This paper examines the report of the "Inquiry into the Teaching of Literacy" (Department of Education, Science & Training (DEST) 2005) and explores the claims it makes about reading pedagogy and the centrality of particular "methods" or "approaches" to teaching backed by "scientific" evidence. Discourse analysis of the report shows that its logics allow only certain kinds of evidence to count in policy, and that it reduces difficult social and political issues to questions of technique. This allows the report to recommend an approach whereby qualitative insights and practitioners' experience can be bypassed through valorising methods developed and verified by scientific researchers. The report's claims are considered genealogically in the light of historical cases from the early nineteenth century, where educational reformers struggled with the issue of how to educate the children of the poor. In one, the monitorial system promoted by Lancaster in England, there was a focus on reading which made teachers or monitors artefacts of a standardised method. By way of contrast, in Scotland, a classroom approach developed by Stow (1854) made the teacher central to the process, as someone who sensitively interpreted and extended students' experiences with texts. Stow's approach would form the model for the modern classroom in compulsory state schooling, while the monitorial system would eventually be abandoned as ineffective. The historical cases demonstrate the dangers of approaches to policy that fail to account for the complex interplay between teacher, student and text in the reading lesson.