Well, okay, also what if it’s not? But also, what if it’s more complicated than that, and we end up with a continuum between scam and non-scam? What if, because of the way humans work, it’s nearly impossible to tell? What do we do about that? Do we even have a meaningful choice here? Maybe. Maybe not.
I grew up wanting to be a physicist.
I’d read a lot of science fiction, see, and I figured it’d be cool to learn a ton of brain-busting math and figure out a loophole in natural laws that enabled faster-than-light travel. Because why wouldn’t I want to do that?
Then I went to college and met scientists. It turned out a lot of their day was spent in political maneuvering–popularity contests, in other words. I wasn’t too discouraged. After all, I was young and smart and strong (wasn’t I? or if not, why not? could I fix that?), so I could get past that somehow.
It got a little bit worse when I read a book by Thomas Kuhn, who popularized the term “paradigm shift,” which in that book referred to how scientific knowledge tends to progress here in the real world. Unfortunately, it amounted to a couple of things that I found discouraging: (1) With competing paradigms, neither can be understood in the context of the other. Which is likely part of the reason people seem so often to talk past each other rather than communicate. Unfortunately this means that, given two competing paradigms, assuming sincerely-held beliefs…few people will convert from one to the other based on logical arguments alone even though all participants may be being entirely reasonable. Therefore, (2) sadly, scientific advances have generally become accepted through attrition: the old guard dies off and the new kids grow up. Were the new kids even right? Maybe! Maybe not! Nothing about the process is inherently more meaningful than any other popularity contest. Oh, and I guess there’s (3): People who disagree with us are not necessarily as stupid as they might appear. Hmm.
What about the real world, though? Don’t better theories produce better predictions, and doesn’t technology keep getting better as a result? Well, yeah. Mostly, with caveats. But when we look at competing theories and try to design something that actually works, we’re not really doing traditional science. We’re doing engineering. I could expand on that, ’cause I think most of the advances people credit to advances in science are actually advances in engineering. I might even claim that engineers typically lead the way, with scientists most usefully employed in trying to figure out why the engineers’ discoveries actually work. But that’s another post entirely.
Anyway. My ambition withstood all that. I could do it! I could become a scientist and change the world! I just needed better math-fu!
But math has no content
Then I started thinking about implications of something I’d originally read in a book by Heinlein. This was the notion that math has no content. In other words (in case you don’t want to read that Wikipedia article), it’s just something we made up. Which is clearly true…the painful implication, though, is that math exists in a sort of logical universe of its own. We set up the rules of this universe however we wish, and then we (well, a few of us) try to figure out clever but non-obvious “truths” about that universe.
Why is this “painful”? Because we always have to fudge things when we apply these logical mathematical rules to the real universe. Generally this process…has strictly limited utility. We can model some very simple systems, and predict how they’ll behave. But if we want these systems to work in the real world, or IOW to behave as our theories say they should, we tend to have to find or build special shelters that minimize “outside” influences. Or start talking about “entropy” as a stand-in for “all that other stuff that’s going on.”
Let’s start with an obnoxiously simple scenario. Take two numbers. Heck, call ’em “1” and “2” and let’s assume they work exactly the way we all expect them to. So, two is greater than one, right? Er…what if I want to sit down in a chair, though? And I have two chairs available? Two might be greater than one, if they’re perhaps in different places and I might want to sit in either. Or two might be less than one, if they’re crowded too closely. And one of the chairs might be greater than the other, because it’s lighter or more comfortable or greener or…
Ridiculous, I know. I just took “greater than” and pretended it meant “better than” and then made it dependent on context. And pretended “one” also meant “one chair.” But this is exactly what happens all over the place when you think numbers are somehow objectively real and try to apply them to things that (probably) are real. You end up having to make excuses for them. Ick.
So I grudgingly decided it was probably impossible to prove anything at all about the real world using math. All I could do was model it, and see to what extent reality matched my expectations. I was wavering in my convictions, but decided to press on. Somebody would figure out cool stuff. Probably lots of people would! Why not me?
The hits kept coming, though.
Math isn’t as cool as we think it is
I’ll get to more extreme examples, but…let’s look at something apparently very simple. Consider the Moon, the Earth, and the Sun. Given their initial mass, known initial positions, and known initial velocities–and assuming there is nothing else but gravitation affecting any of the three–where will they all be in the future? Guess what? There’s no general mathematical solution to this problem. Sound far-fetched? It’s known as the three-body problem, though, and it’s a real limitation. Gets worse when there are more factors involved, too. Yes, I know you can go look up stuff about planetary orbits. Guess what? The predictions you find are based on approximations, not derived from first principles. Which means that, even with a very simple situation like this, theory can only take you so far. Then you have to start fudging.
Then there’s chaos theory…which is essentially the notion that you can take a very simple system, composed entirely of completely understood bits & pieces, and…fail to make useful predictions. Even though they’re in principle “deterministic” (thus predictable) it turns out that at some point the systems begin behaving in a way that is simply not predictable–even in theory. It’s too dependent on initial conditions, meaning…pretty much anything at all can turn out to make a difference. This is also known popularly as “the butterfly effect,” and it’s kind of a big deal when you want to make predictions, based on math, about the real world.
