Why do scientists value evidence? The question may appear absurd, and the answer blindingly obvious: because evidence demonstrates the truth of theories. But is this really true?
I'm currently reading James Ladyman's textbook, 'Understanding Philosophy of Science' which outlines a number of debates on the nature of evidence and the practice of science in general. In this post I'll relate some highlights gleaned from the chapters on 'inductivism' and 'falsificationism' (friendly definitions to follow, I promise).
A popular conception of science sees scientists making a large number of observations regarding a particular phenomena and then proceeding to make generalisations that account for every instance. For example, if time and time again it is observed that metals of every type expand upon heating it would seem sensible to conclude that the statement 'all metals expand when heated' is true. This extraction of generalisations from evidence is called 'induction'. 'Inductivism' is the view that science may be defined by this method.
There is, however, a long-standing debate within the philosophy of science about the validity of induction and inductivism. The various objections to evidence-based generalisations arise from the following, comparatively simple argument: No amount of positive evidence can ensure against the eventuality that a negative instance may yet be found. For instance, it is conceivable that a metal may yet be discovered that does not expand upon heating. Extending the argument into the realm of metaphysics, the eighteenth-century Scottish philosopher David Hume goes so far as to suggest that all scientific generalisations are based upon the assumption that the future will resemble the past, and that we have no rational reason to believe that it will. So even if we had subjected all metal in the universe to heating and found that every sample conformed to our expectations of expansion, we would still - according to Hume - have no reason extract any kind of general law. Perhaps the next time we hold a lump of copper over a Bunsen burner it might shrink. This certainly seems very counter-intuitive - but drawing attention to nature of our intuitions is Hume's intention. We only believe that the future will resemble the past because it has always done so in the past. Though this certainly seems a very good assumption, Hume's point is that we must admit that it is an assumption in that it cannot conceivably be justified by evidence.
What such objections to induction have in common is the challenge they pose to science. If it is true that a theory may at any time be disproved by a negative instance (whether or not this involves the regularity of the universe unravelling, or simply new evidence coming to light), then the ability of scientists to pronounce with certainty upon any subject appears fatally undermined.
Perhaps the most successful rebuttal to this argument is Karl Popper's 'falsificationism'. Popper, a twentieth-century British philosopher, sought to undermine, rather than solve, the problem of induction by suggesting that science is never about proving theories to be true; quite the reverse. Science, according to Popper, should busy itself with the falsification of theories; with the whittling down of the available possibilities. Truth is only ever to be approached and never claimed absolutely. The best theories are those battle-hardened formulations that have survived whatever tests have thus far been devised. They are not to be considered correct; merely the least-wrong.
An important consequence of this argument is that scientific theories must be potentially falsifiable. If one suggests a theory that could not be proven wrong under any circumstances, then no debate can take place and truth is reduced to a matter of assertion. Popper was especially scathing towards Marxism and psychoanalysis on these grounds. If, for instance, a government made some efforts to look after its nation's poor, then a die-hard Marxist could simply explain away this apparent contradiction of their favoured theory as an attempt by the ruling elite to thwart the oncoming proletariat revolution. Likewise, those expressing criticisms of psychoanalysis may be dismissed by its practitioners as suffering from deep-rooted repression.
Critics of falsificationism have pointed out that scientists do, in some cases, appear to believe things for positive reasons. Many successful theories posit the existence of things that cannot be directly observed; atoms, black holes, DNA and so forth. According to Popper, such entities are merely conceptual devices employed by the least-wrong theories to make predictions. A true adherent of falsificationism cannot assert the literal truth of their existence - and yet many scientists do. Speaking personally, I think this splitting hairs: it is possible that in conversations, especially with journalists, scientists may simply use the word 'exist' as shorthand for 'is a reasonable inference of our least-wrong theory'.
Another criticism of falsificationism is that some scientific principles are not falsifiable, and whose apparent violation would send scientists seeking any other explanation but the refutation of the theory. One such example is the principle of the conservation of energy, which states that energy may take different forms but is never created or destroyed. If a system is observed to be creating energy from nothing, scientists would rather question the accuracy of their observations, or postulate the existence of some non-observable energy source interfering with their measurement devices. In such instances, scientists do seem to value the sheer weight of positive confirmations of a theory over one negative instance. The danger here is that such scientists are indistinguishable from self-deluding Marxists. I would argue, however, that it is certainly sensible to make sure that all options have been considered before abandoning a long-standing principle. One may remain open-minded to the prospect of its refutation at the same time as exploring the possibilities for its continued relevance.