Science vs Reason | Hacker News Reader

Science vs Reason

adam.blog.heroku.com

28 pointssant0sk114 comments

I think the author's conflation of empiricism and science isn't quite right. Instead, I would say that science is the combination of empiricism and reason. While ultimately the process of science needs to agree with (and predict) data, the harder part is usually using reasoning to build a model of what's going on.

Yeah, the development of propositions is not really a point of distinction in science. But it was for that reason that I was thinking the author was only contrasting their epistemic value, even though his use of "discover" threatened to undermine that hypothesis.

Oh hell, definitions of science. There goes the planet.

Arguing about definitions can take all day, since all definitions are true by assumption. Let's try and find out what's happening here.

Feynman says science is the belief in the ignorance of experts (in his talk titled What is science?). This is a fun start, but it doesn't really bring us anywhere.

If you look at the world from a positivist perspective, like Mach did, there is an objective universe that we are refining a model of. In this sense, it is the loop

  while(model is not Truth):
    collect data
    adapt model to fit data

With this view, scientific models can claim to be superior to intuition in the sense of being an approximation of an ultimate truth.

A constructivist view, like what Hertz advocated, would be that we construct a model of the world that explains all available data. It's similar to the Machian view, but without the absolute truth. Here intuition can play a role, because any model that explains the current set of observations is right in a sense, and there is no partial order of rightness. Hertz stated that models of the world are to be ordered by correctness (how well they explain the data) and maximally correct models are to be ordered by elegance.

That leaves us pretty much where we started. The author is comparing data-driven decision making to experience and intuition-driven decision making. By Feynman's definition, one would place data ahead of experience. By Mach's, intuition is improved by availability of new data. The Hertzian view would be that data improves correctness and intuition improves elegance. When you are observing (or creating) something completely new, you have to throw away models and work from intuition and analogy. But that is just as data-driven, since the data tells you that it is outside the model. More data allows you to make better models. I don't think the distinction the author makes is all that valuable. Faced with a new situation, all you can do is work intuitively. That doesn't mean you should be ignoring your data, just the intepretations that your standard model has for it. Both elements are necessary for success, and all I got out of the article was that you shouldn't be ignoring data. I find that obvious.

I find the condition in your loop hard to test. A more realistic algorithm would be:

    while (model is not useful in the real world):
        collect data
        adapt model to fit data

The problem here is that "useful" is a quite low barrier. So for the second publication you also need to improve the usefulness of your model compared to the other models around.

That describes the work of scientists quite well.

Well, that was the joke, it was a play on while(True). It's an infinite loop, because you can never detect truth. And neither Hertz nor Mach said anything about usefulness. Also, I wonder why I can only think of physicists when thinking about definitions of science.

The work of a lot of scientists today can be described by

  while not tenured:
    publish
    bully others into citing you
    if you have time:
       do a bit of empirical work

I don't think this post is that interesting since reason is a large part of science, but I do want to share some history on "pure reason" and how it blinded humanity for 2000 years.

Pythagoras was a great mathematician, but also a big promoter of "pure reason" and mysticism. The religious movement he started was called Pythagoreanism (-) and one of the stories about Pythagoreanism is that they discovered irrational numbers, but they hid their discovery from the general public, since irrational numbers did not fit into their "rational world view". Plato and Aristotle are successors of Pythagoras. Much of the western philosophy and religion is built on their ideas. For them pure reason (and belief) was more important and valid than reality.

How much did "pure reason" blind us? Johannes Kepler was largely inspired by Plato and Aristotle and believed for most of his life that the solar system was built by the five Platonic solids (-). And most of his life was spent on building a model for this. Kepler abandoned his idea when he saw Tycho Brahe's excessive observations and data.

Personally, I think the world would have been a much better and more different place if experiments, data and reality were promoted instead of "pure reason" and mysticism. The Ancient Greece had their promoters of experiments (for example, Democritus (-) [the father of atomic theory], but he was hated by "pure thinkers" such as Plato).

* http://en.wikipedia.org/wiki/Pythagoreanism

* http://en.wikipedia.org/wiki/Platonic_solid

* http://en.wikipedia.org/wiki/Kepler

* http://en.wikipedia.org/wiki/Democritus

I can't decide which part of this is more wrong: the part about science being purely empirical with no theoretical component, or the part about triangles having to have sides of equal lengths.

> the part about triangles having to have sides of equal lengths.

Did you miss the equilateral qualifier?

EDIT: ah, it appears this qualifier was added to the blog post between your reading and mine.

Science is primarily about trying to explain the world. That's the theoretical, reasoning part. The explanations aren't derived from data; they begin as guesses.

What makes science special happens when two theories conflict and reason cannot choose between them. That's when experiment is used. (The theory which least explains the results is then rejected.)

Scientists use reason more than they use evidence, and most theories are rejected without bothering to test them.

What isn't widely acknowledged is that evidence itself isn't pure data but rather data + interpretation. Sometimes a better explanation can completely reverse the conclusion initially 'drawn' from a set of data.

Reason and science are not separate things.

Aristotle's argument was possibly valid, however it was not sound, as (at least) one of his premises were incorrect.

http://en.wikipedia.org/wiki/Soundness

I think the point the author is trying to make is that sometimes one must assume that their premises are correct, using one's intuition. However, one should test argument premises as soon as possible because, hey, our intuition could be wrong (just like Aristotle's was).

The section on "when to apply science" seems very disparaging of our predecessors: how does he know that no one tried experiments to prove either of the two ideas?

Both require knowledge of drag (and preferably a vacuum chamber) to perform correctly. Without that understanding, experimental results would probably have "proven" the rational model.

Does anyone know when air resitance/drag were discovered? A quick google indicated some time in the 1600s

I think the real point the author is arguing is closer to the difference between inductive v. deductive reasoning.

Uhm ... does he mean equilateral triangle everywhere he says triangle?

Stuff like this is why I have anti-procrastination on. What a waste of time. Stop voting up please.

Reply on news.ycombinator.com