Tuesday, June 12, 2012

Karl Popper's "Introduction to the Logic of Science": A Summary

Look, there's been a lot of controversy lately about what is and isn't "good philosophy" and what we can and can't say, and opinions on that are like assholes - everyone's got one, and no one knows what to do about Wittgenstein's. But I think we can all agree that a really worthwhile task is understanding the world around us, especially with science. But no one is really talking about how science is supposed to work - they're just going around saying "Oh, science! Look at me, I'm science and I'm the best!" But if everybody is going to be using science, we damn well better make sure we understand exactly what the fuck is going on in science, right? I thought so.

After all, what makes science so great? Everyone seems to think it's induction, that induction is what makes science so great. They're all, "Look at all this data we have that proves that gravity is real! Where you at on the data for beauty and the soul, poets? Yeah, that's wht we thought. Fuckin' poets." Except fuck that, induction is broken. We've known since David Hume that induction is broken, and I'm here to tell you that there's no fixing it. You can't just drop shit a bunch of times and then be like, "Well, all these rocks keep falling at the same speed, and we've been dropping them for like three hours now so obviously that's how it is all the time everywhere in the universe." But for some reason people keep trying to figure out how many times we need to see rocks fall at the right speed before we can assume it's the same everywhere - "Uh, maybe if it happens the same way the first thousand times we try it, we can just go with it." Oh really? A thousand? How'd you get that number? Did we just test a thousand different things a thousand times and they've all worked so far, so we assume a thousand is the magic number? Bullshit, something could go terribly horribly wrong with the sunrise tomorrow and then your arbitrary "thousand" number is right the fuck out the window. You can't inductively justify induction, so you can't pretend like induction is all there is. Knock it off.

"But Karl, without induction, how will science tell us things? Why is is any different from metaphysical bullshittery?" Holy fuck, I'm getting there! Relax! But first, let's talk about how science even starts for a second. Probably because they've got their heads up their asses with induction, these same bros think that maybe we need to know how the scientist thinks, so we can get from one idea to the next and verify it, as if we need some logical breakdown of producing theories so that we can go about proving them in an orderly manner. But what the fuck would that even look like? Sure, maybe psychology can help us have better ideas or whatever, but how would understanding how we have ideas tell us anything about whether those ideas are good or not? Oh wait, right, it wouldn't. Whether a theory matches the world or not has nothing to do with how someone came up with it - we can draw theories out of a hat for all I fucking care.

So how does science proceed, if induction is fucked (which it is) and we can't logically determine how to have new ideas (which we can't)? Easy - just take a fucking guess. No, I don't mea- dammit, you asshole, I don't mean "guess how science works", I mean guessing just is how science works. Just start guessing shit and go from there. Of course you're going to make a couple stupid guesses at first. Seriously, some of the shit you're going to try is going to be genuinely fucked in the head. Remember when we thought heavier objects would fall faster? Boy was that wrong. But we took a guess, tried it out, and it didn't work. Instead of being whiny babies about it, scientists just took another guess and then tested that out, too. That's the process: guess, and then you test that guess. And if the test works, you're like "Huh! That was an even better guess than I thought." And the more tests it survives, the more people are like, "Great guess! I'll bet that's probably it." And then you get to a test that your guess doesn't pass, and you're like, "Welp, close but no cigar. Back to the drawing board."

We'll eliminate the fucking stupid guesses pretty quickly - it doesn't take long to show that we can't move things with our minds. Eventually, you start to build a pretty cool system of things so you can make better and better guesses. and you can totally use data to make good guesses; you don't always have to invent something completely new every time. I'm just saying that's all the data does, helps make good guesses. It doesn't prove shit.

And look! That method is deductive! What incredibly good news! You don't have to derive a universal statement from a bunch of single events, which is great because you can't; instead, you just guess a universal statement and then see if you can't find an event that breaks it. You can't get from "the sun keeps rising" to "the sun will always rise" but if one day the sun doesn't come up, you can be damn sure about "the sun does not always rise." All you need is one bad apple and you know for sure that not every apple is good, no induction needed. QED, motherfuckers.

