Science and truth

Many years ago, a work colleague described being at a dinner party where he heard a remarkable tale. The storyteller had heard that someone was in the centre of the smallish town of Sunningdale and saw a man walking his dog. The town centre is unusual in that a railway line runs across the main road, so there’s a level crossing there. The dog walker was picking up a takeaway from the tandoori restaurant right by the level crossing. He had clearly not thought this through before setting out. He couldn’t take the dog into the restaurant while he collected his food. So he tied the dog’s lead to the level crossing barrier and nipped inside. He was only going to be gone a minute.

What the dog walker hadn’t noticed is that the barriers were in the down position because a train was about to pass by. Moments later, the train rumbled through and the barriers opened. Leaving the dog dangling from its lead high in the air, to the bewilderment of the emerging dog walker. Everyone at the dinner party thought this was hilarious. So did my colleague – even more so than the rest of them, for reasons we will explore in a moment.

If you heard that story, would you believe it was true? There’s an old saying in science that the plural of anecdote is not evidence. You need something stronger than a tale or two told at a dinner party to think that something is a fact. Of course, plenty of anecdotes are true, but just because we hear a story like this is not sufficient to start accepting something. No one at the dinner table had seen this happen – the teller of the tale had heard it from someone else. There was no corroborating evidence, such as a photograph of the dangling dog. Again, this wouldn’t enough in itself – it can be faked easily enough – but it would be a contribution to the balance of likelihood. And there was no way of finding out more, especially in those pre-internet days.

As it happens, my colleague knew something the others present didn’t. When he was at a dinner party and the conversation was a little dull, he had a habit of throwing a fictional tale into the mix. Some months before, he had made up the story about a dog being tied to the Sunningdale railway barrier. This story had spread from person to person – and now he was hearing it back from someone who hadn’t been present when he told it, as if it were fact. His fiction had taken on a life of its own.

When we were thinking about what we need to make it likely that the story was true, we were, in a loose way, indulging in the scientific method. If this had really happened it would have been true. A truth. And assuming we weren’t there, whether or not we accepted that truth would depend on the quality of evidence supporting the assertion.

So isn’t this exactly what Jim Al-Khalili was saying? Why do I consider it odd? This is because science works at two levels, something highlighted in a cruel but amusing fashion by the great physicist Ernest Rutherford, who is said to have remarked ‘All science is either physics or stamp collecting.’ This is clearly intended as a humorous jibe against sciences other than physics – and like most people with a physics background, it makes me grin. But if we take away the subject rivalry, there is a serious underlying point. Science involves two different things. One is to establish what happens. This is the stamp collecting part. It’s vital. This is where we find the facts. The truth about what is observed. But arguably it’s not the interesting part of science. Science only takes off when it attempts to establish how or why something happens. And there, we usually don’t establish a truth, but rather develop a best theory given the present evidence that makes testable predictions and that may change in the future.

Al-Khalili gives as an example of a truth the value of the acceleration due to gravity on the Earth’s surface. This is an easily established fact. Or even simpler, there is the fact that things with a positive weight fall to the Earth if we let go of them in mid-air. But the interesting science is where that value for acceleration comes from, or why things fall in the first place – and there we usually need to have a theory.

Take another example – it is an observational fact that galaxies appear to rotate faster than they should be capable of doing without flying apart. Leaving aside the possibility that there could be an error in the calculation of how fast a galaxy can rotate and remain stable (this has been suggested), it appears to be true. However, what we can’t do is make the leap from this fact to saying that it’s true that dark matter exists. Dark matter is a theory to explain what is happening. As it happens, there are other theories such as MOND based on modified gravity that explain the observations without resorting to the existence of dark matter. Some evidence supports dark matter better, other evidence supports modified gravity. So it is absolutely not a scientific truth that dark matter exists. The stability given the rotation speed is what happens (subject to the proviso above), but dark matter (or modified gravity) is an explanation of why it happens – this is the really interesting science, but it’s not an absolute truth.

Note that this does not mean that we can treat all theories equally. Just as is the case with the dog at the level crossing, we need to examine the evidence. If we aren’t able to do that personally, it makes sense to rely on experts to examine that evidence for us (just as it’s better to have an expert undertake brain surgery on you, or to fly a plane). And in many cases (unlike dark matter), there is a widely supported theory that is the best we have, unless and until new evidence comes up. Human-created climate change is a good example of this. Although it’s fine to continue to look for new evidence which may reinforce or disprove it, the only sensible course is to go with the current theory that is most widely supported by subject experts.

Science, then, is not about uncovering ultimate truths, but concerns something more subtle. And, I would suggest, is all the better for it.