Three types of negated constituent movement during the history of English – NegV1, Negative Inversion and Negative Movement – are linked to the role of NegP in the syntax of English up to the early Modern period, building on proposals made by Haegeman (1995), van Kemenade (2000) and Zeijlstra (2004). NegP is analysed as involving the presence of a Neg Operator, null in languages with a head negator. A high NegP licensed NegV1 in Old and early Middle English, the optionality of a Neg Operator in NegP triggered Negative Movement in Late Middle English until NegP was lost, and the loss of NegP in Early Modern English permitted Negative Inversion. The absence of NegP is proposed for the earliest attested stages of English as well as of other Indo-European languages, as a way of accounting for a stage where NegV1 and Negative Concord were absent.