Tuesday, August 1, 2017

What is Falsifiability (and Why is it Important)?

            The concept of falsifiability is a scientific idea that often comes up in discussions on the existence of gods. Judging by the reactions I sometimes see, it seems to be a badly misunderstood concept. So I figured I’d put up a short post about what it means.
            Falsifiability, in a nutshell, is the ability to frame a test for a hypothesis that would be capable of disproving it (i.e. proving it false). If an idea is unfalsifiable, that means that no test can be devised that would prove it wrong. Here’s where the confusion often comes in: “unfalsifiable,” does not mean “true,” and “falsifiable,” does not mean “false.” In actuality, many facts about reality that we generally accept as true are falsifiable.
            Let me see if I can illustrate the idea with an easy example. Suppose I had a pebble sitting in the bottom of a bowl of water, and I wanted to investigate why the pebble is sitting at the bottom instead of floating. So I come up with the hypothesis that the pebble must be denser than the water. To check, I take the pebble out of the water, dry it off, weigh it, measure its volume, and come up with a density number. Then I do the same with a volume of the water, and I compare the two densities.
            So what happens if it turns out the pebble really does have a higher density than the water? This doesn’t actually prove the hypothesis. It’s a data point in favor of it, but there still might be other factors besides density that are the actual cause of the pebble sinking. More investigation is warranted, and that’s often the case in science; you rarely, if ever, get to say that a given hypothesis is proven true. All you can do is amass evidence that is consistent with it. In that sense, a positive result in this test isn’t all that important on its own, but rather as a building block to a fuller understanding.
            But there is a more important possible result of the test, and that would be if it turned out that the pebble had a lower density than the water. Because if that were the true, then the hypothesis that the pebble sinks because it has a higher density than the water could not be true. The test would have falsified the hypothesis. We’ve never seen that result, we generally accept that it’s true that objects denser than water will sink, but there is a potential outcome of the test that would tell you that the hypothesis is false. This is what it means for hypothesis to be falsifiable.
            But suppose that I had, instead, come up with the hypothesis that the reason the pebble sinks is because there are invisible, intangible water spirits called Naiads that really love pebbles and want to envelop them. Clearly, testing the density of the pebble does nothing to prove this hypothesis true or false, since any possible result can still be explained as the vagaries of Naiad behavior. But here’s the thing: there may very well be no test at all that could ever disprove the Naiad theory. Because no matter what physical observation you make about what traits result in objects sinking in water, it can always be covered by “that’s just what Naiads like to do.” The Naiad hypothesis is not falsifiable.
            So why should we care? What does it really matter?
            Well, as it turns out, an explanation that is not falsifiable is kind of useless. And that’s because the thing that makes a theory unfalsifiable is the fact that it makes no predictions. Going back to the Naiads above, what does believing in the Naiads actually tell you about whether any given object will float? What does it tell you about anything? Does it tell you that you can expect to see Naiads? No; they’re invisible. Does it tell you that you should expect to touch Naiads? No; they’re intangible. Does it tell you that you should expect dense objects to sink? No; Naiads may just not like any particular object even if they seem to have liked every object denser than water in the past. Does it tell you that you should expect less dense objects to float? No; Naiads really could take a liking to anything. Ultimately, there’s no observation you could possibly make that could not be “explained” by saying that it’s just how Naiads choose to behave in that instance. But, by claiming to explain every possible observation, it actually explains none of them. Not in any useful way, anyhow.
            Are Naiads real? Who knows? They’re unfalsifiable – you can’t prove they don’t exist. But since you also clearly can’t make any decisions about the real world that are predicated on their existence, it doesn’t make much sense to act as if they do. How, in fact, would you behave as if they exist, when you can’t test or observe anything about them to know how you should behave?

            Does that sound like anything else you know?

No comments:

Post a Comment