On Science and Theories

I was reading this excellent post over at House of Eratosthenes and I had to add my two cents. And after the fact, I thought my two cents would make a pretty good post on its own. Seeing as how I've been too busy -- and a bit weary of blogging, I might add -- in the interest of efficiency and not wasting work, here's my two cents.

Morgan in his post asks a question about scientists who basically agree on one thing disagreeing on another. That got me to thinking. Here goes:

As far as the scientists disagreeing, scientists do disagree on things. One good scientist can even disagree on something with another good scientist as well. That’s because scientists come up with theories, and then they devise tests to see if the theory holds.

They disagree on the theories and the validity of the tests for those theories for various reasons.

Here’s the sticky thing — a theory can never really be “proven” — it can only be disproven. What can be done is to show that it holds true again and again in a given set of circumstances. And when this is shown, the theory becomes useful. But all it takes is for one case to fail for the theory to be disproven. Still, that doesn’t necessarily mean that the theory is useless. Often the scientist will then come up with another theory as to why the first theory failed in that case. And then he or she tests that theory. See how this works?

If the second theory appears to hold again and again, the first theory can then be modified to account for it. But even then they are both still theories.

Another problem that can come up is that a test or series of tests might not be appropriate — and therefore are invalid, for the theory being tested. For example, “did those two ever fit together?” is a valid test to test a theory that CO2 levels are related to temperature changes. However, it is not a valid test to see whether or not CO2 levels cause temperature increases, partly because it might just be the other way around. (Now how do those two fit together and what is the relationship between the curves is a more useful — and, as it turns out, revealing — approach).

There’s another theory based on blackbody radiation theory that predicts that more CO2 will produce warmer temperatures. But it, too, is a theory. A theory which must be tested before we accept it as a useful predictor.

The question you have to ask yourself is… how has this theory been tested? Followed by, are these tests valid to show that the theory holds? And everybody’s favorite followup question is why or why not? Which basically means, “defend your answer.”

At any rate, it turns out that the CO2 –> Warming is a pretty hard thing to test on the scale we’re talking about. About the best we can do, to the extent we can do it, is go back and try to re-construct temperature and CO2 records throughout climatological history and closely observe their relationships. If they match excactly, that’s not very helpful — because it doesn’t hint at what causes what. But if they don’t … if one increase leads the other by any sort of significant amount, then you get an indicator of which one might be the cause and which one might be the effect.

And as it turns out, when we do this, we see that temperature goes up first and CO2 levels follow.

Why?

Well, there’s a theory that says that cold water absorbs more CO2 (and other gasses as well, but we’re talking about CO2 here) than warm water. And in experiment after experiment, this has never been shown not to be the case. So there’s a theory that says that the increased CO2 in the atmosphere is a result of warmer water (mostly ocean) surface temperatures because the warmer ocean can’t hold as much as it had been holding — so it must release it as gas.

Sounds like a pretty good theory. And whenever we dissolve CO2 in cold water and warm it up, it releases some. Like clockwork. Every time. That’s how real science works. Not by “graduate students, political bodies, and comedian's wives all agree”.

Science is most certainly not about voting.