Occam's Razor (pre-emptive thread split)

[Roaring Biscuit!]

I'll be brief on this, and we can split it into it's own thread, but you don't seem to be using the idea of Occam's Razor correctly.

http://wiki.lesswrong.com/wiki/Occam%27s_Razor

However, you're absolutely right if you consider that what politicians are giving you is a false simplicity, and runs entirely counter to Occam's Razor.

Well, I did have an incorrect interpretation of Occam's Razor, but I still dislike it.  Not because it's a bad idea.

In my eyes, the idea that:

When several theories are able to explain the same observations, Occam's razor suggests the simpler one is preferable.

I completely agree with that, but I think there is one cleverly hidden implication, which is really what I was trying (and failing, as usual  ) to express.  It comes down to that last phrase really:

the simpler one is preferable

SIMPLE IS PREFERRABLE

I'd go as far to say that a lot of new scientific theories fall pray to a kind of simplicity halo effect, like when judging the personality of faces we equate beauty with positive characteristics.  Simplicity is dangerously reinforced in a lot of spheres, and it may even be the case (to stretch the meme/gene parallels a touch) that simplicity of memes is selected for in meme transference, for the obvious ease in spreading an idea and in comprehending.

Hell it might even be that our lazy brains select for simplicity, consider the beautiful = good phenomena in face perception.  What is beautiful?  In very broad strokes we're talking about things like skin clarity (textural simplicity?) and symmetry (geometric simplicity?).  Bloody simplicity is everywhere.

Now I'm gonna jump the gun a bit here, and suggest that someone will say something along the lines of "why use a paragraph when you can use a word", which is absolutely right (sort of).  But there's no point in doing that if you're using the wrong word.  So to bring it back to the original subject, Occam's Razor for me represents a subtle trend towards simplicity for the sake of simplicity.

xx
Edd

p.s. @LMNO, it may or may not clarify my point somewhat to know that I study psychology, as such the theories in my mind when talking about this sort of thing tend to be psychological theories, which are tend to be a lot messier than in other sciences.

[LMNO]

Yeah, I DEFINITELY think you're approaching this from a different angle.

Maybe think of it this way.  Occam's Razor says "make the answer only as complicated as it has to be."  I think you're bugging out over the word "simple", maybe adding in the alternate definition of "simple" as "dumbed down".  All it asks is for you to throw away the useless terms.  It's not "choose the answer with the fewest words or ideas," it's "take out all the meaningless shit from whatever answer is working."

[Precious Moments Zalgo]

Yes, what LMNO said.

The word simple, in the context of Occam's razor, is used more in the sense of elegance or clever simplicity, rather than in the sense of being dumbed-down and devoid of complexity.

If several theories are capable of explaining all of the observed phenomena, the one that makes the fewest unnecessary assumptions is preferable.

[Roaring Biscuit!]

I completely agree with that, but I think there is one cleverly hidden implication

Stated again, I understand what Occam's Razor actually is, and why it's good, but it is also a symptom of the simple is good phenomena, which is far more interesting to discuss.  I think there is a general trend in softer sciences, that want to be hard sciences, to look at the simple elegant equations of classical physics and think "If we want to be considered proper scientists we should have simple elegant models" rather having models that are actually accurate/true.

Let's change tack!  Pros and cons of "simplicity = good" meta-meme?

xx

edd

[LMNO]

Linguistics/Semantics.


Ok, fine.  That's usually my default positon.  But hear me out.  First off, it sounds like you're saying the softer sciences are using Occam's Razor horribly incorrectly, in order to make their theories "as elegant" as physics.  Which is something you should never, ever say around a Scientologist, because they'll go batshit insane with glee.  And just because they don't understand what it means, and are distorting words, doesn't make Occam untrue, or unhelpful.

Second, there seems to be another semantic fart when people try to use "simple" as a synonym for "easy".  Because what is more simple?

As an example from Less Wrong, lightning could either be electrical atmospheric discharge, which sounds complicated, or the god Thor, which sounds simple.  But in fact, positing Thor is insanely complicated, because now you have to explain and prove Thor. 

Simply (ha!) put, just because an explanation sounds simple, that doesn't mean it is simple.  Words are incredibly dense signifiers, packed with meaning that most people don't even consider.  Sure, we can sum up a massively fundamental difference in the way the universe has been shown to work by saying "Thor".  But that doesn't mean that it's simpler.  It just means it's shorter in that form.  But once you start unpacking the word "Thor", you'll end up saying, "you know, can we go back to that atmospheric discharge thing?  That's starting to make a lot more sense."

That's what I was saying before about "false simplicities".  They're like icebergs, you only see 10% of them above the water.

Now, the problem you seem to be addressing is the human trait of not wanting to think too hard.  They want to go with first impressions, clear narratives, and obvious themes.  They avoid nuance, they want order not chaos.  That, I believe is partly hard wired into our monkey brains, because to dwell too long on something increases the chances of getting eaten by something large and with a lot of teeth.  So, I don't think that's a meme so much as a good chunk of the OS.

[Adios]

Linguistics/Semantics.


