"science works" -- Not entirely clear, to me at least. Surely science works in its limited domain. We have rockets to the moon and cable tv. But we still have war, and hate, and poverty, and human misery. So science doesn't "work" for all human problems; it only works for those problems that are amenable to science. Isn't something else needed? Or does your claim that "science works" mean that if we'd only increase the science budget, we'd eliminate all human problems? The twentieth century proved that applied science is really great for killing people. That the best science can do?

Literature and fine arts work too, the problem is that artists and scientists are talking to each other less and less. Philosophy used to mediate but for some reason doesn't anymore.

@user4894 - Scientists select problems amenable to solution via the scientific method. "All human problems" are not within reach--not now, possibly not ever. But infant mortality down to about zero? Lifespans routinely in the 80s? Talk as if by telepathy to practically anyone anywhere in the world on a whim? All possible only because of the cumulative knowledge of science. (Product development and such is often necessary as well.) Of course, given other human problems, we're also all too happy to use our vast power to make vast stick to smite the other guy with.

@Conifold - Literature and fine arts do not work in the same sense. You don't have reality coming back and telling you, "Oh, yes, that is such an evocative painting" the same way that reality tells you, "Oh, the spring equation! Yup, yup, you got it." So you don't have to listen to people nearly so much when you're doing science. You just have to be right (i.e. create reliable predictive models).

@RexKerr As someone who has been both a mathematician and a psychologist, I have to object. Scientists also choose other problems, more closely related to peace on earth, and those problems are just harder. Also, classical psychology has contained those closest to bridging this very divide, like Jung and Lacan. There is just very little space in our world for the middle ground. Our technological bias has even reached inside clinical psychology and split it into those who 'do physiology' (psychiatrists) and those who don't, and decided that people who do both are just too expensive.

@jobermark - There is markedly less progress in peace-on-earth research than e.g. elliptic curve cryptography or high-efficiency LEDs. I think the jury's out on the actual effectiveness of science in these areas, though I agree that in principle it ought to work given appropriate conditions. (For example, psychology is unreasonably full of fads driven by strong personalities compared to e.g. membrane biophysics.) I agree that lacking space for a middle ground is an ongoing problem (and may be getting worse).

The objection is that you dodged user4894's question unfairly. It is not that scientists just choose different problems, it is that it is better at some things than others. For instance, in psychology, it does not motivate folks to work in the middle ground, and that it does not deal well sociologically with all the ambiguity and bias that allows strong wills to dominate the field. He is right. It may or may not actually work in general as long as we look at it the way that we have traditionally.

@jobermark - Er, doesn't my first sentence imply that science is better at some things than others? (Along with implying that scientists work mostly on those?) One place where philosophers of science could make a positive impact is rejecting the idea that a lot of the stuff done under the name of psychology is "science". Very many psychology papers would, if subject to the same standards in other fields, go approximately like this: "We had several interesting ideas that we tested poorly and inndirectly, and the results were ambiguous and inconclusive. But we still like our ideas!"

We need to draw a distinction between "science" and "scientism". "Science" is a method for gaining knowledge based on experiment and observation. "Scientism" is the philosophical belief that all questions can be solved through science, or that anything that cannot be analyzed scientifically is not real. I would think that it is obvious that scientism is false. If you ask whether Julius Caesar really lived, for example, there is no possible scientific experiment we can do prove he did or did not, because the claim is that he lived in the past and not the present, and we cannot perform ...

... experiments in the past, but only in the present.

@RexKerr Hmm. That sounds like a domain intermediate between science and literary criticism. Too bad we have, broadly as a culture, rejected the value of all such domains. And no, you defended 'science works', which is sheer bias masquerading as observation. It works when it works, and when it doesn't, we pretend it is badly done, rather than difficult.

+1 Re, scientists losing interest in intellectuals, western culture as a whole has done this. Back when intellectuals held more sway in the public sphere, they were more relevant to everyone <- tautology ;). There are periods of Western history when it was normal for statemen to be philosophers (and even scientists as well, e.g. Ben Franklin). Today, the last thing they want to appear as is an intellectual or philosopher -- they will get laughed off stage by the populus.

..."Intellectual" is of course a word born of history, as is (somewhat later) anti-intellectual.

@Jay, that is one definition of science, and, I would claim, an excessively narrow one. Neither math nor about half of psychology as studied today are sciences by its standards, and I think that definition is part of the problem. Your Julius Caesar example even rules out geology as a science. Clearly there are not experiments in mountain formation currently going on... Basically there is not science and scientism, there is an obsessive, modern definition of science, and a classical one.

@RexKerr I finally wrapped my defensiveness around my two 'non-scientific' sciences up into an answer. I hope to avoid the hairsplitting that we ended up in last time by laying my position out on a philosophical basis.

"One problem is that science works."—Are you acquainted with Feyerabend's criticisms of the idea that there is a 'method' (or 'system of rationality') behind the practice of science? He also has some criticisms of your "science works". Furthermore, it isn't clear that science is working all that well when it comes to the topic of 'human nature'; see for example Donald E. Polkinghorne's Narrative Knowing and the Human Sciences. There's also Alasdair MacIntyre's "Crises, Narrative and Science".

