Have I been whooshed? If not… here goes!
Assuming that our self-awareness counts for nothing, we’d be relying on the word of something that hasn’t been proven to exist (us) that aliens exist. That’s the same as saying “I know aliens exist because the fairies told me so!”
godzillatemple: Every scientific advance has come about by someone proving that Occam’s Razor had initially lead them to the wrong answer in a specific case. Every single one. How? Because they showed that the old explanations simply couldn’t account for all new data, and that a new explanation was required.
So, how did they search for this new explanation? Did they postulate a million completely new forces, a thousand little green men on Mars, and a heretofore completely unknown branch of mathematics? No. They simply tried to explain what they saw in the simplest possible terms. Then they used that explanation to make a prediction. Often, the prediction was utter crap. They revised the explanation, made a new prediction, ran the tests again, and found that they’d been wrong again. Maybe, after twelve dozen times of being wrong, they made the explanation more complex. Maybe they found a way to make it significantly less complex, instead. But the point is this: They didn’t go off half-cocked and try to create the most complex explanation possible. Instead, they were constructively lazy.
Lazy? Yep. They knew they had to test this new explanation. They knew that every part of it had to hold up under scrutiny. They also knew that life is short, grants run out, and nobody remembers the crackpots. So they tried to get away with as little new theory as possible. So far, that approach has worked astoundingly well.
There’s a principle in engineering, especially software design: When trying to fix something, only make one change at a time. Don’t completely redesign the system until you’ve proven that you need to. There are various technical reasons related to interlocking systems and complexity increasing as an exponential function and so on, but it comes down to laziness: Why change something that still works, and why test more than you need to? Throwing the baby out with the bathwater is arrogant, foolish, and makes you work too hard. Simplicity is our business, and business is good.
How do you know the people around you aren’t cleverly manufactured homunculi built for the sole purpose of deceiving you into believing that the crop circles were made by humans? Worst case scenario of course, but it’s just something to think about.
There’ve been many, many instances where had Occam’s razor been applied to a situation, it would have been wrong. As an everyday tool for situations, it is useful, but I don’t believe it’s close to being infallible. It’s simply a lax guideline as opposed to a hard rule.
Perhaps a better way of looking at it is that the simplest is always wrong, but that it’s the best to work with in the meantime. For instance, suppose you observe that light reflects, refracts, and interferes like a wave. The obvious conclusion is that light is a wave. We know it’s not going to be that simple - what in physics is ever that simple? But until there’s a way to choose between the infinite number of other possibilities, going with the obvious answer seems to work best.
It seems to me that the most important reason for this is human nature. As the evidence mounts that would tend to refute a particular theory; or as the evidence supporting that theory turns out to be non-credible; the people who believe that theory have a tendency to dance, dodge, and weave by modifying the theory. In doing so, they make their theory progressively complicated.
Ultimately, they end up presenting a confusing and convoluted theory requiring conspiracies, ad hoc assumptions, etc. Common sense indicates that such a theory – which is informed more by wishful thinking than reality – is highly suspect. Occam’s razor, and the related notion of “falsifiability” capture this idea and allow us to say “get real.”
**
I’m sure it must happen regularly. But there are so many people out there who believe so many crazy things, I suspect it’s the exception not the rule.
Just MHO of course.
The problem here is that Occam’s Razor does not say “The simplest explanation is always (or usually) right.” I don’t know where this version originated, but it is only a very rough approximation of what William was trying to say. William’s actual words (in Latin, of course) were “Pluralitas non est ponenda sine neccesitate,” or “plurality should not be posited without necessity.” You can find this, and a good short article about its history, online in The Skeprtic’s Dictionary.
