Beauty of Gray

Friday, August 16, 2002

Democracy and the Middle East

Well, this is a positive development. The US is now tying additional aid to Egypt to democratic reforms. Hopefully this will be the beginning of a trend—the US is a major aid supplier to several countries in the Middle East, especially Egypt and Jordan. As such, this should give the US substantial leverage to press for reforms in those countries, just as our aid to Israel gives us leverage to pressure their government. However, up to now that particular lever has never been tested.

There are two caveats that have to be mentioned, though. First is that this decision only affects additional aid to Egypt; it doesn’t touch the existing aid of almost $2 billion a year. As such, Egypt is putting on a brave face and vowing they won’t give in to pressure. If not, then perhaps the US needs to start chipping away at that pot of regular aid, actually decreasing it rather than just freezing the level. There might be some diplomatic issues there—I believe the aid was agreed to as part of the Camp David Accords, and I don’t know the exact terms of the agreement.

But one area that the US has consistently fallen short has been holding our nominal allies accountable. It was true during the Cold War, and it remains true in the Middle East—if you are willing to play nice with the US, we will look the other way when you torture and imprison your own citizens. Obviously, in regions where there are no palatable options, if you want any friends you need to make some compromise of ideals to reality, but that doesn’t mean you must simply accept the status quo without exerting some pressure for reform. Hopefully this action is the first step towards a new policy of encouraging reform among Middle Eastern (and other) governments.

The second caveat is more serious. And that is, as mentioned in the article, that Egypt has been an ally in the US effort against terrorism. And the US would prefer not to lose that cooperation. However, as Daddy Warblogs has discussed, the two goals of promoting democracy and opposing fundamentalist Islam may not, in the short term, be compatible. There is no reason to think that democratic regimes in Egypt, Iraq, or Iran would be very sympathetic to the US. The major opposition groups to the Egyptian government comes from hardline Islamicist groups, so chances are any democratic reforms will have the effect of giving these fundamentalist groups greater influence.

However, as I’ve argued before, I think a push for more freedom is, in the long run, the best policy. Right now, the autocratic regimes are quelling any dissent, which tends to channelize opposition to the governments towards the more militant and extreme groups, namely the Islamic parties. This repression also will make the break, when it comes, more violent. Like levees blocking floodwaters, keeping dissent suppressed will make it manifestation that much worse when it finally breaks through, as it did in Iran.

Liberalization and democratic reforms would, first of all, hopefully help drain some of the appeal and power from hardline opposition groups, lowering their chances of ultimately seizing power. While this might not work, at the very least getting the US on the right side of the issue will increase the sympathy of the populace. There are two big issues that alienate many Arabs from the US. The first is the US’s backing of Israel. The second is the US support for the indigenous dictators. The financial and military support the US provides for these autocratic regimes naturally makes any leaders in the street that oppose those dictators anti-US.

Ultimately, if we want to see any long-term reform in the region, we must be willing to risk short term instability and setbacks. Dictatorships don’t transition seamlessly and peacefully to democracies. They didn’t in the west, and they probably won’t in the Middle East. They need to go through their own revolutions to fight through to reform.

The status quo is obviously not a good situation, so simply attempting to preserve it is not an appealing prospect. Once that fact is accepted, the US can hopefully position itself to really be the beacon of and promoter of freedom and democracy that we like to style ourselves as. And in the long run, I have faith that democratic regimes will make peace with the West. People don’t like war, they don’t like being repressed, and they don’t like poverty.

Update: Combustible Boy has also commented on the topic.

Link posted by Doug Turnbull at 12:45 PM

Tuesday, August 13, 2002
Science, Falsification, and Validity

Most people are familiar with the canonical method by which science proceeds—through the proposal of a hypothesis and experimental tests to attempt to disprove it. What’s sometimes less clear is why this is the method chosen—are scientists perverse in that they prefer to disprove things rather than prove them? Leaving aside for a moment the fact that this description of scientific progress is rather idealized and not all that accurate when compared to what many scientists actually do, Demosthenes discusses in a long post why the method of falsification is a better grounds for the advance of knowledge.

He explains it in terms of the idea of satisficing, which is accepting limitations in data and seeking an outcome that is "good enough," rather than optimal. A finite data set doesn't allow you to achieve a perfect optimization--there are always multiple possible courses of action that will satisfy your goals. Applying the idea to the scientific method, this boils down to the fact that no matter how much evidence you have that supports a given hypothesis, this is not enough to absolutely prove that hypothesis is true, because there might be other hypotheses that fit the observed data just as well. On the other hand, a single good piece of data that violates a hypothesis is enough to disprove it.