@jobermark Yes, math is not "science". Scientists use math, of course, but that does not make math a science. Not everything that a geologist says is science. Speculation about things that happened in the past may be consistent with science, but if a theory is not testable experimentally, than it is not science by the traditional definition. You could, of course, insist on inventing a new definition that includes all sorts of untestable theorizing and philosophical speculation. And indeed many people today do that, for example when they want to call their political or social beliefs ...

... "science". But such a definition of science makes the word virtually meaningless. At worst, it's the classic logical fallacy of guilt/honor by association: I define "science" as the sort of knowledge that lets us build moon rockets and cell phones, then casually lump in "... and distinguish superior races", and now the fact that science has proven to be so effective means you must believe me when I declare that such-and-such race is inferior because "it's science". I'm not saying that that's what you're trying to do, just to say that's the danger of a fuzzy definition.

@labreuer Science DOES work. It has proven to be an incredibly powerful tool ... but not every tool is useful for every job. A hammer is a very useful tool: you can use it to put in nails, break things open, etc. But I wouldn't try to use one to put in a screw. Certainly not to remove a splinter from my finger or measure out a dose of medicine. If we're asking, say, "Should marijuana be legalized?", science can be helpful in telling us the medical effects of marijuana. But it is useless at answering the fundamental moral and political questions.

@Jay, if science == experimental progression against falsifiable goals then 1) String Theory is not a science, it only explains everything we already knew before it came into being and 2) Alchemy is science, there was a lot of experimentation, the theory was just bad. People improved it, it grew and had economic and material success. Read someone like Lakatos or Feyerabend relate a real instance of theoretical paradigm shift, and you will not recognize this notion of science any more.

Your definition is also far from 'traditional' given that it took Popper, in the middle of the last century to even state it. And he meant for that to be descriptive, not prescriptive definition. It is part of the hegemonic obsession you are calling scientism to even accept this definition as you are using it -- to say what is and what is not science.

@labreuer - Feyerabend is correct in calling out the excesses of what works and what doesn't, but he nearly throws the baby out with the bathwater, often failing to notice why it only worked accidentally: people somehow had the right ideas, but they were not justified, and used tricks to get their ideas adopted before justification came along. Human nature is really complicated. It's really hard to do proper science on really, really complicated things without awesome tools. So it's not doing all that great now (though we are learning a few interesting things), but we ought not expect it to.

@Jay - All that is required to do science even in a Popperian sense (which is quite strict) is that you can make predictions, do stuff, and see if those predictions come true. You can learn an enormous amount about geology, history, astronomy, and so on, under those "look but don't touch" constraints. There's no epistemological requirement to poke something and see a reaction; that's just a convenient way to find what you're looking for. If there is adequate data, but only if, you can do Popperian science on past events, including geometry. (Math is not science, I agree.)

@RexKerr We seem to be having a massive quantification problem here. Feyerabend does not contest that the particular method works, but that it is a reasonable model of what has worked historically and that it is the thing most likely to work in all domains. He does not throw anything out by demanding we critically entertain continually modifying these rules, instead of stating them prescriptively.

Popper says 'this is what has always worked, and statistically that makes it really likely it will always work'. What is more true is that 'this is what scientists tend to have success with during this century' and I would claim that is only true if the notion of 'success' is incredibly biased toward increasing laziness, distraction and destructive over-consumption.

('This century' for Popper would obviously have to mean the century before he wrote, which is the ascent of industrialization in Europe.)

@jobermark - The whole puzzle with science is why it works in the fashion that it does when nearly everything else does not. Feyerabend is correct in dampening the enthusiasm for narrowly prescriptive methodology, but in my opinion he mistakes success for correctness.

Much of the trick in doing science well is in being able to detect success, and that is where appropriate methodology comes in. And almost every time, it looks remarkably similar to what Popper outlined originally: you find something that will help greatly increase your confidence of which of two hypotheses is wrong.

It's not the only way to do it; I agree with Feyerabend there. But how far can you stray and have reliable success? I'm not interested in case studies; people get lucky now and then no matter how ill-advised their approach is. I'd like to know, systematically, what reason have we to believe that anything else works well? And that to the extent it works it is not because it is essentially falsificationist?

(Where I would extend falsification to include anything that would cause a Bayesian to substantially change his ratio of probabilities of two hypotheses.)

OK, but for most of that history, no one thought about falsification, and then suddenly, it becomes an absolute requirement of everything that works. This is not a definition we are 'straying from' it is a definition that modernists are attempting to impose. Popper himself does not believe it addresses, for instance, the success of Darwin. But we are suddenly mad for this theory and start treating it like a definition.

Moving back from falsificationism to some theory of induction, or leaving it out of the notion of paradigm comparison would cause no harm. Because it is not a real component that scientists themselves use to judge their own work -- or no one would do String Theory, which much of physics currently loves. This is a case of trying to make life imitate art.

We like it because previous theories of induction are all circular, but it turns out at the point of application, this is equally circular, because no one has ever rejected a new paradigm based on falsification. It is an historical illusion. When you look at the actual supposed events in history, they just did not go that way.

If historical analysis is not part of your way of deciding what works well, you are not deciding, you are pretending.