Then there’s the scientific method
Okay, this one seems pretty straightforward. No brain-stretching involved. In a nutshell, science is about making predictions and testing them against the real world.
Which means, of course, that if you’re either (a) not making predictions, or (b) not validating those predictions vs. real-world events, you’re…not doing science. At all. Which doesn’t mean you’re a bad person. But it does mean that “scientist” is a strange label for people to use when talking about you.
Oh, and about modeling
Another career choice? For some, perhaps. Probably not for me, though. Curse my luck!
Moving on. At first I’d figured “natural laws” were what science was about, and we knew some of ’em, and we just needed to find more. Then I realized that what we generally thought of as “laws” were in fact approximations. Always and forever, amen.
So, okay. I could deal with that. Science is about building models and testing them. Building the model has to do with coming up with a notion that is, or may be, compatible with known past events–and testing it against unknown future events.
The process of building a model compatible with past events is sometimes known as curve-fitting (and in some contexts this term is pejorative). Unfortunately, even perfect curve-fitting offers no guarantee whatsoever that the model has predictive value. That’s what testing, aka “experimentation,” is all about.
Science is what? Testing.
Wow. To do what I wanted to do as a theoretical physicist, I’d have to learn all sorts of crazy math skillz, come up with a notion that might or might not be valuable if tested, get people to listen to me, divert other people’s resources to testing my notion(s)…and then, even if I were ultimately successful? And I actually convinced people of that too? Then I wouldn’t have discovered a law of nature, or proven anything. I would simply have come up with a way to make predictions in some area where previous predictions either hadn’t been attempted or had been unsuccessful. Or perhaps they’d been a shade less successful.
I’m not saying that isn’t noble work. But I am saying it had less appeal for me in my late teens than the idealized (somewhat obscure pun fully intended) notions of science I had had when I was younger.
But…what’s science not?
Lots of things!
Astronomy, for instance, isn’t a science at all by any definition I’d consider reasonable. It’s a field of study, sure. Interesting stuff! But where are the experiments? How do competing explanations, both compatible with observed data, get tested? Answer: they don’t, really. Except by luck, if we get to see some new phenomenon that helps to resolve the question. So why is one notion more popular than another? I dunno. It’s kind of like wondering why one person won an election and another one didn’t. There are lots of competing theories out there, but guess how many tend to be taught in schools? Hmm.
Speaking of that stuff, what about psychology/psychiatry? To the extent it’s about making predictions, it’s a science. When it’s about curve-fitting, though, it’s just not. When it’s about untestable theories, it’s really not. When applied to individuals, as it so frequently is, it tends to be essentially noise. Another book got me all mad about this one…Thomas Szasz was a smart feller.
Where’s the scam, though? You promised a scam!
Also, lots of things.
Let’s look at weather forecasts. There’s clearly stuff to study there! And wouldn’t it be neat to create a tornado on demand? Okay, that link is slightly lame and ends with a silly quote, but still! Cool!
Thing is? Weather’s complicated. Very. It’s one of the classical examples people point to when they start talking about chaos theory and complexity. So…what does that mean? It means that accurate predictions are probably impossible in principle (if we eliminate luck as a factor, anyway).
This is very far from saying there’s nothing to study, or nothing to learn. But it does mean we should be very, very careful about believing anyone who claims to know what will happen next week. Or tomorrow, even.
Generally weather models are engineered to fit past data…at least an attempt is (often, and I’d like to think always) made to do so. This process necessarily involves simplifying both the data input and output. Given the chaotic/unpredictable nature of the underlying system, how useful is the output likely to be? What sort of predictive value will it have? In, you know, theory?
So “climate scientists” take this a step further and talk not only about accurate forecasts they can’t produce, but about how those forecasts change in relation to some specific change in the input. To which I have to say: wow. That’s really nifty. You can predict an average temperature 100 years out even when you’re clueless about next month.
It gets worse when testable stuff like “global warming” goes away in favor of undefined “climate change,” ’cause another thing about useful science is that a hypothesis is defective unless it’s falsifiable. Meaning: if it can’t be disproven, it’s not about science. Because…how can it be tested? So “climate change” needs specific numbers, or it’s just noise.
I have a theory. It’s this: anybody who actually understands what a complex system is, and who is involved in publicly predicting climate changes today, is in it for some reason unrelated to scientific discovery. Because honest people who “get it” will choose some other line of work. Which, if I am correct, means that the entire field is very likely chock-full of people who either (1) don’t understand the basic problems they’re dealing with, or (2) are willing to lie about them.
Unfortunately, my theory isn’t easily testable. Damn. So in all honesty I have to content myself with saying that it seems plausible but I can’t prove it. Therefore you should feel entirely free to discount it. Oh well.
Weirdly, this will be seen by some as a political position on my part. Here’s what I have to propose: can we quit thinking computer models have validity because of the credentials of their creators, and instead measure the accuracy of their predictions? ‘Cause that’s actual science, there.
Meanwhile I have a friend who got his Ph.D. via a dissertation that, among other logical issues, used math that assumed the human body was a perfect sphere. This apparently was no problem for him at any point. So does he enjoy my company because my body has in the past closely matched his theory? Possibly. It’s an interesting notion…maybe I should follow him around and see who else he hangs out with. For science.
Meanwhile? Have fun out there!
Be First to Comment