There is the question of content and that of procedure. The question of content is mainly a matter of definition and custom. When a native English speaker says "science" they almost always mean "natural science." On occasion they include the social sciences too. Even rarer so mathematics. The German word for science is "Wissenschaft" and in its usage is much closer to the Latin root "scientia."
According to the Online Ethymology Dictionary
- Science from Latin scientia "knowledge," from sciens (gen. scientis), present participle of scire "to know," probably originally "to separate one thing from another, to distinguish"
The German "Wissenschaften" include besides the natural sciences not only the social sciences and mathematics, but also "Kunstwissenschaft," "Musikwissenschaft," "Literaturwissenschaft," etc, literally the science of art, the science of music, the science of literature. It speaks for itself that if you Google "Kunstwissenschaft" the first two suggestions are the completions "in English" and "translation." In the following I want to leave the content of "science" as open as the German and Latin expressions leave it, and let it be constrained by procedure, which for me is the more interesting aspect.
As for the procedure, I have come across these three points of view:
- A: Science is what proceeds by the scientific method
When pushed, the usually well-educated defender of this opinion will without hesitation produce a definition for scientific method along the lines hypothesis, experimental test, falsification or gradual acceptance as established fact.
The problem, as Feyerabend pointed out, is that a lot of progress in science did simply not come about this way. Worse, requiring a universal method may in the long run stifle progress for the reason that the scientific method itself can't adapt to changing circumstances. (I'm not sure if Feyerabend said that, but I just did.) Requiring people in a field in which creativity is of vital importance to obey certain rules, however sane they seem, begs for somebody to break the rules - and succeed nevertheless.
There are many examples of studies that have been pursued for the sake of scientia without the possibility or even intention of experimental test, and they have later become tremendously useful. A lot of mathematics falls into this category and, until not so far ago, a big part of cosmology. Do you know what will be possible in 100 years? Prediction is very difficult, especially about the future, as Niels Bohr said.
The demand of falsifiability inevitably brings with it the question for patience. How long should we wait for an hypothesis to be tested before we have to discard it as unscientific? And who says so? If you open Pandora's box, out falls string theory and the technological singularity.
Finally, let me mention that if you sign up to this definition of science, then classifications, that make up big parts of biology and zoology, are not science. Science however are literature studies, for you can well formulate a hypothesis about, say, Goethe's use of the pluralis majestatis and then go and falsify it.
- B: Science is what scientists do
This definition begs the question who is a scientist. The answer is that science is a collective enterprise of a community that defines its own membership. Scientists form, if you want to use a fashionable word, a self-organizing system. They define their own rules, and the viability of these rules depends on the rules' success. The rules cannot only change over time, allowing for improvement, there can also exist different ones next to each other that compete in the course of history.
I personally prefer this explanation of science. I like the way it fits into the evolution of the natural world, and I like how it fits with history. I also like that it's output oriented instead of process oriented: it doesn't matter how you do it as long as it works.
In this reading, the scientific method, as summarized in A, is so powerful for the same reason that animals have the most amazing camouflage: Selection and adaption. It does not necessitate infallibility. Maybe the criteria of membership we use today are too strict. Maybe in the future they will be different. Maybe there will be several ones.
The shortcoming of this definition is that there is no clear-cast criterion by which you can tell what of today's efforts are scientific, in much the same way that you can't tell whether some species is well adapted to a changing environment till they go extinct, possibly because they fall prey to a "fitter" species. That means that this definition of science will inevitably be unpopular in circumstances that require short and simple answers, circumstances in which the audience isn't expected to think for themselves.
Given the time to think, note that the lack of simple criteria doesn't mean one can't say anything. You can clearly say the scientific method, as defined in A, has proven to be enormously successful and, unless you are very certain you have a better idea, discarding it is the intellectual equivalent of an insect dropping its camouflage and hoping birds don't notice. Your act of rebellion might be very short.
That having been said, in practice there is little difference between A and B. The difference is that B leaves the future open for improvement.
- C: Science is the creation, collection, and organization of knowledge
"All science is either physics or stamp collecting," said Ernest Rutherford. This begs the question whether stamp collection is a science. The definition C is the extreme opposite to A; it does not demand any particular method or procedure, just that it results in knowledge. What that knowledge is about or good for, if anything, is left up to the scientist operating under this definition.
The appeal of this explanation is that scientists are left to do, and collect what they like, with the hope that future generations find something useful in it; it's the "You never know" of the man who never throws anything away, and has carefully sorted and stored his stamps (and empty boxes, and old calendars, and broken pens, and...).
The problem with this definition is that it just doesn't overlap with most people's understanding of science, not even with the German "Wissenschaft." There is arguably a lot of knowledge that doesn't have any particular use for most people. I know for example that the VW parked in front of the house is our upstairs neighbor's, but who cares. Where exactly does knowledge stop being scientific? Is knowledge scientific if it's not about the real world? These are the question you'll have to answer to make sense of C.
(img sources: click on image)