Against faith: The case for science

It appears that the current covid-19 pandemic has determined a decrease in trust and affection to science, and to medicine in particular, because of the lack of consensus which has been observed among scientists (virologists first) and the many mistakes for which doctors and policy makers have been considered responsible, rightly or wrongly.

I think that the problem stems from an incorrect view of what science really is. Science is not a faith, and as such should not be trusted. Hypotheses and theories are scientific precisely if it is possible to falsify them, and possibly easily. That is, if it is possible to test them and eventually prove them to be wrong. Faith instead is needed in order to believe in things which cannot be tested, as it is the case with metaphysics and religion. Political ideologies may eventually be tested, but not so easily, and therefore it is difficult to falsify them. It may be possible to assess ex post the consequences of some government decisions, but it is difficult to confront them with alternatives, to make experiments. Also, very often the supporters of a political idea resort to ad hoc hypotheses in order to corroborate it. Therefore, politics can hardly be considered scientific.

This does not imply per se any negative judgement on non-scientific endeavors, as there are many beautiful and positive things which we can enjoy and that are not scientific, like arts and democratic freedoms. But the definition of science should be narrowed down to those ideas which can be tested empirically by means of rigorous experiments. Only one is necessary to reject a hypothesis as false, and a single well conducted study can falsify all the previous ones. And in most cases (Popper would say always), a scientific hypothesis will be falsified, because we cannot rule out that sooner or later it will not pass a test.

I do not fully agree with the Austrian philosopher here, since any problem is a problem of time (Kant, more or less). Since we do not have an infinite amount of time ahead of us, we cannot be certain that any hypothesis will be falsified, but we may remain with the doubt that it is true. Also, I do not agree with him on his assertion that making mistakes will always lead us closer to the truth. It may happen in most cases, but not necessarily, as the mistakes we can make are virtually infinite. Therefore, testing ideas diminishes the possibility to make the same mistakes, but not new ones.

So, science does not produce truth. It searches for it, by means of trial an error, per speculum et in enigmate (St. Paul), but it is not the truth, which is not evident to us. Nonetheless, we may accept a hypothesis at least temporarily as plausible, i.e. compatible with the observed evidence. If it is not, then it should be rejected as false. Why should we even bother about science then? Because even if scientific theories are not the truth, some are better than others and definitely better than those ideas which cannot even be tested. The best hypotheses - the most robust - so far are those which have withstood the most rigorous and challenging experiments, i.e. attempts to falsify them.

It is therefore important to understand the real meaning of systematic reviews, which are cited so often: their aim is not to determine the most robust hypotheses, but those with the largest consensus. Beside the fact that the lack of consensus is the norm in science (as it is in democracy), robust hypotheses do not always have a large consensus: they are exceptions at some point in time. If we could make a "review" at the time of Copernicus, most astronomers sticked to the Ptolemaic or geocentric hypothesis (with some rare exceptions, of course), which was later proved to be wrong. Then, the Copernican or heliocentric hypothesis was believed to be true, but later on astronomers realised that even that was wrong, since the sun roams around and it is not at the center of the universe. Similarly, after Newton published his laws, people (including scientists and philosophers like Kant) believed that they were true beyond any doubt (and that is why they were called "laws"). We had to wait for the XX century and for Einstein to prove that Newton's laws were wrong, even if they were a very robust theory and an excellent approximation in most cases (within limited speed and distance).

This does not mean that we should stop doing reviews, as they are both useful and interesting, in order to understand the direction that current research has taken on a topic (I have recently published two reviews on covid-19, here and here). But to support the opinion that reviews are capable of sorting out the most robust hypotheses or - even worse - that science tells the truth does not do any good to science. On the contrary, it will definitely lead to a loss of interest in it, at least from the general public, when they will realise that both the above statements are false and cannot be related to science.

It should be clear at this point that I am not questioning faith in religion (which is fine) but faith in science. Science is not an ideology (like scientism) or a system of belief, akin a religion. Science must be doubted in principle, because only falsifiable theories are scientific, and science progresses when a new more robust hypothesis replaces one which has been demonstrated to be false. As Novalis once said: "If you want the truth, you must fight it".