Science is supposed to be (among other things) a self-correcting enterprise. Rather like the legendary invisible hand of the free market gravitating towards fairness and efficiency, science is supposed to gravitate towards the factual. Theories get tested continuously, and sometimes replaced with better theories. You can think of science as rather like Evolution in that sense: just as in Evolution, where the fit survive (and the unfit do not), good scientific theories and models replace the old. Like Evolution, the process is ever upward and onward; a continuum of relentless improvement.

When viewed this way, it's easy to see why some people express so much faith in science as a cure (or eventual cure, at least) for all ills. By definition, the whole process is one of improvement: things can only get better.

If only it were that easy.

Although the selection of scientific theories has certain parallels with Natural Selection, an investigation of the details may reveal that the parallels aren't all advantageous to science. It's clear enough that Natural Selection occurs on the basis of the ability of the organism to perpetuate itself, but what form of selection happens with regards to scientific theories? What sort of selective pressures do scientific theories face?

The most obvious thing that is expected of a scientific theory is the ability to make correct, accurate, and useful predictions. We expect theories of electricity to be accurate models of how electricity behaves so that we can produce nifty electronic devices; we expect theories of physics and gravity to be accurate so that we can put rockets into space and predict the motion of heavenly bodies. In principle, if someone comes up with a model of gravity that explains the facts better than the existing model, the new model will be preferred.

Even this simple and obvious selection criterion conceals some ugly complexities. Sometimes theories interlock into a broad paradigm, and even when one particular theory has been identified as weak and inferior to another, it may still be preferred because it can't be thrown out without also throwing out a bunch of other useful theories. It's been suggested, for example, that gravity travels faster than light, and this seems like a good model for explaining what actually happens, but it contradicts Einstein's General Relativity (which states that nothing can travel faster than light). We can't accept mutually contradictory theories, and the question of one theory being "better than" another is not always clear-cut.

But science isn't just about the practical issues of "what works": there are other, more philosophical issues involved. One very popular criterion for selection of scientific theories involves non-appeal to supernatural causes. Whilst reasonable in a general sense, when taken to an extreme it results in the flat denial that there is such a thing as a supernatural at all: that is to say, it assumes the truth of philosophical materialism. Such materialism may or may not be true, but if compatibility with philosphical materialism is overemphasised, our science will reflect that compatibility more than it will reflect fact, as such.

And this is more or less the point. Natural selection of living organisms occurs on the basis of reproduction: the usual definition of Natural Selection (I make no claim that it's the only definition) is synonymous with reproduction. Selection of scientific theories, on the other hand, is an inherently artificial process, based on some set of values held by scientists about what makes good science. Science will only gravitate towards fact when the scientific selection process causes it to do so. Without due care and attention, we run the serious risk of mistaking some current theory or philosophical ideal as fact. We expect science to illuminate facts, but its so doing is contingent on us practicing good science.

The existing culture of science seems to be ever vigilant against the threat of contamination by supernaturalist explanations, almost to the extent that you can't suggest any theory that would be compatible with God in a general sense (with particular reference to scientific theories of origins). Even "Intelligent Design Theory", which attempts to be utterly non-specific about the identity of the designer, tends to get hurled out immediately as thinly veiled creationism (where creationism is deemed bad because it refers to a supernatural creator). The converse is not true, however: naturalism has been accepted as such a good thing that more naturalism must be better. Naturalistic Evolution is preferred over Intelligent Design Theory not primarily as a matter of compatibility with the evidence, but as a matter of the definition of science; a loss of distinction between science and philosophical materialism.

In short, we appear to be in a state where most scientists believe in philosophical materialism because it is most compatible with scientific facts, and scientific fact is determined by (among other things) its compatibility with philosophical materialsim. Whilst not quite a case of circular reasoning, this is rather similar to the old Scooby Doo sight gag where Scooby and Shaggy leap with fright into each other's arms, each failing to realise that the other is no longer making contact with the ground.

So what is really going on here? Has philosophical materialism become a contaminating (rather than helpful) factor in scientific selection, or is it somehow inseparable from the scientific method? If the latter, then do we know that philosophical materialism is a true belief and tends to lead towards fact, or are we just creating a view of what the world would be like if it were true? Something has to be supporting something here: if Scooby and Shaggy leap into each other's arms and neither of them is touching the ground, it's only a matter of time before they look down and the camera reveals they are hovering, unsupported, in mid-air. We all know what happens next.

Science can be a self-correcting process, but it's not automatic. We create the rules of science, and unless we ensure that they do result in self-correcting behaviour, we may well end up in situations straight out of a Scooby Doo script.