What pluralitas non est ponenda sine neccesitate means is that when you have an explanation, you can stop explaining. It doesn’t mean that new evidence is unlikely to or will not appear. It just means that you don’t believe something without a reason. The website I linked to above makes the point that William of Ockham used it to argue against Platonic realism, the idea that reality consists of metaphysical “forms” of which our world is but a shadow. Ockham argued that saying these “forms” exist doesn’t actually tell us anything about the world, so there is no reason to suppose they exist–people believed in them because they wanted to, not because the evidence supported it. Today, much against Ockham’s intent, the exact same arguement is used against God’s existance–it doesn’t actually explain anything. Even most theists (inteligent design proponants excepted) agree that the existance of God fails as a scientific hypothosis–such a belief must rest on something besides natural evidence (for Ockham, Scripture; for most modern theologians, faith). Belief in God may be true, but it isn’t rational. Using this principle to counter claims of extra-terrestrials making crop circles is like using a Buick to swat a fly: it’s overkill, but it works–and it produces the same result as before: People who believe in aliens crop circle makers may be right, but unless they have evidence the rest of us don’t have, they aren’t rational.
Dammit! The link above is to The ** Skeptic’s** Dictionary, although skeprtics may also find it useful. 
Hey – if I misspelled “has” in the title, you’re entitle to misspell “skeptic”.
Speaking of such things, I just realized that I also misspelled the link the www.chirobase.org in my OP…
Make that “entitled” :smack:
Alan Smithee: excellent post. If I may, I’d like to add that I consider Occam’s razor to be specifically aimed at superfluous entities, theoretical elements that don’t really serve an explanatory purpose.
This “academic” completely misses the point of Occam’s Razor. Occam’s Razor can’t be right or wrong and can’t be proven or disproven because it’s not saying that the simplest answer is usually the correct one.
All it’s saying is that “one should not increase, beyond what is necessary, the number of entities required to explain anything.” All it means is that if there is a satisfactory explanation that already exists, then there is no point in making it more complex. If data mining requires non-linear math then it requires non-linear math. It doesn’t prove that OR is wrong because data mining requires something more complex than linear math. You wouldn’t say that OR is “wrong” because finding the area under a curve requires something more complex than simple geometry.
Alot of the confusion around OR is a result of giving common sense a fancy nickname. Maybe what they should have said is “If you strike oil, you can stop drilling.”
Yes, but why shouldn’t one do so? Is it because the simplest answer is usually the correct one? Is it because we don’t want to make additional work for ourselves and should therefore stop as soon as we find a simple explanation, whether or not it is the correct one? Or is it just an a priori principle that everybody should just accept (in spite of the fact that it wasn’t actually put into words until some time in the 19th century)?
In other words, why does everybody think that Occam’s Razor is such a wonderful guiding principle?
Barry
I think you are mis-interpreting what Occam’s Razor means. It doesn’t mean pranksters are any more likely to cause crop circles than plasma vortices. Its simply a preference for a direct approach over a convoluted approach. A better analogy would be two mouse traps - one basic and one a Rube Goldberg machine. Occam’s Razor would favor the former.
Another way to view it is as an esthetic. A simple explanation is more beautiful.
I have never interpreted it as predictive or empiracle.
I lump it with maxims such as “Its better to shoot with a rifle than a shotgun.”
godzillatemple, Occam’s Razor has unfortunately become vastly overrated in the popular imagination. Alan Smithee and KidCharlemagne rightly pointed that out. Occam’s razor is not a theorem or a law or even a ‘wonderful guiding principle’. It’s merely common sense.
Weinberg, in A short history of medieval philosophy (p. 239) describes it accurately and without ado as the ‘principle of parsimony’. There is no sense in complicating a perfectly good theory by adding new entitities to it. So if you add something, it should do something useful.
Let’s take quantum theory (as far as I understand it). What would you say of a theory that, like others, assumes quarks and all those other subatomic particles and stuff, then adds Invisible Pink Unicorns. You ask the scientist: what’s the use? He says, well, I kinda like 'em. At that point you can apply Occam’s razor.
This doesn’t mean, a propos, that the best theory is simple. The simplest theory of physics is that God causes everything. Quantum theory is anything but simple, AFAIK.
Basically if Occam’s razor really was ‘the simplest theory is probably right’ it would of gone out the window a long time ago as for instance I’ve never heard anyone support general relativity over Newtonian gravity, because it’ simpler. What Occam’s razor does i sget rid of unnecessary terms, for example instaed of the famliar F = ma, I could have a new equations: F = ma + - where is MC's constant and the two explantions of force are still exactly equivalent, but Occam's razor tells me that '’ is unnecessary.
