bosons--to strings. If both theories have the same predictive power, OR says that string theory is more likely to be correct in the future.

[ QUOTE ]
I agree with your point that its a very useful working methodology. Trouble is people keep using it to insist that their theory is better than someone elses and it's rarely a justified use.

[/ QUOTE ]
I think in some cases it's very useful, and in some cases it's pretty worthless. /images/graemlins/wink.gif

[ QUOTE ]
and its Ockham, named after William of Ockham. Occam is a programming language.

chez

[/ QUOTE ]
That's what I've been using for several years, but my two best friends Google and Wikipedia disagree with me. I'm sure some English scholar knows the etymology. (edit: apparently, "Occam" is the Latin spelling of the English "Ockham").

[guesswest]

[ QUOTE ]
Seems fair enough. I don't think the value if there is one, in OR is avoiding mistakes. The impossibility of avoiding mistakes isn't really a skeptical argument - skeptical arguments are of the form that even if a method is applied perfectly the answers are not necessarily true. (that's a digression but I ignored it once before and can't resist this time).

[/ QUOTE ]

That's not really accurate RE skepticism. There are as many formulations of the skeptical attack in the world as there are starbucks, but they primarily take two forms. There's the attack from below, which asks 'how do you know that' and receives the response 'because of/with reference to x', and so on in regress, until there's nothing to base a position or method on. And there's the attack from above, which holds that we've applied the same method before in similar circumstances and made mistakes, so there's no way of guaranteeing we haven't done the same this time. I personally think they're both compelling, but the latter is if anything the strongest in so far as the former requires a rejection of some form of foundationalism - the latter requires nothing.

[ QUOTE ]
Sometimes methods that avoid mistakes are much more complicated - double entry book-keeping springs to mind.

[/ QUOTE ]

I'm not going to go out on a limb and say there are no examples out there in support of that suggestion - but the one you chose isn't one. Double entry book keeping is just a case of repeating the same process twice.

[ QUOTE ]
If there is value in OR then its because there's something inherently better about simpler theories. It could just be that we find them appealing but I think the value is that simpler theories focus on the nub of the issue allowing better understanding and more chance of extending the theory.

[/ QUOTE ]

I don't think that's right, or even internally consistent. I wouldn't have been sure of that initially, but having just done a bit of reading on OR I'm fairly convinced OR is just meant procedurally. It's not an assertion that a simple explanation is more valid necessarily, it's an assertion that it's more 'likely' to be valid - and the only explanation for a quantitive idea like 'likely' is the removal of opportunity for error. If it's anything else but capacity for mistake, why is nobody able to say what it is that makes a simple theory better or more valid?

[chezlaw]

[ QUOTE ]
[ QUOTE ]
I agree with that but except for very simple cases where something extra that adds nothing can be shaved, its not at all clear which ontologies are simpler.

[/ QUOTE ]I agree, but a lot of simple cases are not trivial. For example, a typical overdetermination argument in philosophy of mind goes something like this: physical interactions are sufficient to explain events. Nonreductivists posit that there are also mental interactions which are sufficient. We have no physical evidence of mental (extraphysical) phenomena, and both have the same predictive power; thus by OR we should reject nonreductivism.

[ QUOTE ]
In your Einstein vs Newton example its not at all clear that Einstein's theory is simpler. It seems more elegant but its all a bit touchy-feely. If they both predicted the same evidence then it would be tough to demonstrate that Newton's was more complicated.

[/ QUOTE ]
Okay, I'll give you that one. Take string theory, though. The mathematics that goes into it is exceptionally complex, but it reduces all of these other entities--minds, bodies, atoms, quarks, gluons, bosons--to strings. If both theories have the same predictive power, OR says that string theory is more likely to be correct in the future.

[ QUOTE ]
I agree with your point that its a very useful working methodology. Trouble is people keep using it to insist that their theory is better than someone elses and it's rarely a justified use.

[/ QUOTE ]
I think in some cases it's very useful, and in some cases it's pretty worthless. /images/graemlins/wink.gif

[ QUOTE ]
and its Ockham, named after William of Ockham. Occam is a programming language.

chez

[/ QUOTE ]
That's what I've been using for several years, but my two best friends Google and Wikipedia disagree with me. I'm sure some English scholar knows the etymology.

[/ QUOTE ]
The rest is all trivial so lets deal with the tough bit.

From wiki:
[ QUOTE ]
Occam's razor (also spelled Ockham's razor) is a principle attributed to the 14th-century English logician and Franciscan friar William of Ockham.