And - AND - now we know what is and isn't science! Holy fuck I am on fire here. Not actually. Just- look, I'm solving lots of things, is my point. Scientific theories are falsifiable - they're incompatible with certain things we could observe. They predict shit, and then we see if that shit really happens. Back when we thought Newton's Laws were totally, completely true, Mercury had this weird fuck wobble in its orbit that said we should find another planet. Except we looked and no planet. And now we know for sure that Newton wasn't completely right. Einstein? He was a patent clerk for fuckssake, and he came up with a fucking incredible guess. And we just keep devising more and more complicated tests to check it out, and it keeps on passing. When it does finally fail, we'll fucking know. There won't be aaaaany confusion whatsoever. Souls? How the fuck would we go about testing for souls? "Well, we cut him open, and we didn't find a soul, so..." "Yeah, but you can't see souls! That's the whole point!" So you're saying we can't ever test for souls? That's fine, just, it means souls can't come to the science party. They're not falsifiable. You must be THIS FALSIFIABLE to ride the science ride, and souls just aren't.

I'm not saying crazy shit doesn't happen from time to time - sometimes someone's tools are broken, sometimes some clumsy fucking lab monkey burns everything down, sometimes your fiber optics report that neutrinos are fast as shit, too fast, really. So we can just lay down some rules for doing science, rules like "make tests that we can repeat to double-check" - don't be all, "Well, there was no fire last Tuesday, so, that's it for the theory!" We need to make sure that we justify our guesses with tests that in principle anyone can understand and perform. And we can totally come up with those rules later - I'm just saying we can't pretend like science is straight logic and everyone else is wrong. When that happens, bros start to get real uppity with people who disagree with their theories, like "don't you know how logic works, you simple simpleton?" But really, we should never act like we know exactly what the fuck is going on in science. All we've got are some super-good guesses, and as soon as we lose sight of that metaphysical bullshit starts creeping into our theories. BAD. BAD UPPITY SCIENTISTS. Yeah, I wish we could be more certain, too, but it's a damn sight better than that inductive bullshit you were using. Seriously, you guys, what the fuck was up with that?

---
"Introduction to the Logic of Science" is Part I of The Logic of Scientific Discovery, which you can get at Amazon. Part II is just a much longer and more careful laying out of his theory of falsifiability - he handles specific objections and technical questions.

7 comments:

  1. Bayesianism: Kicking the shit out of induction, falsificationism, and everything since 1763.

    http://plato.stanford.edu/entries/epistemology-bayesian/
    http://shawnslayton.com/open/Probability%2520book/book.pdf (Is that a free pdf of Jaynes' masterwork Probability Theory: The Logic of Science? Yes. Yes, it is.)

    ReplyDelete
  2. Brah, you're such a fucking boss, your work has never failed to blow my mind.

    (using induction:: you will always blow my mind. wait a sec)

    ReplyDelete
  3. This is great! I remember reading Popper as an undergrad. Now that I've left academia and don't get my 'philosophy' fix I really enjoy reading your posts!!

    ReplyDelete
  4. Well done. Thanks again for helping people sift through the arrogant writing of most contemporary philosophers to deliver the message.

    ReplyDelete
  5. I read Popper as an undergrad too. I actually majored in Science, and at the time, it was all good and made helluva sense in a clean-cut, neat Science-y sorta way. So I appreciated this. Thanks, Bro.

    Also: my favourite line from this piece was "QED, motherfuckers". Haha.

    ReplyDelete
  6. Woah bro, I just learned more from this post than I did in an entire semester of Philosophy of Science. Thanks!

    ReplyDelete
  7. "Look at all this data we have that proves that gravity is real! Where you at on the data for beauty and the soul, poets? Yeah, that's "wht" we thought. Fuckin' poets."

    Might wanna change wht to what.

    ReplyDelete