Topics Science

Question Are there ever good reasons to believe in hypotheses which are not falsifiable (i.e., "scientific")? Is it implicit in the idea of a falsifiable hypothesis that we should throw non-falsifiable claims by the wayside?

Accepted:
May 11, 2007

Comments

The notion of a "falsifiable" hypothesis is very difficult to make precise in a way that allows us to count as "falsifiable" all and only those hypotheses that we are inclined to regard as "scientific." Take the hypothesis that the total amount of energy in the universe is the same at every moment in the history of the universe ("energy conservation"). No observations would ever be enough, all by themselves, to demonstrate that the total amount of energy has changed. There would be no way to prove, on the basis of observations alone, that we have failed to take into account some previously hidden form of energy that would suffice to "balance the books." This is not a craftily chosen example. Typically, in mature sciences, when an individual hypothesis (such as energy conservation) is tested against observations, other hypothesis are needed in order to conduct the test. So if our observations fail to go as we expect, there is the possibility of putting the blame on those other hypotheses, rather than on the hypothesis that we were officially testing. Therefore, it is rare in mature sciences for an observation to be able to "falsify" a given hypothesis, if falsifying it requires proving it to be false.

Of course, it is important in science to have hypotheses that are able to be confirmed or disconfirmed by possible observations, where "confirmation" is much less difficult to achieve than "proof". But even a hypothesis that we have no way to confirm or to disconfirm at present may still be scientifically very valuable in leading us to formulate hypotheses that are more easily testable. A hypothesis like the atomic theory was very fruitful indeed, and so it would have been folly to "throw it by the wayside" in those days before it could be subjected to anything like a powerful confirmation or disconfirmation.

It seems to me that there is no very good way to distinguish "scientific" from "nonscientific" hypotheses. Some hypotheses are better confirmed than others at present, some hypotheses have been very strongly disconfirmed (or very strongly confirmed). But no hypothesis about the world is in principle immune to being confirmed (or disconfirmed).

Question Are there ever good reasons to believe in hypotheses which are not falsifiable (i.e., "scientific")? Is it implicit in the idea of a falsifiable hypothesis that we should throw non-falsifiable claims by the wayside? Accepted:May 11, 2007

Question Are there ever good reasons to believe in hypotheses which are not falsifiable (i.e., "scientific")? Is it implicit in the idea of a falsifiable hypothesis that we should throw non-falsifiable claims by the wayside?

Accepted:
May 11, 2007