Bas van Fraassen is not your typical scientific anti-realist. Against the positivists, van Fraassen doesn’t want to admit that two different theories which predict all the same observations and experiences, are semantically equivalent, and thus he rejects the positivist construal of science according to which science needs to be ‘properly’ understood because science is to be ‘properly’ construed (where properly means according to the positivist theory of semantics). Instead, he agrees with the realist that science is to be literally construed. He says that a good definition of scientific realism might go as follows:
Scientific Realism: Science aims to give us theories which are literally true, and the acceptance of a scientific theory commits one to it’s literal truth.
On this definition somebody could be a scientific realist without accepting all (or even any) modern/current scientific theories. However, it does mean that once a scientific theory is accepted, the one accepting the theory is committing herself to the literal truth of it’s story. van Fraassen is concerned, however, about belief in unobservables being difficult to justify epistemically. Though we have empirical evidence for electrons, for instance, it is not the case that we have ‘access’ via the five senses to electrons. Some might suggest that if we have no access to some entities via the five senses then the entities are not, strictly, empirical, and to broaden empiricism beyond the realm of the five senses is to broaden empiricism, potentially, to everything metaphysical, so that even numbers, sets, or God, are empirical entities.
Thus, van Fraassen promotes a form of scientific anti-realism which he has called Constructive Empiricism. The best definition of it might go something like this:
Constructive Empiricism: Science, which should be literally construed, aims to give us theories which are empirically adequate, and acceptance of a scientific theory commits one only to that theory’s being empirically adequate.
Now, on the one hand, this is an attractive construal of science. It has the advantage of not being open to error, such that no scientist who commits him or her self to some theory is, in so doing, erring, even if the theory turns out to be wrong. Science, instead of having a history of failure after failure, as one theory is superseded by another theory and so on apparently ad infinitum, has a history of creating more and more empirically adequate theories. The disadvantage is that none of these theories aim to be literally true, but perhaps van Fraassen could say that there is some terminus to how empirically adequate some theory can be: the terminus is exhaustive empirical adequacy. Science aims to terminate there, in a final scientific story which is exhaustively empirically adequate. Moreover, he could then say that he is a kind of realist if he only said that the exhaustively empirically adequate theory, literally construed, is literally true. He could also simply adopt the view that even the commitment to the most empirically adequate scientific theory, which should be literally construed, need not entail being committed to its truth qua scientific commitment or scientific practice.
I’m not sure if I accept anything like this Constructive Empiricism. I certainly don’t want to be a naive realist who thinks any or all the scientific theories of today are closed to reform, or are not tenuous. However, I also want to uphold a confident scientific realism which is optimistic about science. Perhaps I could say, then, that as theories become more and more empirically adequate, or perhaps to the degree that theories are empirically adequate, they approximate to the truth. For example, I myself have no trouble believing in unobservables, like electrons or angels, since I think we have good reasons for believing in both. However, perhaps my belief in the reality of unobservables like electrons is a commitment over and above my ‘scientific’ commitment to atomic theory. If my commitment to atomic theory involves no more than a commitment to it’s empirical adequacy then my commitment to the reality of the entities which the theory postulates becomes a philosophical commitment.
I rather like that.