The question "what is science?" always becomes more pressing when debates about evolution and creationism are going on. Even though the question is actually a bit of a mess, it suddenly becomes tempting to try to offer a short, concise description of science that can be used to guide decisions about what should and should not go onto high school curricula. Often, the first thing people draw on is Karl Popper's account of science, based on the idea of falsifiability. For Popper, a hypothesis is scientific only if it has the potential to be refuted by some possible observation. There are serious problems with this formula, and hardly any philosophers would accept anything as simple as this. It is a fantasy to think that big theoretical ideas in science are set up in such a way that they can be knocked out, with logical certainty, if some single crucial observation is found. For example, all scientific ideas, especially the big theoretical ones, only make predictions about observations when assumptions are made about many other matters (for example, the experimental apparatus and the circumstances of observation). But no one has come up with a reasonably simple alternative formula to Popper's one that does much better. So is it hopeless to try to say something simple and general about how science differs from other kinds of inquiry? What should philosophers say when judges in court cases (like the recent one in Pennsylvania) are looking for a way of deciding whether a controversial idea counts as genuine science?
To me, Popper was onto the right general idea, but he simplified the story too much. He also tried to express his test for science in terms of a test applied to the content of scientific hypotheses themselves. I think it is better to start from the idea that there is a distinctively scientific way of handling ideas and hypotheses.
Most ideas, especially big ones like evolution and divine creation, can, in principle, be handled both scientifically and unscientifically. The scientific way of handling a theoretical idea is to look for ways to expose it to observation. This does not mean that the idea has to be formulated so that a single observation could knock it out. Often, what it means is that scientists start with a very simple version of the idea, and look for ways to modify, develop, and extend it in response to what is observed. Crucially, when one version of the idea turns out to be inconsistent with what we seem to be seeing, and a modification is needed, the next move is to a version of the idea that can be "exposed" to observation in the same sort of way. With an idea as big as biological evolution, this tends to lead to the development of a whole range of specialized research programs, each looking at the role of evolution in some specific context. Some people look at evolutionary processes in model organisms (like bacteria and fruit flies) in the lab, others look at the patterns in the fossil record, and so on.
At each stage in the process, there will be unanswered questions and puzzles. At each stage there are "gaps," as the anti-evolutionists like to say. Of course there are gaps! The whole point of the process is to push the idea into new areas, and get it to make contact with more and more phenomena of the kind we can observe. If the theoretical idea is no good, this process will grind to a halt before long; it will be found impossible to develop it and hang onto it without continually insulating it from observation, as opposed to exposing it.
So for me, it is important to look not just at single hypotheses, but at the development and modification of ideas over time. Really big ideas like biological evolution, that atomic theory of matter, the Marxist theory of history, the Freudian struggle in the unconscious, and an intelligent designer of the universe all have the potential to be handled scientifically. Some of them are, and some of them aren't. The most important objection to the Intelligent Design movement is not that the very idea of intelligent design is linked to supernatural causes in a way that makes it intrinsically unscientific. The problem is that the idea has in fact been handled in a way that gets less and less scientific as time passes.
Peter Godfrey-Smith is a philosopher of science at Harvard and Australian National University. He'll be our guest on Tuesday's upcoming show, "What is Science?"