/ 5 April 2002

The science of the matter

I recommend that Thabo Mbeki read The Logic of Scientific Discovery (1959) by Karl Popper and The Methodology of Scientific Research Programmes (1978) by Imre Lakatos.

It is logically impossible to verify or “prove” a scientific hypothesis. This is why Popper proposed “falsifiability” as a criterion for good science: a hypothesis is scientific if it can in principle be tested and falsified. We can never know if a hypothesis is true (even if it is true), but we can at least reject hypotheses that are false.

Lakatos points out that we tend to have more faith in a hypothesis that can predict novel facts that competing hypotheses cannot predict. So if a hypothesis predicts that a child with HIV will die of Aids, we tend to believe that the hypothesis may be true if the child then dies of Aids.

If empirical data show a strong correlation between the use of a particular medication and a reduction in transmission of HIV from mother to child, we tend to believe that it works, even if we cannot “prove” it.

Yes, we cannot “prove” that there is a causal link between HIV and Aids (even if it is true). But in the light of an overwhelming empirical correlation between HIV and Aids, it is rational to act on the assumption that HIV causes Aids.

Given the possibility that millions of people may die of Aids, politicians cannot wait for “scientific proof” that can never be achieved. Until someone comes up with a better hypothesis, politicians must act on the assumption that the HIV/Aids hypothesis is the best we have.

Mbeki must stop wasting time on futile academic debates and get on with doing his job. Louis Liebenberg, Noordhoek