Richard Marshall interviews Bas van Fraassen in 3:AM Magazine:
3:AM: A naïve realist might assume that science is about finding out what the world is really like and that its theories are either true or false. But you have this idea of ‘empirical adequacy’ that doesn’t take that view, don’t you? Can you first say something about the anti-realism you presented in The Scientific Image so that we get an idea of where you’re thinking is coming from?
BvF: Scientific activity is an important and ubiquitous cultural phenomenon, and has its own criteria of success in practice. Philosophers differ, however, on what those criteria are.
Roughly speaking, scientific realists take the basic criterion to be ontological, they take the ultimate aim to be the construction of true theories, accurate representations, not just of what appears in our measurement outcomes but of what is going on ‘behind the phenomena’. On this view to accept a scientific theory as completely successful would involve belief that everything it postulates is real and everything it says about that is true. This is how it stood in the seventies, since then that first fine careless rapture has been qualified in terms of approximation, asymptotic approach, similarity, and at the extreme end, belief only in significant structure with agnosticism as to a material carrier of this structure, if any. But the basic ideology remains intact: science is all about discovering the true blueprint of the universe.
Empiricists in the philosophy of science take the basic criterion of success in science to be empirical, with the ultimate aim of empirically adequate theories, accurate representation of what is accessible to human observation and manipulation. Theories and the search for explanation are important, but instrumentally, as roads toward greater empirical knowledge. Empiricism is not skepticism, nor anti-realism in any general sense, it is just anti-scientific-realism. Empiricism as it is now can be combined with a ‘common sense realism’ that sees no difficulties in our reference to trees, rocks, persons, lasers, electron microscopes, interferometers, radio telescopes …. or, in their own way, optical phenomena such as rainbows and images produced by microscopes, to give some examples.
Constructive empiricism, the position I advocate, is an empiricism of this sort, a view about what science is that sees overall truth, or truth about what is not observable, as irrelevant to the basic operative criterion of success in the sciences.
More here.