Philosophy of science

One of my courses this semester is a reading/discussion class, and we're doing a unit on the philosophy of science. This week, we read two papers, one on the merits of multiple working hypotheses, and one on "strong inference", which dealt with inductive reasoning and falsification of hypotheses. We talked about whether or not we as scientists and ecologists and wildlife managers use these concepts in our research and how we could or should apply them.

I think it's a discussion that every scientist and aspiring scientist should have - our group started out making excuses for why we don't use the elegant designs and "crucial experiments" that the strong inference paper talked about. Government regulations, messy systems, and the issues of fundamentals versus nitty-gritty details all came up and were good points. But as a scientist, if you don't ask the questions in the right way, you will waste time and resources trying to solve your problem. This is especially relevant in my field, as boats and field seasons in general are not cheap, and permits to "take" whales are rare and difficult to come by.

My advisor solves this problem by drawing logical trees as she plans a project or an analysis. What is our main question? What data do we need to answer it clearly? What methods could we use to gather this data? (or, in the event that we have a dataset already: what questions can we answer with this data set? Do we need additional information? How do we get that? What steps do we need to take before we can answer the question?), but I don't know that she's ever thought of falsifying a hypothesis rather than proving one. Because as the paper points out, it is impossible to actually PROVE anything in science. Can you prove that the sun will rise in the east tomorrow? No, though logically it probably will, as it has always done so before. However, if the sun happens to rise in the west, the hypothesis that the sun always rises in the east is shot down and disproved. There can be overwhelming evidence that backs up a hypothesis or theory (think evolution by natural selection), but it is still a theory. We might find something that disproves it someday. Lack of evidence is NOT evidence of lack.

On the other hand, something that cannot be disproved is NOT a scientific hypothesis (think "intelligent design"). "Certain things are so complex that they must have been designed by a creator that we can't perceive except through his works" is not falsifiyable. You can't prove that the creator doesn't exist. etc etc etc. This bit has been covered so frequently that I'm not going to go further into it here.

The point of the discussion was to think critically about the design of your experiments and your hypotheses before you actually get into the nitty gritty work. Simple and elegant is the thing to strive for, even if your work deals with messy, restricted systems that don't easily lend themselves to uncomplicated tests.