This text presents a defence of scientific realism against sceptical and positivist attacks. It relies on the importance of models in scientific work. Using the "type hierarchy", a technique of knowledge representation drawn from Artificial Intelligence, it provides an account of these models.
