It is a long-standing and commonly held belief in the United Kingdom and elsewhere that the use of elite forms of language reflects superior intellect and education. Expert opinion from sociolinguistics, however, contends that such a view is the result of middle-class bias and cannot be scientifically justified. In the 1960s and 1970s, such luminaries as Labov and Trudgill were at pains to point out to educationalists, with some success, that this "deficit" view of working-class children's communicative competence is not a helpful one. However, a close reading of recent think-tank reports and policy papers on language and literacy teaching in schools reveals that the linguistic deficit hypothesis has resurfaced and is likely to influence present-day educational policy and practice. In this paper, I examine in detail the findings, claims and recommendations of the reports and argue that they are biased, poorly researched and reflect the vested interests of certain specialist groups. I further argue that we need to move away from the pathologisation of working-class children's language patterns and, once again, inject a sociolinguistic perspective on language and educational failure into the debate. (Contains 7 notes.)