Occam’s Razor

[RichardR]

A criticism of Occam’s Razor is that it doesn’t necessarily lead to the right conclusion - it’s not infallible. Of course, that’s correct. But is that a reason to ignore it? Surely if you apply Occam’s Razor, it just means you don’t invent unnecessary explanations? So only by ignoring the razor are you taking an additional step – inventing something. And why take that additional step?

To ask the question a different way - for what reason would any skeptic ignore Occam’s Razor when evaluating a claim?

 

[MRC_Hans]

Occam's razor is to be applied after evaluating all evidence. I don't think one should ever ignore it, but it will often be wise to validate it. If the second-least complex explanation is not very unlikely, it is a good idea to try and investigate it.

Hans

 

[athon]

The problem is, science is not about finding truth. You can't find the 'right' model. It's a method that endeavours to identify the model that has the best probability of describing outcomes. It's about predicting likely effects.

Occham's Razor does this. It suggests the most likely model of a selection based on the premise that a model which uses the most evidence with the least speculation is more useful in predicting outcomes than any of the rest.

Add in the fact that science is self-correcting, Occham's Razor has a pretty strong place in the scientific method.

Athon

 

[Darat]

Aren't there practical reasons for following a guideline like Occam’s Razor? What it tells us is that if you have two theories, one that can be tested in 30 minutes with a bar of soap and a candle and another that requires 46 years of work with two particle accelerators and a space-based neutrino detector, is "Test the first theory first you idiot!".

This is a reason why when looking at paranormal events it makes much more sense to check the easy to test for stuff first - so instead of measuring slight variations in the surrounding EM field when someone claims they can read without using their eyes Occam's Razor tells us "check they aren't peeking".

When you think about it the million dollar challenge is an example of Occam's Razor applied to the (almost) real world.

 

[Batman Jr.]

A criticism of Occam’s Razor is that it doesn’t necessarily lead to the right conclusion - it’s not infallible. Of course, that’s correct. But is that a reason to ignore it? Surely if you apply Occam’s Razor, it just means you don’t invent unnecessary explanations? So only by ignoring the razor are you taking an additional step – inventing something. And why take that additional step?

To ask the question a different way - for what reason would any skeptic ignore Occam’s Razor when evaluating a claim?

I think the problem is that people use Occam's Razor incorrectly. A common misapplication of it is to invoke it to justify belief in the absence of things unseen. Occam's Razor was never meant to prove the negatives of the unobserved or untestable. Its actual purpose lies in the need to keep unessential elements out of a theory, to keep us from arbitrarily assuming the existence of unknowns. For instance, I think the common Invisible Pink Unicorn demonstration of the Razor is actually wrong because the arguer is using Occam's Razor to say that the Invisible Pink Unicorn actually doesn't exist. What you could use Occam's Razor to say in that particular instance is that there is no evidence of the unicorn interacting with the observable world and thus no reason as of yet to consider its existence or nonexistence at all. In other words, it is, for the time being, irrelevant. This is a better interpretation of the precept because we realize that science is constantly changing and expanding in its knowledge, so a rule which disproved the unknown would have been itself disproved many times over already.

 

[c4ts]

A criticism of Occam’s Razor is that it doesn’t necessarily lead to the right conclusion - it’s not infallible. Of course, that’s correct. But is that a reason to ignore it? Surely if you apply Occam’s Razor, it just means you don’t invent unnecessary explanations? So only by ignoring the razor are you taking an additional step – inventing something. And why take that additional step?

To ask the question a different way - for what reason would any skeptic ignore Occam’s Razor when evaluating a claim?

It works when you apply it to dual realities.

 

[steenkh]

For instance, I think the common Invisible Pink Unicorn demonstration of the Razor is actually wrong because the arguer is using Occam's Razor to say that the Invisible Pink Unicorn actually doesn't exist. What you could use Occam's Razor to say in that particular instance is that there is no evidence of the unicorn interacting with the observable world and thus no reason as of yet to consider its existence or nonexistence at all. In other words, it is, for the time being, irrelevant. This is a better interpretation of the precept because we realize that science is constantly changing and expanding in its knowledge, so a rule which disproved the unknown would have been itself disproved many times over already.

I do not think that Occam's razor can prove or disprove anything. I believe it is a guideliune that tells us if we should use our energy on some hyporthesis or not. In the case of the invisible unicorn, we have two hypotheses: One where such beings exist, and one where such beings do not exist. Occam's razor does not prove or disprove either theory, but, considering that the world is perfectly explainable without such beings, it tells us that we would probably waste our ressources on the invisible unicorns.

 

[Batman Jr.]

I do not think that Occam's razor can prove or disprove anything. I believe it is a guideliune that tells us if we should use our energy on some hyporthesis or not. In the case of the invisible unicorn, we have two hypotheses: One where such beings exist, and one where such beings do not exist. Occam's razor does not prove or disprove either theory, but, considering that the world is perfectly explainable without such beings, it tells us that we would probably waste our ressources on the invisible unicorns.

