Topics Science

Question What's the status of the so-called "scientific method" among philosophers of science these days? I realize that there are and have been many different methods employed in what we call or want to call scientific investigation, so I appreciate how misleading the singular term might be. But, with that caveat in mind, in school and elsewhere you hear all about this great 'method' we've established. And certainly scientists take themselves to know and share some activity. To put a finer point on this question, let me sketch what I get the impression this 'method' looks like: 1) It's empirical, that is, it involves observation and experimentation. 2) The scientist makes some initial observations, forms a hypothesis, deduces some predictions from it, then designs and performs a "controlled experiment" to "test" them. This experiment is done by attempting to identify variables, some independent, one dependent to ensure (obviously with fallibility) that the appropriate relationship/conditions are being tested. 3) The scientists then draws conclusions about other phenomena based on these results, including positing causal relations and general laws to "explain" the predicted regularities observed. Unless his hypothesis is "falsified", in which case it's back to the (holistic) drawing board. I notice, so sketched, that "the method" involves both 'inductive' and 'hypothetico-deductive' reasoning. What do you make of this? Is that what I've sometimes heard called "rational empiricism"? Are there meaninful generalizations we can make about "scientific method" that might help us identify it and understand the meaning of its results?

Accepted:
August 11, 2010

Comments

"The scientific method" is often poorly or incompletely or misleadingly described in science classes (especially high school science classes). So I'll say a little about that first, and then something about recent philosophical discussions of scientific method.

As you (and many others) describe the method, it begins with "the scientists makes some initial observations and forms an hypothesis." This is typically understood as the inductive part of scientific method. And, while it is true that scientists sometimes start with inductive generalizations, most of the time they start with a deeper hypothesis, one that offers a causal explanation for what is observed. If all we ever did was make inductive generalizations, we'd never get beyond the observable--never get to atoms and magnetic forces and osmotic pressures and all those other invisible entities or abstract concepts that form part of scientific theories. So really, the scientist starts with a hypothesis that is arrived at by abduction ("looking for the best explanation"), then the testing is, as you say, making predictions from the theory, and doing experiments or observations to see whether the predictions are correct.

What I have described is the standard "hypothetico-deductive" account of scientific method. Many philosophers of science still think it is a correct account of good scientific reasoning. Others find it either too idealistic (scientists fail to follow scientific method yet still get good results) e.g. Thomas Kuhn, or insufficient (does not say enough about what is needed for scientific reasoning) e.g. Helen Longino or even incorrect e.g. Ronald Giere says that what scientist work with is models, not theories.