[/ QUOTE ]
The simplest explanation is that Ockham spelt ockham ockham not occam. So by ockhams razor the original spelling is ockham.

chez

[CallMeIshmael]

nietzreznor,

In general, nice posting.


Specifically, Im one of the few who understands the greatness that is:

Loc: are you a two-jar slave?


One of the all time funniest things ive ever seen.

[bunny]

[ QUOTE ]
This is not my interpretation of OR. In particular, I think that applying OR to A and B is a misapplication of the principle, because neither A nor B posits more entities (assuming that both structures are the same). I don't think OR is used because we are less likely to make mistakes, but rather because simpler ontologies have generally shown to be right more often.

[/ QUOTE ]
This is essentially my point - our reason for using Ockham's razor is that it has worked before. However, I would claim that it is only justification for using it in similar situations (ie in investigating the natural world). I dont think whether there is a creator or not fits into that category - I think it is a question outside of nature and inaccessible to science.

[chezlaw]

[ QUOTE ]
Quote:
--------------------------------------------------------------------------------

Seems fair enough. I don't think the value if there is one, in OR is avoiding mistakes. The impossibility of avoiding mistakes isn't really a skeptical argument - skeptical arguments are of the form that even if a method is applied perfectly the answers are not necessarily true. (that's a digression but I ignored it once before and can't resist this time).


--------------------------------------------------------------------------------



That's not really accurate RE skepticism. There are as many formulations of the skeptical attack in the world as there are starbucks, but they primarily take two forms. There's the attack from below, which asks 'how do you know that' and receives the response 'because of/with reference to x', and so on in regress, until there's nothing to base a position or method on. And there's the attack from above, which holds that we've applied the same method before in similar circumstances and made mistakes, so there's no way of guaranteeing we haven't done the same this time. I personally think they're both compelling, but the latter is if anything the strongest in so far as the former requires a rejection of some form of foundationalism - the latter requires nothing.

[/ QUOTE ]
Maybe we're talking a cross-purposes.

We can always say 'you don't know p because you make mistakes' but that's not skepticism (or if it is its a particularly dull version).

We know there's no largest prime number. Saying that we don't know because we might have made a mistake in the proof is kinda childish and not what philosophical skepticism is about.

For a skeptic to undermine knowledge of that type requires a demonstration that deduction applied without error can lead to false results.

chez

[guesswest]

That's not true - you're right of course when you said earlier that skepticism concerns itself with the reliability of processes and their purported conclusions. But it doesn't challenge that reliability based on dogma or assumption, it's based on specific arguments. Which primarily take the two forms I mentioned - either the reliability of a process in terms of it's basis or lack thereof, or in terms of it's tendency to err.

Take for instance the skeptical all-you-can-eat-buffet idea of intermediary sense-data in epistemology. It's attacked from below in terms of it's basis, ie it's questioned to what extent we can verify that sense-data mirrors anything. And from above in terms of our history of error via illusion etc. That's very different to the pedantic (but probably not incorrect) suggestion that we can't know 2+2=4 because we may have added it wrong, the difference being that we can consciously articulate all the factors involved in 2+2.

[chezlaw]

[ QUOTE ]
That's not true - you're right of course when you said earlier that skepticism concerns itself with the reliability of processes and their purported conclusions. But it doesn't challenge that reliability based on dogma or assumption, it's based on specific arguments. Which primarily take the two forms I mentioned - either the reliability of a process in terms of it's basis or lack thereof, or in terms of it's tendency to err.

Take for instance the skeptical all-you-can-eat-buffet idea of intermediary sense-data in epistemology. It's attacked from below in terms of it's basis, ie it's questioned to what extent we can verify that sense-data mirrors anything. And from above in terms of our history of error via illusion etc. That's very different to the pedantic (but probably not incorrect) suggestion that we can't know 2+2=4 because we may have added it wrong, the difference being that we can consciously articulate all the factors involved in 2+2.

[/ QUOTE ]
The skeptical argument against relying on sense data is not that we make mistakes in how we deal with the sense data but that for sense-data to allow knowledge requires knowledge that we are not dreaming (or something similar).

Again someone can just say maybe you've made a mistake and what you decided was an apple was in fact a pear but that sort of mistake is roughly equivilent to adding up wrong.

Edit: I don't mean to suggest that that is the only skeptical argument but I think they all rely on exposing a fundemental flaw in the perfectly applied process rather than making mistakes.

chez

[guesswest]

Right, but assuming you accept for the sake of argument that sense-data is a reliable mirror, what is illusion and hallucination but an instance of error?

