What Is Science?

Following on from last night’s post, I’ve been wondering today how to distinguish science from non-science. Or maybe how I distinguish science from non-science.

I suppose that one way that I tend to crudely distinguish between the two is by whether a text uses words or numbers. If I open a book and flip through it, and it’s all words, then it’s going to be painting word-pictures, and it may well be a work of fiction. But if it has some numbers in it, and some mathematics, and some graphs, then it’s quite likely to be science. Or something that is having a shot at being science.

For as far as I’m concerned, science begins with measuring things, and measurements always produce numbers – metres, kilograms, seconds, etc -, and when you start manipulating numbers – adding them up, etc – you’re doing mathematics.

And when you’ve produced these measurements – say of gas pressure and volume and temperature – you might find that the numbers have some constant relation to each other, and that Pressure times Volume = Some constant times Temperature ( P.V = R.T, which is one of the Gas Laws). And this is how empirical science is done. You measure things, and you look for regularities in your data. That’s how it was found that the planets go round the sun in constant periods of time, and that their paths are elliptical.

But then, when you’ve got these experimentally-produced laws, you might start wondering why gases and planets behave that way, and you start trying to build models of gases and planets. At which point you’ve started doing theoretical science (or theoretical physics). In the case of planets, Isaac Newton found that if there was a force exerted between two planetary bodies which varied with their mass, and inversely with the square of their distance, they’d go round each other in elliptical paths, just like real planets do. And in the case of gases, someone called Ludwig Boltzmann imagined that gases were made up of atoms all flying around in different directions, and showed that if this was the case, then the pressure they exerted on the walls corresponded to the empirically derived gas laws. And both Newton and Boltzmann used some pretty advanced mathematics to prove these things.

And then, once you’ve built some theoretical model of whatever it is you’re studying in the real world, and your model behaves like the real world does, you can start to use your model to predict how the real world is going to behave – where the planets will be next year, and what the gas pressure will be if you heat it, and so on. I’ve actually built my own (rather clunky) simulation model of the planets in the solar system, and the earth actually did go round the sun in about 365 days.

So, for me, science is about building models, and these models are usually mathematical models. And if you open a physics textbook on any page, you’ll find these models being described in mathematical terms. There are models of light (either as something that travels in straight lines, or which acts like ripples in a pond), and electricity and heat and so on. When Crick and Watson discovered the structure of DNA, they did it by building models of its double helix. Chemists are forever building models of molecules with theoretical atoms held together by theoretical bonds, and their models work just like the real world.

But then, if this is how science is done – measuring things, manipulating numbers, and building models – then, to my mind, whenever someone isn’t doing this, they’re not doing science. They’re painting pictures. Although very pretty pictures sometimes.

So when I read something about, say, Freudian or Jungian psychology, and it’s all words the whole way through, and nothing is measured, and there is no model, then I’m inclined to think: this isn’t science. It may be very interesting, but it’s not science.

And I feel the same about pretty much the same about all the ‘life sciences’.

Take economics, for example. Economists measure all sorts of things – prices, sales, receipts, taxes, interest rates, and so on. So they’re doing the measurement bit. And they’ve even got a few empirically-derived laws. But when it comes to building models of economies, they’ve got lots and lots of different ones. There are Keynesian models, and there are monetarist models, and Marxist models. And they all predict different things, and none of them ever seems to predict them right. The economists all disagree with each other. So it seems to me that none of them have got any really good, working models of economies. And accordingly it seems to me that economics as a discipline isn’t quite a science yet. They’re still guessing. And this is why it’ll be no surprise to me if the whole economy goes belly-up, and everyone starves, because the economists don’t really understand how economies work, and can’t predict what’s going to happen, and thus prevent it from happening.

But even when what ‘scientists’ are doing looks like the best theoretical physics, it still may not be science. Take climate science. That’s theoretical physics on a planetary scale. I’ll bet they use the gas law I mentioned earlier in their computer simulation models. I bet they also use Newton’s gravitational equation too. And lots of other ones. And using these models, they’ve predicted that the addition of carbon dioxide to the atmosphere will result in the atmosphere’s temperature rising. That’s what their models predict. But for the past 15 years or so, the planet actually seems to have been cooling slightly. So it looks like their model isn’t a very good one. And this might be because there are all sorts of processes – like cloud formation – that aren’t very well understood. And this means that the bits of the model that are modelling stuff that isn’t well understood are quite likely to be producing wrong answers which are being added to lots of right answers to produce the wrong answer. And it’s really because all these climate models are at the leading edge of theoretical physics, and if you’re at the leading edge, you’re probably getting things wrong more than you get them right. So climate science is another not-quite-science. They’re nearly there. But not quite yet.

And what about antismoking ‘science’. Well, in my view, it’s nowhere near science at all. Because, in the first place, nothing is measured at all accurately. They don’t really know how many cigarettes anyone has smoked or for how long, or what killed them, or anything. And when they produce their dodgy numbers, and try to derive empirical laws from them, they draw the wrong conclusions. For example, in the very first UK study – the Doll and Hill London Hospitals study – it was found that 99% of the lung cancer patients were smokers, and so they concluded that smoking caused lung cancer. But what they didn’t mention was that 96% of the patients who didn’t have lung cancer were also smokers, and that pretty much everybody in London hospitals back in 1950 was a smoker. It’s quite likely that pretty much everybody was also a Londoner.  And even if smoking didn’t cause lung cancer at all, you’d still expect to have found that about 98% of lung cancer patients would have been smokers. And Londoners. But it’s even worse than that. It’s that they only have one model of cancer causation – that smoking causes lung cancer -, and they never try any other model. They refuse to consider anything else. So it’s no wonder that cancer research seems to have got more or less exactly nowhere over the past 6o or 70 years. And it’s a scandal.

Moving on, somebody wrote in the comments today:

Just as your respect for good judgement has been lost by Sir David Attenborough with his misguided hot air over global warming, our respect for your good judgment is lost with your out dated sophistry on 9/11. Tower 7 is a text book demolition. 6 seconds of video is all you need to absorb and compare to any other controlled demolition.

And I thought about this. It’s not just that people say that Tower 7 is a textbook demolition: they also say the same about the twin towers. But just seeing a video of something falling down doesn’t prove anything either way.

The real question with the twin towers (or Tower 7) is: how would you expect them to fall down? Because clearly this commenter thinks they would have fallen down that way. Maybe he thought they would have fallen down like trees or lamp posts (like the lamp post I knocked down once). Or maybe he thought they shouldn’t have fallen down at all.

And the way that you’d expect them to fall depends on your model (maybe just a mental model) of how they’d fall down. And it might be the wrong model.

So today I started thinking of building a mathematical model of the twin towers, using Newton’s gravitational equations, and the strength and density of concrete and steel, and then flying a plane into one side (which would knock out quite a few columns) and starting a fire (that would weaken many of the rest of them), and watch what happened.

I’ve actually already got some simulation models that do most of the physics. What I haven’t got is code that can model the collisions between objects that fall on top of each other. But that may not be too hard to figure out.

So maybe I’ll do that one day. Because that seems to me to be the right way to go about it. Build a model. Build a model as near you can to the real thing. And see how it behaves. And see if it behaves completely differently to the way the real building behaved. And then do a simulation of a textbook demolition, and see how that behaves.  And, in the end, maybe my commenter will be proved right.

But until I’ve actually built a model, I’m just guessing how the twin towers should have fallen down. And just guessing isn’t science either.