You keep saying that word. I do not think it means what you think it means.
– Inigo Montoya
This is a staple of pseudoscience. Not quoting The Princess Bride – everyone does that too much, regardless of their scientific credibility. I mean anomaly hunting. But the anomalies that woo-mongers think they’re looking for often aren’t anomalous in any useful, scientific sense of the word.
A scientific anomaly is a fact that is strange or unusual, in that it doesn’t fit into the model suggested by a particular theory. It’s some piece of data which genuinely oughtn’t to be there, if our present understanding is completely correct.
A scientific anomaly is emphatically not any event or occurrence that makes you go, “Oooh, that’s spooky“.
For instance. If biologists ever observed a modern chimpanzee giving birth to human offspring, that would be an anomaly totally irreconcilable with the current theory of evolution. This is true despite the persistently ignorant insistence of some creationists, who think that this is exactly what would be needed to finally prove Darwin right. Similarly, a verifiable discovery of those famous rabbits in the Precambrian would be entirely anomalous, and could not be accounted for within evolution.
If psychics exist, they would presumably be able to demonstrate their powers under controlled experimental conditions. If their rate of success at telling me what number I’m thinking of was sufficiently above what you’d expect from chance guesswork, then this would be an anomalous result, incompatible with the current scientific worldview which does not admit psychic powers. So, we would need to update our picture of the universe to accommodate this. This kind of anomaly can’t simply be left hanging.
One real anomaly, which intruded into astronomy in the mid-19th century, concerned the orbit of the planet Uranus. We had a wonderful theory of how everything in the solar system moved, and could predict where all the known planets would be at future times with fantastic accuracy, using Newton’s law of gravitation. But Uranus wasn’t quite behaving. People had checked and double-checked the numbers, but the seventh planet was definitely wandering very slightly off course, if the information they were plugging into the calculations was right.
So, this anomaly prompted people to start wondering what was going on that we weren’t seeing. For the most part, we had a pretty good theory going, and it turned out that it could be saved if we supposed that there was another planet further out, tugging on Uranus’ orbit a little with its gravitational pull. Then the numbers would all work beautifully again.
Crucially, though, they weren’t just assuming that some other massive body must exist out there, because the theory just had to be true. They were refining the theory, adding new elements to it, and in so doing they made a new prediction, by which they could test whether the new version of the theory was any good. Theories do that. If it can’t predict specific future observations, it ain’t a theory. And in this case, the Newtonian model of the solar system predicted a new planet of a specific mass, in a specific place, with a specific orbit.
They worked out where it should be, aimed their telescopes thataway, and, lo and behold: Neptune.
So, looking for anomalies and ways to account for them can be productive. But if you go chasing after things that aren’t truly anomalies in this sense, you’re not going to be doing anything as awesome as finding new planets. It just becomes pseudoscience.
The kinds of anomalies that some people go hunting for don’t hint at improvements to good scientific theories, but consist simply of any result which stands out in some way. Anything that looks a bit weird can be seen as an “anomaly” – even though weirdness is often a fundamental and entirely expected feature of the universe. Not every theory should be expected to immediately explain every observation. To suggest that a theory needs to be entirely thrown out, and replaced with some entirely new paradigm, is a common overreaction to one small “anomaly” being found.
So, when anomaly hunters approach an idea that’s actually pretty solid and widely accepted – say, that 9/11 was perpetrated by a band of Islamic extremists, or that ghosts don’t exist – they might pick up on some small factors that seem at first glance not to fit perfectly with the established explanation – say, that “fire can’t melt steel”, or that there’s something strange in your neighbourhood – and use these to call the established explanation into question. The very fact that anomalies exist – in this sense of strange-seeming things that can’t be immediately explained – is held up as evidence of the weakness of the prevailing theory.
But it may well easily be shown, with a little more work, that the prevailing theory is entirely consistent with everything we’ve seen – say, by slowly explaining how chemistry works, or by just growing up. These aren’t genuine anomalies, in that they don’t really need any new phenomena to be invoked to explain them. They fit just fine into a description of the world that we already have.
The kinds of anomalies that people latch onto might be things that we really don’t know the answer to, and can’t explain with certainty to everyone’s absolute satisfaction. But y’know, those are actually okay too. The unknown is pretty consistent with a lot of good ideas. Failing to absolutely nail every single detail of everything that’s going on is not scientifically anomalous at all. There’s no problem if it’s just an uncertainty; it’s only when something is truly inexplicable that your theory needs to be re-worked.
Every so often, a person might see some strange-looking lights in the sky which they can’t accurately identify. These reports are exactly the types of anomalies that UFO-enthusiasts go hunting for, but they’re not comparable to the problem with the orbit of Uranus. There’s nothing about a world free from alien visitors which implies that everyone will know exactly what they’re looking at every single time they spot a thing in the air. People occasionally squinting up at the sky and going “Wassat? I dunno… some geese maybe? Helicopter?” doesn’t undermine the skeptical position, because that could easily happen if there weren’t any aliens around. It would take much more than that – a genuine scientific anomaly, entirely lacking in plausible naturalistic explanations – before their case is supported.
This actually relates to Ockham’s razor, which I’ve apparently neglected to provide its own entry yet. These supposed “anomalies” are often held up as being evidence of some new and strange phenomenon, but if that phenomenon is something completely unproven, then a more mundane explanation might be far more reasonable to assume, even if we can’t be sure of all the details. There was no plausible mundane explanation – one that didn’t introduce some new assumption – as to why Uranus’ orbit shouldn’t fit the calculations; but people thinking they see stuff in the sky can easily be explained without bringing aliens into the equation. The Moon confuses some people. We know that boring stuff is often what causes these things. Saying that it might do so again, even without absolute proof, isn’t much of a stretch.
To see someone getting this particular point really wrong, check out Steve Novella‘s blog on this topic, in the section where he mentions Richard Hoagland. The “anomalies” that guy finds have only the flimsiest connection to his pet crazy ideas, and have very easy explanations already that don’t require massive leaps of logic to some totally new concept. When you have to invent vast alien civilisations and sinister, all-encompassing government cover-ups to account for the fact that there’s no other evidence for what you’re saying… at what point do you decide that maybe some mountains just happened to make a kinda interesting shape that one time? It’s a quirk, but not an anomaly.
Exploring the limits of a prevailing scientific theory’s power to explain the available evidence is one thing. But anomaly hunting, tracking down any slightly funny-looking result or interesting quirk of data, and using it to bolster the standing of your alternative hypothesis, however tenuous the connection might be, regardless of whether it matches with any of your own predictions, and without exhaustively checking whether it can be reconciled with the original theory, is not good science. It’s a wander into crazyville.
Read Full Post »