So if you want to be sure of yourself, the best way to proceed is to come up with as many possible hypotheses that explain the data as you can, and then get more and more data to try and eliminate them. At the end of the day, the hope is that you will find one single theory which fits all the observed data. You don’t stop there, but you can now provisionally accept that theory. It’s not proven true, but it’s proven useful, which is ultimately almost as good.

There remains the possibility that future discoveries will expose shortcomings in your existing theory, which will require its modification or the shift to a new explanation that differs in significant ways. However, this isn’t a wholesale overthrow—a theory only becomes accepted if it is the simplest explanation that fits the data over a wide range.

Normally, new data that contradicts the existing theory comes from a different realm of experience (an obvious example here is quantum mechanics.) The new theory that explains this data is usually more complicated, but still reproduces the existing theory for the previously measured regimes. It must, since the data from those regimes was consistent with the old theory.

Since this is the case, for many applications it will make more sense to use the previous theory, even if it is not completely correct, because it is accurate enough in the realm you’re interested in. (Relativity is another example—strictly speaking, every calculation in mechanics that doesn’t compensate for relativistic effects is wrong. But it’s usually not worth the trouble to make the more complicated calculation when it only changes your answer by 0.00001%.)

Getting back to the idea of satisficing, I’ve seen this concept used as the basis for a critique of science and the scientific method. We’ve taken lots of data, but that data could be fit by an infinite number of hypotheses, so we are unable to draw any conclusions from it. Therefore, science is not a valid method of arriving at truth, all knowledge is contingent, postmodernism reigns supreme, and Toni Morrison is a better author than Tolstoy. While the final conclusion is obviously flawed, I’ll limit myself to the critique of science, which I think fails on two counts.

The first is that I think it radically underestimates the amount of specification that even a relatively limited data set can give. There just aren’t that many hypotheses, in most cases, that are available to explain the data. The great majority of scientific work is in the details of applying relatively well understood broad theories or phenomena to new situations. If you see a novel energy transfer effect in a solid, there are really only a half dozen or so processes that could explain it, each of which is fairly well understood in some contexts. So you’re not starting from a blank canvas with infinitely many potential explanations. Most often, you’re working withing the framework of existing theories, which severely constrain the possible explanations. (And even in completely groundbreaking work, the data tends to be pretty specific. There just aren’t that many different kinds of mathematical functions or differential equations that can be used to model the data.)

As a side effect, this limitation of hypotheses also means that many working scientists don’t have to rigorously look to disprove alternatives—if your choices are relatively limited, looking for evidence in favor of one hypothesis often amounts to the same thing as attempting to disprove all the others. It’s a bit of sloppy thinking, really, but the paucity of possible explanations allows many scientists to get away with it.

The second point was hinted at above in describing the way a new theory supercedes an old one. Science, by its own admission, is not an attempt to arrive at Truth. It’s an attempt to arrive at a valid explanation which has predictive power. That is, I want to come up with a theory that explains existing data and will allow me to predict the result of experiments in the future. A scientific theory only gains acceptance if it works, in this sense. Any alternative hypothesis which equally fit the data is, by definition, equivalent to the accepted theory. They produce the same predictions, and hence are interchangeable. So even if there were infinite alternate hypotheses, they’d all be equally good explanations of the world, rather than being equally bad.

At worst, a scientific theory that has gained acceptance is a really good approximation to the situation in the real world. It makes predictions which are as accurate as we can measure, which is functionally equivalent to a perfect theory. But even once we are able to measure situations where the theory does diverge from measurements, the old theory is still good as a rough (or not so rough) estimate. It doesn’t suddenly become totally wrong; it becomes slightly wrong, but usually still very useful.

In fact, in physics, a lot of work is directed towards generating theories that people know in advance are wrong. A theorist will make lots of approximations—treat all bodies as speheres, all distributions as Gaussians, etc.—in order to try and generate a bottom line result that will give a rough approximation. Because you’ve fudged a lot of details, you know the fit won’t be perfect, but it’s still a victory is you can figure out that, say, the heat capacity of the solid should depend linearly on the temperature, rather than on the square root of the temperature. Sometimes this works, sometimes it doesn’t. Sometimes, the devil is in the details and you can’t get good results by assuming everything is spherical. But sometimes you can, and that’s a victory for science, rather than an evidence of its shortcomings.

Link posted by Doug Turnbull at 11:27 AM

A Renaissance blog: Politics, sports, literature, history, and whatever else strikes my fancy.


Sneaking Suspicions
Crooked Timber
Unqualified Offerings
Asymmetrical Information
Cranky Professor
The Poorman
newsrack blog

Powered by Blogger