3 ms·

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. Wh

• zach 17y ago
  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.
• Kliment 17y ago
  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.
  • beza1e1 17y ago
    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.
    • Kliment 17y ago
      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