I’ll answer the last question first. The invocation of Occam’s Razor (OR) is a great rhetorical strategy that can be used to bludgeon your opponent into intellectual submission with a intimidating term that can be misappropriated to serve just about anyone’s argument. Creationists use it to “debunk” evolution while evolutionists use it to debunk creationism. It also sounds kind of cool.
Other people have already described the use of OR in shaving off unnecessary elements of a theory (something about pink unicorns). In this case, absolutely nothing else is required to explain the phenomenom so anything else is superflous. I call this the “You can stop drilling because you struck oil” use of OR.
There is another for use for OR where a theory doesn’t completely explain a given phenomenom, but adding more elements could be detrimental. This is where OR earns what little respect it deserves. Here is an real life example:
As a commodity trader, I used to create trading systems that would incorporate certain algorithms (hypotheses) that I hoped would explain market behavior just enough to profit from it. I would test and build the system using market data from 1981-1998 and when I felt I had it right, I was allowed to run the system on out-of-sample data (1970-1980) just once. A system’s efficacy is determined how it fairs in the out-of-sample data, not the in-sample data. Now in theory, I could have fit the rules of the system to never have a losing trade in the in-sample data. Let’s say I started with a general, moderately profitable algorithm that said, “Buy X commodity if it makes a 100 day high.” This would be an example of the “simple explanation.” Now let’s say I look over the trades that this algorithm takes and find that I can reduce many of the unprofitable trades by not taking any buy/sell signal if it came on a Friday. If I wanted to get rid of every unprofitable trade I may eventually find myself adding ridiculous algorithms just to screen out a few bad trades i.e., "Don’t buy if the last unprofitable trade was closed out on Wednesday before a holiday weekend. " The problem is, I’m not trying to describe the movement of this sample of prices, I’m trying to make intelligent generalizations about market behavior that will work on any market data, not just the data that I’m currently working with. The fewer the number of rules that result in a profitable system the more likely those rules accurately reflect basic principles of the market rather than the vagaries of the given sample. So if you have two systems that make the same profit but one system has one rule(explanation) and the other has two rules (a more complex explanation), you should prefer the former. A simple explanation has fewer degrees of freedom than a complex explanation and therefore has fewer chances of being right simply by chance. This doesn’t mean that a more complex system wouldn’t actually be more profitable, but unless their is a significant reason (change in profitability) to add another rule, it should be avoided.
FWIW, I found an interesting article on Occam’s razor on Wikipedia. It might be worth reading just so that we can all agree on what we’re talking about.
I’m not sure I agree with this, KidCharlemagne. That sounds like exactly the sort of misapplication that the researcher in Shalmanese’s link tried to use. The “common sense” application is pretty clearly what Ockham had in mind. What’s hard to realize is that it wasn’t always common sense. Nor is it today! Strict reductionists have used Occams Razor to shave off not just God and objective morality, but all aspects of human conciousness, emotions, and free will. The Skeptic’s Dictionary isn’t exaggerating when it says it’s more like a chainsaw when applied to ontological theories. All rational people will agree that Occam’s Razor is an appropriate tool in some circumstances, but to apply it to your own thinking universally is a) really damn hard and b) leaves you with a thoroughly reductionistic world that most people cannot accept. Berkeley (as pointed out by The Skeptic’s Dictionary again) used it to shave off all of external reality! After all, I don’t need to posit anything but my own feverish and delusionary mind disembodied.
It seems to me that Occam’s Razor needs to be applied here. By applying Occam’s Razor we can see it is clearly simpler to assume Occam’s Razor is a correct principle, as by pointlessly multiplying the entities explaining Occam’s Razor we needlessly complicate Occam’s Razor.
One problem I have with Occam’s Razor is that it can be difficult to objectively assess which of two theories is simpler. So yeah, it’s easy for people to use it as a rhetorical device.
In my humble opinion, the concept of falsifiability is a more useful criterion. In many debates, one can ask “What evidence would cast serious doubt upon your position?”
If one theory is difficult or impossible to falsify and a competing theory is easy to falsify, it seems to me that the latter theory is preferable.