If our scientific picture is changing rapidly, and large parts of what we currently believe in science are wrong or incomplete, then why rely on the current state of scientific knowledge?
The answer: it’s better than nothing, and it’s the best we have.
This is true in some areas. Yet, for a large swath of science (nutritional science, psychology, biology, archaeology, and so on) there are other significant sources of possible knowledge. The first and most obvious, which applies to an area like psychology, is our intuitions about our own mental life. In areas like nutritional science, there are traditions about how to eat. In archaeology, there are myths or stories which involve archaeological aspects (consider the possible discovery of The Iliad‘s Troy, after critical history deigned it a legend). Beyond this, there are large amounts of anecdote (an + ecdote = un + published) or written works which impinge on various areas science studies, but haven’t been developed up into ‘scientific’ case studies, say (Darwin’s reading an animal breeder pamphlet to suggest ideas on change in animal form is an example of this in action).
(Indeed, the way science changes is often by noticing things that are outside science, and then working to incorporate them. If science is largely wrong or incomplete, then the importance of such things as a set seems greater than the current scientific beliefs.)
In these latter sorts of cases, the question becomes more complex: how much do we weight the current picture of science, and how much do we weight these other sources of possible information?
I think there is no simple answer, but if there’s a seeming divergence between science and some other seemingly important source of information, there seem to be a few questions that can be used as rules of thumb when evaluating scientific claims: 1. What is the robust domain of this? For example, Newtonian physics is robust when applied to certain kinds of causal processes on the surface of the planet, say. However, going from that to “the universe runs on Newtonian processes” is probably unwarranted (and as it turns out, we no longer believe this). Being robust means that the findings have been tested and confirmed extensively for a given set of phenomena. Very often, we find out that a theory that worked well to explain one domain doesn’t work as well when expanded to other domains. (In some cases, it turns out there is no robust domain – it’s just mistaken.) 2. What is the space between evidence and theory or presentation? (Pretty much every presentation of evidence contains some theory about what the evidence means.) Consider ‘brain maps’ which show neurological activity: how exactly are these maps created? What exactly are they showing? How much do the researchers’ own beliefs affect the presentation of the so-called evidence? 3. Following on that, given that there is always a leap from evidence to theory, could there be another way to interpret or explain this evidence?
Combining this with what the other source of evidence might be telling us can lead to a better view of things than just the current scientific view alone.