The difference is, unlike mathematics etc, we can't deconstruct the process such that we can verify if and where we've erred. The skeptical criticism here does not say that veridical perception is unreliable, rather that we can't ascertain in what instances we've correctly applied it.

[atrifix]

[ QUOTE ]
This is essentially my point - our reason for using Ockham's razor is that it has worked before. However, I would claim that it is only justification for using it in similar situations (ie in investigating the natural world). I dont think whether there is a creator or not fits into that category - I think it is a question outside of nature and inaccessible to science.

[/ QUOTE ]
I can see why you would think that, although I don't know if I am convinced. I certainly agree that OR is not the kind of be-all and end-all that someone might take, say, the law of noncontradiction to be. I don't know whether the fact that the creation of the universe is extraspatiotemporal simpliciter is enough to say that OR is not justified in this instance.

The more typical way a theist is going to attack this objection is chez's, e.g., it's not clear that other things are equal if we get rid of the assumption of God.

[bunny]

Agreed - and that is clearly my view (that universe-God doesnt explain everything that I find so rejecting Ockham's razor isnt necessary to my beliefs). I was more asking whether others would apply ockham's razor in the case of two competing theories for the universe's creation (or, more accurately, how they would justify it). I dont think we're there yet but even if we were, ockham's razor doesnt seem applicable to me. I think we apply it because it has worked before in similar situations.

[chezlaw]

[ QUOTE ]
Right, but assuming you accept for the sake of argument that sense-data is a reliable mirror, what is illusion and hallucination but an instance of error?

The difference is, unlike mathematics etc, we can't deconstruct the process such that we can verify if and where we've erred. The skeptical criticism here does not say that veridical perception is unreliable, rather that we can't ascertain in what instances we've correctly applied it.

[/ QUOTE ]Okay, we're talking about the same thing with different words.

There's two types of mistakes.

Mistake's where the process is used incorrectly (adding up wrongly, seeing JTo and believing JTs)

and mistakes where the process is used perfectly but incorrect conclusions are reached due to limitations of the process (dreaming of a royal flush).

I don't call forming an incorrect belief the second way a mistake but I'm not going to argue I'm right about that.

Back to Ockham and simplicity.

Solipsism seems ontologically minimal to me. That something that appears to be me is doing some sort of experiencing seems certain and I can dispense with everything else.

Does ockham -> Solipsism is most likely to be true.

chez

[guesswest]

[ QUOTE ]
[ QUOTE ]
Right, but assuming you accept for the sake of argument that sense-data is a reliable mirror, what is illusion and hallucination but an instance of error?

The difference is, unlike mathematics etc, we can't deconstruct the process such that we can verify if and where we've erred. The skeptical criticism here does not say that veridical perception is unreliable, rather that we can't ascertain in what instances we've correctly applied it.

[/ QUOTE ]Okay, we're talking about the same thing with different words.

There's two types of mistakes.

Mistake's where the process is used incorrectly (adding up wrongly, seeing JTo and believing JTs)

and mistakes where the process is used perfectly but incorrect conclusions are reached due to limitations of the process (dreaming of a royal flush).

I don't call forming an incorrect belief the second way a mistake but I'm not going to argue I'm right about that.

Back to Ockham and simplicity.

Solipsism seems ontologically minimal to me. That something that appears to be me is doing some sort of experiencing seems certain and I can dispense with everything else.

Does ockham -> Solipsism is most likely to be true.

chez

[/ QUOTE ]

I don't disagree with your initial division, but I don't think perceptual error is necessarily an example of the second kind. The process (veridical perception) is, or at least could be argued to be, reliable. The problem is that we can't ascertain in what instances we identify veridical vs some other kind of perception accurately. That's no more an issue of process unreliability than saying logic is a fallible process when we look at 2+2 and get 5 - it's simply a mistake. You can argue that its a process somewhere higher in a superstructure - ie the process by which we apply logic or somesuch, but I don't think that's a very useful way of looking at it, because that eventually amounts to our entire consciousness.

RE: your last point on solipsism - I can't tell if you're asking or stating?

[chezlaw]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
Right, but assuming you accept for the sake of argument that sense-data is a reliable mirror, what is illusion and hallucination but an instance of error?

The difference is, unlike mathematics etc, we can't deconstruct the process such that we can verify if and where we've erred. The skeptical criticism here does not say that veridical perception is unreliable, rather that we can't ascertain in what instances we've correctly applied it.