Ok, fine.  That's usually my default positon.  But hear me out.  First off, it sounds like you're saying the softer sciences are using Occam's Razor horribly incorrectly, in order to make their theories "as elegant" as physics.  Which is something you should never, ever say around a Scientologist, because they'll go batshit insane with glee.  And just because they don't understand what it means, and are distorting words, doesn't make Occam untrue, or unhelpful.

Second, there seems to be another semantic fart when people try to use "simple" as a synonym for "easy".  Because what is more simple?

As an example from Less Wrong, lightning could either be electrical atmospheric discharge, which sounds complicated, or the god Thor, which sounds simple.  But in fact, positing Thor is insanely complicated, because now you have to explain and prove Thor. 

Simply (ha!) put, just because an explanation sounds simple, that doesn't mean it is simple.  Words are incredibly dense signifiers, packed with meaning that most people don't even consider.  Sure, we can sum up a massively fundamental difference in the way the universe has been shown to work by saying "Thor".  But that doesn't mean that it's simpler.  It just means it's shorter in that form.  But once you start unpacking the word "Thor", you'll end up saying, "you know, can we go back to that atmospheric discharge thing?  That's starting to make a lot more sense."

That's what I was saying before about "false simplicities".  They're like icebergs, you only see 10% of them above the water.

Now, the problem you seem to be addressing is the human trait of not wanting to think too hard.  They want to go with first impressions, clear narratives, and obvious themes.  They avoid nuance, they want order not chaos.  That, I believe is partly hard wired into our monkey brains, because to dwell too long on something increases the chances of getting eaten by something large and with a lot of teeth.  So, I don't think that's a meme so much as a good chunk of the OS.

This is bordering on genius.

[LMNO]

I can't quite take credit for it; it's extrapolated from some of the Less Wrong stuff.  But thanks!

[Kai]

The original wording of Occam's Razor is "Do not multiply entities beyond necessity". IOW, if a particular assumption is unnecessary for the explanation of a particular phenomenon, it should be eliminated from the explanation.

Example: Angels bowling is not a necessary assumption to explain thunder, as it can be explained adequately by air superheated through high voltage static electricity. Therefore, it can be eliminated.

[LMNO]

Master Kai, I'm glad you entered in to this.

Beyond Occam, RB noticed a behavior for humans to ignore nuance in favor (and often counter to reality) of false simplicity.

Though you don't necessarily subscribe to meme theory, does this sound like a conditioned behavior, or does it seem to be a deeper impulse?

[Cain]

The idea is that a simple, abstracted theory will provide an overall better guide to judgement and behaviour than a possibly contradictory but more complicated and nuanced system would.

I know this because I've been reading Mearsheimer's theoretical work recently, Mearsheimer being a prime offender of someone who is totally wrong on a factual level, but insists he is correct on a higher, theoretical/abstract level.  He argues that we need theories as a guide to action (giving the example of a theory of crossing the road, "you look left, then right, then again, then you start to cross, then you make it to the other side."  Often in reality it doesn't work like that, because of traffic, ice, animals presenting a hazard etc, but that is the theory from which we derive practice).

The problem is this removes what I would consider the most basic element of science - fact-checking.  If a theory, like Mearsheimer's, continues to fail to explain past actions and predict future actions accurately, while insisting it is theoretically correct, it is no better than religion.  He argues that where his theory fails, it is because more nuanced theories (mid-level or micro-level) which complement but do not contradict his overall thesis can explain the phenomena in question, something which is not readily apparent.  He also cherry-picks his examples from the historical data, with no good reason even from his own theoretical standpoint for excluding certain data, and so does not explain the counter-examples to his argument.

He also derives an ought from an is which, as a fan of Hume, amuses and infuriates me.

I think it is hard for a social science to ever be like the hard sciences, because they're much more complicated and we don't have the necessary mathematical or conceptual tools to reduce and express such behaviour at this time (and we may never have).  That said, it shouldn't stop people from trying to adhere to basic methods of science, like observation, including anomalous data, not putting theory before facts and so on and so forth.

[LMNO]

This Mearsheimer guy sounds like a total ass.  Are there any upsides to his work?

[Cain]

Well...you could give his stuff to undergrands to look for the logical flaws.

Sadly, this kind of "reasoning" permeates the social sciences, though usually not in such a flagrantly incorrect manner.

[LMNO]

IT'S ALWAYS SUNNY DURING THE DAYTIME.*













*Except when it rains.**

**Or it's generally overcast.***

***Or if you live in Seattle.

[Cain]

Social scientists seem to have not so much science envy as Nobel Prize-winning Grand Unified Theory envy.

They seem to have forgotten Grand Unified Theories were worked up from centuries of observation and testing, and most science started out as small, bounded theories that gradually fit into a more coherent whole as more commonalities or underlying rules of those minor theories were discovered.

Not by some bloviating asshole with tenure making shit up and then shoehorning reality to fit in with his theoretical construct.

[LMNO]

They also seem to have forgotten that the Grand Unified Theory for physics hasn't been found yet.