Flights of fancy: a functional linguistic interpretation of semantic gravity and semantic density in secondary school history teaching. J R Martin Erika Matruglio University of Sydney Australia Martin (to appear) discusses four phases of dialogue between Bernstein's sociology of education and systemic functional linguistics - coding orientation, pedagogic discourse, knowledge structure and most recently identity. This paper develops a theme from the third phase, knowledge structure - drawing on discussions collected in Christie & Martin 2007. In particular it focuses on one dimension of Maton's Legitimation Code Theory, semantic gravity and semantic density, in relation to SFL work on technicality and abstraction (especially in relation to grammatical metaphor). This dialogue is a development of Bernstein's late work on horizontal and vertical discourse (Bernstein 1996/2000), and within vertical discourse, his distinction between horizontal and hierarchical knowledge structures. Muller (2007) proposes ‘verticality’ to describe how theories progress - via ever more integrative or general propositions (cf Bernstein's strong/weak internal grammar) or via the introduction of a new language which constructs a ‘fresh perspective, a new set of questions, a new set of connections, and an apparently new problematic, and most importantly a new set of speakers’ (Bernstein 1996:162). In addition Muller proposes ‘grammaticality’ to describe how theoretical statements deal with their empirical predicates (cf. Bernstein's strong/weak external grammar). The stronger the (external) grammaticality of a language, the more stably it is able to generate empirical correlates and the more unambiguous because more restricted the field of referents. Though these concepts highlight features of knowledge structures, they do not analyse their underlying structuring principles. Addressing this problem, Maton (in press) considers the form taken by theories and knowledge structures along two dimensions: (i) semantic gravity, or the degree of context-dependency of meaning; and (ii) semantic density, or the degree of condensation of meaning. He notes that we can talk about of processes of "weakening semantic gravity, as one’s understanding is lifted above the concrete particulars of a specific context or case, and strengthening semantic gravity, as abstract or generalised ideas are made more concrete; and of strengthening semantic density, such as when a lengthy description is ‘packaged up or condensed into a term or brief expression, and weakening semantic density, when an abstract idea is fleshed out with empirical detail.” Processes of this kind are obviously essential to effective teaching and learning in any subject area. Consequently, as part of a project funded by the Australian Research Council investigating secondary school biology and history teaching, we have been concerned with interpreting such processes from a functional linguistic perspective, unpicking the linguistic