[/ QUOTE ]Okay, we're talking about the same thing with different words.

There's two types of mistakes.

Mistake's where the process is used incorrectly (adding up wrongly, seeing JTo and believing JTs)

and mistakes where the process is used perfectly but incorrect conclusions are reached due to limitations of the process (dreaming of a royal flush).

I don't call forming an incorrect belief the second way a mistake but I'm not going to argue I'm right about that.

Back to Ockham and simplicity.

Solipsism seems ontologically minimal to me. That something that appears to be me is doing some sort of experiencing seems certain and I can dispense with everything else.

Does ockham -> Solipsism is most likely to be true.

chez

[/ QUOTE ]

I don't disagree with your initial division, but I don't think perceptual error is necessarily an example of the second kind. The process (veridical perception) is, or at least could be argued to be, reliable. The problem is that we can't ascertain in what instances we identify veridical vs some other kind of perception accurately. That's no more a process error than saying that logic is a fallible process when we look at 2+2 and get 5 - the error is simply a mistake. You can argue that its a process somewhere higher in a superstructure - ie the process by which we apply logic or somesuch, but I don't think that's a very useful way of looking at it, because that eventually amounts to our entire consciousness.

RE: your last point on solipsism - I can't tell if you're asking or stating?

[/ QUOTE ]
I think there's a difference between erroneous beliefs that could in principle have been avoided with more effort and those that are beyond any effort.

[ QUOTE ]
Does ockham -> Solipsism is most likely to be true.

[/ QUOTE ] How can that not be clear? /images/graemlins/blush.gif

I was raising it as a problem for discussion. If it's ontologicaly minimialist then does OR make it more likely to be true?

Just trying to clarify some stuff about OR and simplicity.

chez

[guesswest]

I understood the proposition, it's just the 'does' made it read like a question, but then the lack of a question mark made it sound like a statement.

I'm not sure is the answer. I can see why you'd suggest it, but considering solipsism in the first place requires a fairly in depth epistemological construct, before it crosses anyone's mind - you could argue that everyday experience of a constant reality outside of perception would trump solipsism in application of OR.

[chezlaw]

[ QUOTE ]
I understood the proposition, it's just the 'does' made it read like a question, but then the lack of a question mark made it sound like a statement.

I'm not sure is the answer. I can see why you'd suggest it, but considering solipsism in the first place requires a fairly in depth epistemological construct, before it crosses anyone's mind - you could argue that everyday experience of a constant reality outside of perception would trump solipsism in application of OR.

[/ QUOTE ]
The /images/graemlins/blush.gif was me not you. My statement made no sense.

The trouble I have with OR is that there seems no criteria for simplicity. Ontological minimialism has been suggested and I'm trying to see if that holds water and I don't see that it does.

Your response that 'it could be argued ...' is the problem. My fear is that it can always be argued ... in which case how can OR ever be objective.

chez

[guesswest]

I'd agree - it's only really appropriate to apply OR in formal logic. You can compose symbolic arguments for epistemology questions and the like, but they'd get extraordinarily complicated and it's likely people would end up disagreeing on the terms.

[atrifix]

[ QUOTE ]
The trouble I have with OR is that there seems no criteria for simplicity. Ontological minimialism has been suggested and I'm trying to see if that holds water and I don't see that it does.

Your response that 'it could be argued ...' is the problem. My fear is that it can always be argued ... in which case how can OR ever be objective.

chez

[/ QUOTE ]

I think ontological minimalism doesn't hold water only because it fails to explain phenomena as well as a nonminimalist position. I don't think that it has to do with a problem in the application of OR.

[chezlaw]

[ QUOTE ]
[ QUOTE ]
The trouble I have with OR is that there seems no criteria for simplicity. Ontological minimialism has been suggested and I'm trying to see if that holds water and I don't see that it does.

Your response that 'it could be argued ...' is the problem. My fear is that it can always be argued ... in which case how can OR ever be objective.

chez

[/ QUOTE ]

I think ontological minimalism doesn't hold water only because it fails to explain phenomena as well as a nonminimalist position. I don't think that it has to do with a problem in the application of OR.

[/ QUOTE ]
I think that we then restrict OR to scientific theories(which isn't a problem). I saw your string theory example but that's way beyond me. How about these:

Newtons particle theory of light vs huygens wave theory.

Einsten's cosmological constant

Force being required to keep something moving vs it not being needed.

putting ourselves in the shoes of those considering these problems before they were resolved by different predictions, can we wield OR?

chez