You're missing the point: Judgments of unneeded variables, whether positive or negative, should be omitted. There are actually three competing theories: The world exists and there also exists an Invisible Pink Unicorn of which we have no evidence, the world exists and there does not exist an entity which fits the description of the Invisible Pink Unicorn, and the world exists, period. Obviously, the most parsimonious explanation of things must be the explanation which pertains only to the observed and nothing more, so the third option would be the best. Also, the first two theories would be subsumed into the one of the greatest brevity, so given that any one of these theories is true, the third must be true. You have a 100% chance of being correct if you choose it. In the case that the unicorn exists, the first one would be wrong, but the second and third ones, correct. In the case it does not exist, the first and third would be right and the second, wrong. Both of the theories which make pronouncements about the unicorn have the pontential of being incorrect given that a certain condition of unknown probability occurs; however, there is no case in which the one which remains ambiguous is wrong.

 

[steenkh]

I do not see how you can ever disprove certain things like invisible unicorns or gods. Occam's razor tells us that since the world can be perfectly explained without gods, there is no need to believe gods exits. How can it ever become a proof with 100% certainty? One day somebody might find a way of interaction with gods that noone had noticed before (like if praying could heal).

I am quite certain that Occam's razor is not a 100% proof but a proof with XX% certainty.

 

[athon]

Science is not about proof, though. Proof is a mathematical term.

It's about accumulation of evidence in support of a model. Complicated models are better at addressing complicated observations, however they do this at the cost of needing to address multiple variables. Occham's razor is a way of compromising this; finding the most suitable models that address variables economically.

This is where woo's get all mixed up. Science cannot provide absolute truth. It isn't about that. It is about making models that can predict outcomes with a comfortable degree of certainty. Occham's Razor is a tool with which we can do this and add to that certainty.

Athon

 

[Melendwyr]

There are always an infinite number of possible explanations for any set of data. Occam's Razor makes it possible to select between them.

Additionally, it's just good sense. If you can explain something by making a certain number of assumptions, why would making more or more extravagant assumptions be better?

 

[Batman Jr.]

I do not see how you can ever disprove certain things like invisible unicorns or gods. Occam's razor tells us that since the world can be perfectly explained without gods, there is no need to believe gods exits. How can it ever become a proof with 100% certainty? One day somebody might find a way of interaction with gods that noone had noticed before (like if praying could heal).

I am quite certain that Occam's razor is not a 100% proof but a proof with XX% certainty.

Right. The problem I have with Occam's Razor is when people use it to make probabilistic judgments of things of which there is insufficient evidence to find a probability. In the case of an invisible unicorn, by its definition, the unicorn couldn't be observed. We couldn't figure out its possible causes to induce the likelihood of its existence. The best thing we could do is to say that the existence of the unicorn is irrelevant to us because it is not required to explain anything around us. Belief in it has no pragmatic value.

 

[bruto]

We shouldn't forget in all this that Occam himself was a theologian too. He applied his principle of parsimony in more than one way. When he cut science loose from god he also cut god loose from science. God or the Spaghetti Monster or that pink unicorn become articles of pure faith. Science cannot invoke them, but neither can it address them. The skeptic who uses his wisdom of cause and effect, science and logic, to argue against God is as out of place in Occam's world as the creationist who attempts to stick God into our genes. For Occam, the incompatibility of science and theology not only purifies science, but confirms and demonstrates God's transcendence of necessity.

It's ironic that some of those now trying their hardest to de-positivize science and blunt Occam's razor are those whose theology would best benefit from actually understanding it.

 

[Melendwyr]

Of course science can address it. Things that, by their very nature, cannot potentially be the subject of scientific inquiry, are unreal. There is no meaningful way in which they can be said to exist.

 

[bruto]

Of course science can address it. Things that, by their very nature, cannot potentially be the subject of scientific inquiry, are unreal. There is no meaningful way in which they can be said to exist.

I disagree. God cannot be disproven, or proven. He can of course easily be doubted, or denied, for various reasons, including the inaccessiblity of divinity to empirical verification. It seems reasonable to conclude that something so remote and unaccountable can be done without and not missed. However, to suggest that there is no meaningful way god can be said to exist is, to my mind, logical positivist ummm... [which rule is that?]. When my crackpot cousin tells me he believes in Jehovah, I know he is a crackpot, but I also know what he means. His ramblings are devoid of good sense and intellectual rigor, but not of meaning in any really useful common-sense sense of the term.

 

[Festus]

Definition?

Ockham's razor. The rule which states "plurality should not be assumed without necessity". In other words if you can prove a postulate without relying on an assumption, such an assumption should be excluded. Two things of interest here. First, Ockham's razor is a rule or guideline for setting the basis of an argument, the supportive information assumed by the author. Second, nothing in Ockham's razor implies in any way that it is related to the scientific method or that it is appropriate for the evaluation of any phenomenon that can be proven or disproven in fact.
There seems to be a notion that Ockham's razor can be taken to mean that the argument with the fewest assuptions tends to be the more correct. Given the quoted definition above this meaning is an extrapolation of the original that ignores the critical value of necessity.

