tom thinks

date 2001-03-10:14:36
Creatures There's something fundamentally endearing about a dogs rolling in snow. Ebb the Dog not only roles in the snow, but snuffles along with her nose buried in it, sending up a plume of powder on both sides. I have no idea what she thinks she's doing, but it sure looks like fun.
Anomalies Not everything makes sense.

There are a lot of reasons for this. First, we have incomplete information. Second, we make mistakes, both of judgment and in action. Third, we tend to think that simply because something has happened it must be fairly likely to have happened.

Of these reasons, the latter two are by far the most difficult to deal with because they're nearly invisible. We know we have incomplete information--very few of us really think we "know it all."

But we often forget we--and others--make mistakes. Artistic portrayals of the history of science show how prevalent this misunderstanding is. There's a terrible movie about the history of the atomic bomb that contains a lovely example. A group of physicists is in a seminar and someone rushes in and says something to the effect of, "Fermi has discovered that hydrogen moderates neutrons." The physicists all look at each other "with wild surmise" and then rush out of the room, all having immediately realized the significance of this discovery.

In the real world, it doesn't work like that, because one of the things scientists get trained to do is always consider the possibility they've made a mistake, and to assume everyone else has made them too. This is one of the big differences between scientists and people who pursue "alternative" approaches to the world; the alternative community tends not to consider the possibility of mistake very seriously. This has advantages: they sometimes see things that overly-skeptical scientists dismiss. But it has big disadvantages too.

Consider the farcical scene described above. In the real world, the first response would have been the immediate recollection of the last mistake Fermi made, and everyone would have a laugh. Then there'd be some discussion as to what on Earth the claim could possibly mean, and the person who brought the message would be asked impossibly detailed and marginally relevant questions about the apparatus and technique that were used. Experimentalists would offer possible alternative explanations for any suggested apparatus, along the lines of, "Well, you know with a counter of that design you've got to be really careful with electrical noise, and they just put a new elevator in that building, and I bet he's measuring the surge from the electric motors." An older member of the faculty would recollect that Rutherford had tried something similar years ago and seen nothing. There'd be a side-debate about what Rutherford actually did, or didn't do, and where he was when he did it. A theorist would break in, having finally understood what experimental claim was being made, with a proof that it couldn't be done. The gathering would break up into pockets of minor bickering and argument, as the import of the claim slowly settled into each individual's mind.

It would be days before anyone really understood what Fermi had done, much less how he had done it and what significance it had.

The reason why real scientists wouldn't go rushing out of the room to follow up on such a momentous announcement is that they know perfectly well that nine times out of ten such announcements are wrong. I say this as someone who was briefly swept up on the cold fusion flap, and knew people who built apparatus to test the early claims within days of the original announcement. One of my colleagues devised a variant apparatus that would have produced protons rather than neutrons had cold fusion been occurring, and let it run overnight. When he came in in the morning there was a HUGE peak in the proton spectrum from the surface-barrier detector he'd set up beside the apparatus, and for about two seconds he thought it was all true. Then he looked for the mistake, and found that electrolysis had eaten through one of the electrodes and leaked heavy water onto the surface of the detector, causing it to short and produce the observed peak.

We've all had experiences like this--the investigation of nature is hard, and we learn to discount data that isn't rock solid. It's too easy to build a castle of theory on the sand of error, and embarrassing when it falls down. Sometimes there's no escaping it--we don't detect every mistake made, and nature may conspire against us as it did in the 17 keV neutrino mess, where multiple independent experiments all appeared to suggest the existence of a particle that in fact does not exist.

This is where the third cause comes in: we tend to think that just because something has happened, it's likely. But billions of events of different kinds occur every day. The odds of some pretty improbable things occurring are therefore remarkably high. If I flip a coin twenty times the odds of twenty heads coming up in a row are a million to one. If I flip it a couple of million times, twenty heads in a row will come up somewhere in the sequence nine times out of ten.

So when people point to some remarkable phenomenon, like the face on Mars, and say, "What are the odds of such a thing happening?" we have to ask, "The odds in how big a sample?" Consider all the surface area of all the terrestrial planets in the solar system. That's a big area. What are the odds that some random formation in all of that will look a bit like a human face to humans, whose visual apparatus is so good at seeing faces that they are amongst the first objects babies are able to discriminate from the background?

Despite all this, anomalies are not just important, but fascinating. Careful investigation of anomalies sometimes leads to substantial discoveries. The precession of the orbit of Mercury failed to agree with the calculations of Newtonian gravity by some very small amount, but was taken seriously enough that explaining it became one of the first triumphs of general relativity. Scientists sometimes are too dismissive of anomalies, because they know how much work it takes to investigate them properly. Non-scientists sometimes are too enthused by them, and fail to realize how much work it is to understand them, and fail to accept that because we have incomplete information there will in some cases be a core of anomalous behavior that we just aren't going to understand right now.

The appropriate response to this is not to ignore such anomalies, but to put them on a shelf somewhere and wait for more information to accumulate. In the fullness of time, perhaps generations hence, it will, and by accumulating anomalies this way our understanding of the world will slowly but surely increase.

I have some favorite anomalies of my own, and will talk about them in this section over the next little while.

Find Enlightenment