Hacker News new | ask | show | jobs
by Barrin92 2043 days ago
I consider Feyerabend's Against Method to be one of the most important books I've ever read. The article primarily addresses beauty as an important feature of science (and it probably is, as Tupolev said, "an ugly plane won't fly") but one can go much farther.

Not only is beauty valid in science, anything that advances science is. Epistemological anarchy is what we should strive for, because everything else inhibits science. In a sense the irony is directly visible in the modern scientific attitude. Empirical data and experimentation are supposed to reign supreme, the only thing that is verboten is experimentation with the scientific doctrine itself, which is supposed to be taken almost on faith, full of complex rituals that everyone needs to abide by.

The mainstream story about Galileo often implies that it was the sinister church at odds with the enlightened scientist. But the reality is much different. It wasn't just the church, but actually also establishment science at the time that was very unconvinced. Galileo's heliocentric worldview could not account for his own empirical observations. (Diffraction was not yet known, so he estimated that observed stars were closer by a factor that was off by a thousand, prompting everyone else to ask why, if his theories are true, observed stars don't change drastically in size if the earth indeed moves around the sun. He didn't really help his case by essentially writing what you could today call trollish treatises in which he was just mocking everyone for their criticism.

But as it turns out in the end he was right, despite what (even relatively sound) science had to say. It shows that science is fundamentally uneven, that ignoring existing 'truths' for the purpose of radical shifts, even for dumb reasons, might advance science.

2 comments

Agree: Against Method is a great work, and the argument for "anything goes" is compelling. IMHO the core argument can be used to dismiss software dev methodologies too. I recommend Feyerabend's Farewell to Reason as well.
I am reading this right now. A lot of awesome insight in the book. I just can't help but get the feeling that 90% of it is going over my head without having read Kuhn, Popper, etc.
That is indeed a problem. He's responding to a century's worth of epistemological thought, and that reply makes only limited sense without knowing what he's replying to.

I see it as a kind of conversation: Popper: Scientists invent hypotheses, then try to falsify them.

Kuhn: Turns out that's not really what scientists do most of the time, especially the most important ones. The inspiration comes out of nowhere.

Feyerabend: And that's a good thing, because they'd miss out on pretty much everything. You can never know for sure what's going to be a revolution and what's just crankery.

I find that Feyerabend is a little too broad in that last assertion. But it's a really important observation nonetheless, especially when a lot of scientists (and more importantly, science-adjacent types like most HN readers) are stuck back on Popper.