The following sites have similar references.
skepdic.com/occam.html
wotug.kent.ac.uk/parallel/www/occam/occam-bio.html
math.ucr.edu/home/baez/physics/General/occam.html

 

[bruto]

Ockham's razor. The rule which states "plurality should not be assumed without necessity". In other words if you can prove a postulate without relying on an assumption, such an assumption should be excluded. Two things of interest here. First, Ockham's razor is a rule or guideline for setting the basis of an argument, the supportive information assumed by the author. Second, nothing in Ockham's razor implies in any way that it is related to the scientific method or that it is appropriate for the evaluation of any phenomenon that can be proven or disproven in fact.
There seems to be a notion that Ockham's razor can be taken to mean that the argument with the fewest assuptions tends to be the more correct. Given the quoted definition above this meaning is an extrapolation of the original that ignores the critical value of necessity.

The following sites have similar references.
skepdic.com/occam.html
wotug.kent.ac.uk/parallel/www/occam/occam-bio.html
math.ucr.edu/home/baez/physics/General/occam.html

Nothing in Ockham's razor as stated above implies that it is related to scientific method, etc. but I think if you look a little further into Ockham's philosophy you will find that this was his intention. I grant I have not read the primary sources, but it appears he also stated it more than once, and with slight variations. As far as I've read, it seems clear that he intended quite explicitly to secularize science, and used this principle of parsimony to suggest that if natural explanations suffice, supernatural ones are unneeded. In the process, being a good Catholic scholastic, he used his arguments to exempt theology from both scientific proof and parsimony.

 

[drkitten]

Of course science can address it. Things that, by their very nature, cannot potentially be the subject of scientific inquiry, are unreal. There is no meaningful way in which they can be said to exist.

This is simply wrong. Let me give a counterexample from The Science of Discworld. In this book, which is basically a recaptitulation of the history of the Earth from an outsider's standpoint, the outsiders find that "civilization" (broadly defined) has been invented and re-invented many times throughout history, starting with some statue-constructing crabs in the Carboniferous and extending through tool-using dinosaurs.

Of course, "science" doesn't know about these. Furthermore (and this is part of the authors' point), science would not be expected to know about these civilizations -- the sand-and-spit statues, or the wooden tools, could not possibly survive to the present to be found by modern scientists. But just because we cannot possibly inquire about dinosaur tools would make them no less real.

Of course, this example is fictitious. But that's the point. Suppose that I took seriously these ideas and wanted to investigate them scientifically. I couldn't. But saying "I don't know whether or not any dinosaurs could have made wooden tools" is entirely different than saying "Dinosaurs could not have made wooden tools, because there is no way to subject the question to scientific inquiry."

 

[Festus]

Further Elucidation

I don't agree that Ockham's aim was necessarily to secularize science, perhaps to apply a logical system to the arguments of his day, similar to the work of George Boole.
I have read some secondary sources and it seems like Ockham's Razor may not have been directly attributable to the man himself. Rather it may be that the concept stated however many ways was identified with Ockham in the fashion of an idiosyncrasy. Not that the concept was not his but that it was not formally stated as a rule by Ockham. This is important in that it leaves open what Ockham's intention actually was.
I take the stand that the relative number of assumptions between two opposed arguments is irrelavant to the fitness of either argument.
My opinion is that William of Ockham was more of logician than a scientist and in that light Ockam's Razor is more appropriately applied to a single argument as a measure of the fitness of the assumptions than in evaluating the relative worth of opposing arguments based wholy on the raw number of assumptions. If an argument with many accurate and neccesary assumptions is in opposition to an argument with a single wholly false assumption an incorrect interpretation of Ockham's Razor would validate an argument based on falsity and therefore by nature unsound. I therefore submit that it is more fitting that an interpretation of the Razor that reduces unnecessary assumptions in a single argument would be more useful and in keeping with the nature of Ockham's work than a rule that would accept or reject one of many opposing arguments leaving the opportunity for acceptance of arguments supported by specious or unsound assumptions.
In conclusion I would like to add that rules of thumb such as Ockham's Razor are simply shortcuts to conclusions that would be better served with close reading and critical evalution of the position in all parts and as a whole. Shortcuts may save time but they should not substitute for critical thinking.

 

[Festus]

Unreal oxymoron drkitten?

Drkitten,

"Things that, by their very nature, cannot potentially be the subject of scientific inquiry, are unreal." -Melendwyr

"the outsiders find that "civilization" (broadly defined) has been invented and re-invented many times throughout history"- drkitten

"Of course, "science" doesn't know about these." -drkitten

Your example is self-contradictory. If these 'outsiders' are able to find the evidence of these previous civilisations then the potential exists that 'science' can discover the same.
Melendwyr's statement is correct. If there are no potential means of scientific inquiry into the existence of a thing then for all purposes that thing is unknowable and cannot be said to exist. This is deferenciated from the idea of a thing that is completely unknown yet potentially could be discovered.