Falsifiability and Science

This blog entry over at NeuroLogica regarding SETI (the search for extra-terrestrial intelligence) brings up some interesting issues regarding the philosophy of science.

What determines whether or not a researcher should spend time looking for a previously unobserved thing?

What's the difference between trying to observe the following unobserved things:

1) Bigfoot
2) Extraterrestrial intelligence
3) Subatomic particles
4) Leprechauns
5) Ghosts

Assuming we don't want to waste our time looking for something that has a very low chance of actually existing, how do we make that determination? Presumably, we should have some justification for thinking that a given thing might exist even though it has not yet been observed. In the case of extraterrestrial intelligence, we know that life and human-level intelligence arose at least once before. We don't understand well at all how either event happened, so our estimation of how likely either would be to happen again is very murky. But it's reasonably not zero.

I don't know physics very well, but presumably the people looking for unobserved subatomic particles have some justification for thinking that they might exist, based on some particular theory.

What about things like Bigfoot, leprechauns, and ghosts? Well, if we count eyewitnesses and other people's beliefs as evidence, then we have a decent amount of evidence for all of these. Here's a whole bunch of eyewitnesses saying they saw a leprechaun in Alabama:



They even have a sketch of the little guy! I guess when it comes to testimony, we have to actually have some way to assess credibility, yes?

And finally, I thought this comment on the original post was interesting:

One way you can think about the “falsifiability” objection is to turn it on its head.

If the hypothesis is: “There is intelligent life in the universe other than humans,” then no amount of evidence can prove this wrong.

But, if the hypothesis is: “There is NO intelligent life in the universe other than humans,” then you just need a single example to disprove the hypothesis.

It seems like SETI is assuming the latter, and doing their best to disprove it. That’s good science.

So a hypothesis instantly becomes scientific when you add a negative modifier?

According to that logic, the following hypothesis is unscientific: "Leprechauns exist."

While this one is scientifically sound: "Leprechauns don't exist." Because, after all, the observation of a single leprechaun would falsify it. And if a hypothesis is falsifiable, then it has to be scientific, right?

Uh. This is why I'm not an adherent to Karl Popper and the whole falsifiability criterion. A hypothesis is either good or bad depending on whether or not one can justify reasons for proposing it and the extent to which one can find evidence that either supports it, weakens it, or both.