involved − policy makers, deliverers and recipients − are integral. The study found that the actors involved are not involved in any conspiratorial or deliberately oppressive way, but they are mostly well-intended and the outcomes unintended. Actually, the actors are involved in a confusing amalgam of more or less conscious contributions that comprise something like, in Foucault’s analogy, ‘immense pieces of machinery, full of impossible cog wheels, belts which turn nothing and wry gear systems: all these things which “don’t work” and ultimately serve to make the thing “work”’ (cited in Gordon, 1980: 257). This paper reveals the centrality of Aboriginal education policy in this machinery that works to produce some patchy improvement in educational outcomes and the problematic outcomes suggested by Pearson.