摘要：A spectre is haunting linguistics and education: the prematurely buried legacy of Basil Bernstein, Michael Halliday, Ruqaiya Hasan and Karl Marx. On the one hand, this spectre seems to demand that we treat both language and learning as natural wholes, instead of reducing them to natural sciences like phonetics or neuroscience and social sciences like discourse analysis and curriculum studies and then trying to hook them up again through interdisciplinary links. On the other, this spectre requires us to work in a world where knowledge is already profoundly fragmented, where choosing to learn one fragment rather than another does make a difference in the health and welfare of the teacher and especially of the learner. Let this paper serve as a manifesto defending the maligned work of Halliday and Hasan, and also as a manual to demonstrate how their analytical approach exposes the fragmentation of knowledge and shows how it can be overcome.
